Electronic supplementary material

Groundwater arsenic removal using granular TiO2 Integrated laboratory and

field study

Jinli Cuia middot Jingjing Dua middot Siwu Yub middot Chuanyong Jinga middot Tingshan Chanc

aState Key Laboratory of Environmental Chemistry and Ecotoxicology Research

Center for Eco-Environmental Sciences Chinese Academy of Sciences Beijing

100085 P R China

bGuizhou Electric Power Testing amp Research Institute Guiyang 550002 P R China

cNational Synchrotron Radiation Research Center HsinChu 300 Taiwan

Corresponding author Chuanyong Jing

Tel +86 10 6284 9523

Fax +86 10 6284 9523

E-mail cyjingrceesaccn

Co-corresponding author Tingshan Chan

Tel +886 3 578 0281 Ext 7304

Fax +886 3 578 3805

E-mail chantsnsrrcorgtw

1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

Table S1 Summary of As removal from geogenic groundwater using granular adsorbents in column studiesa

MediaEBCT

(min)

Media

massBV10

q(BV10)

(mgg)

As(III)

(μgL)

As(V)

(μgL)

Total As

(μgL)

Sili

ca

P(μg

L)

Alkalinity

(mgCaCO3L)Ca Mg Fe pH Reference

GTiO2 011 252 0 0 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 022 267 158 013 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 032 279 247 018 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 054 285 527 04 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 108 336 968 096 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 5 1000 843 058 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 101 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 78 374 291 665 89 - 387 383 100 047 82 This study

TiO2 25 84 3460 029 56 14 70 33 20 36 16 73 (Hao et al 2009)

TiO2 granulates 3 41500 28 52 52 21 lt001 41 26767-

836(Bang et al 2011)

TiO2 granulates 3 45000 17 43 43 21 210 41 26lt00

2(Bang et al 2005)

TiO2 granulates 097 32000 19 1977-

82(Gupta et al 2010)

MetsorbG 048 158 21000 06 43 43 37 15 87 30000

572 (USEPA 2008)

MetsorbG 028 15 15000 051 28 28 25 lt01 8(Hristovski et al

2007)

MetSorbG 5 28 14000 02 25 25 78(Westerhoff et al

2006)

MetsorbG 057 94 16000 02 215 215 NA 54 342 54003

877 (USEPA 2008)

2

21

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974 (USEPA 2008)

Adsorbia GTO 038 45 10000 225 13 355 51 24 64 16lt00

2585 (USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572 (USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 37 10500 134 96 24 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 37 5800 135 170 50 220

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 37 4300 187 320 90 410

80-

82(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 62 12 23 370 79 000 75 (USEPA 2008)

3

6 3

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585 (USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572 (USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478 (USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 51 10 162 83 51 000 74 (USEPA 2008)

4

7 9

GFH 3 20900 52 52 21 lt001 41 2678-

81(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375 (USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572 (USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478 (USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877 (USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244 012 75 (Maji et al 2012)

5

3

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375 (USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18 lt00 86 (USEPA 2008)

6

25

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478 (USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 2000 52 52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 7300 52 52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite 12960 gt900 01667250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72(Maiti et al 2013a)

Treated laterite 2080 2000 192 1027 1027d20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 785 998 001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

7

oxides

MediaEBCT

(min)

Media

massBV10

q(BV10)

(mgg)

As(III)

(μgL)

As(V)

(μgL)

Total As

(μgL)

Sili

ca

P(μg

L)

Alkalinity

(mgCaCO3L)Ca Mg Fe pH Reference

GTiO2 011 252 0 0 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 022 267 158 013 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 032 279 247 018 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 054 285 527 04 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 108 336 968 096 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 5 1000 843 058 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 101 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 78 374 291 665 89 - 387 383 100 047 82 This study

TiO2 25 84 3460 029 56 14 70 33 20 36 16 73 (Hao et al 2009)

TiO2 granulates 3 4150028 52

52 21 lt001 41 26767-

836(Bang et al 2011)

TiO2 granulates 3 45000 17 43 43 21 210 41 26lt00

2(Bang et al 2005)

TiO2 granulates 097 3200019

1977-

82(Gupta et al 2010)

MetsorbG 048 158 21000 06 43 43 37 15 87 30000

572 (USEPA 2008)

MetsorbG 028 15 15000 051 28 28 25 lt01 8(Hristovski et al

2007)

MetSorbG 5 28 14000 02 25 25 78(Westerhoff et al

2006)

MetsorbG 057 94 16000 02 215 215 NA 54 342 54003

877 (USEPA 2008)

8

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974

(USEPA 2008)

Adsorbia GTO 038 45 10000 22513

355 51 24 64 16lt00

2585

(USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572

(USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 3686 10500 134

9624 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 5800 135

17050 220

80-

82

(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 4300 187

32090 410

80-

82

(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 6212

623 370 79

000

375

(USEPA 2008)

9

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585

(USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572

(USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478

(USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 5110

7162 83 51

000

974 (USEPA 2008)

10

GFH 3 20900 52 52 21 lt001 41 2678-

81

(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808

(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375

(USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572

(USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478

(USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877

(USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244012

375 (Maji et al 2012)

11

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Table S1 Summary of As removal from geogenic groundwater using granular adsorbents in column studiesa

MediaEBCT

(min)

Media

massBV10

q(BV10)

(mgg)

As(III)

(μgL)

As(V)

(μgL)

Total As

(μgL)

Sili

ca

P(μg

L)

Alkalinity

(mgCaCO3L)Ca Mg Fe pH Reference

GTiO2 011 252 0 0 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 022 267 158 013 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 032 279 247 018 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 054 285 527 04 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 108 336 968 096 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 5 1000 843 058 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 101 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 78 374 291 665 89 - 387 383 100 047 82 This study

TiO2 25 84 3460 029 56 14 70 33 20 36 16 73 (Hao et al 2009)

TiO2 granulates 3 41500 28 52 52 21 lt001 41 26767-

836(Bang et al 2011)

TiO2 granulates 3 45000 17 43 43 21 210 41 26lt00

2(Bang et al 2005)

TiO2 granulates 097 32000 19 1977-

82(Gupta et al 2010)

MetsorbG 048 158 21000 06 43 43 37 15 87 30000

572 (USEPA 2008)

MetsorbG 028 15 15000 051 28 28 25 lt01 8(Hristovski et al

2007)

MetSorbG 5 28 14000 02 25 25 78(Westerhoff et al

2006)

MetsorbG 057 94 16000 02 215 215 NA 54 342 54003

877 (USEPA 2008)

2

21

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974 (USEPA 2008)

Adsorbia GTO 038 45 10000 225 13 355 51 24 64 16lt00

2585 (USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572 (USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 37 10500 134 96 24 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 37 5800 135 170 50 220

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 37 4300 187 320 90 410

80-

82(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 62 12 23 370 79 000 75 (USEPA 2008)

3

6 3

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585 (USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572 (USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478 (USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 51 10 162 83 51 000 74 (USEPA 2008)

4

7 9

GFH 3 20900 52 52 21 lt001 41 2678-

81(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375 (USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572 (USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478 (USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877 (USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244 012 75 (Maji et al 2012)

5

3

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375 (USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18 lt00 86 (USEPA 2008)

6

25

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478 (USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 2000 52 52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 7300 52 52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite 12960 gt900 01667250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72(Maiti et al 2013a)

Treated laterite 2080 2000 192 1027 1027d20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 785 998 001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

7

oxides

MediaEBCT

(min)

Media

massBV10

q(BV10)

(mgg)

As(III)

(μgL)

As(V)

(μgL)

Total As

(μgL)

Sili

ca

P(μg

L)

Alkalinity

(mgCaCO3L)Ca Mg Fe pH Reference

GTiO2 011 252 0 0 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 022 267 158 013 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 032 279 247 018 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 054 285 527 04 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 108 336 968 096 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 5 1000 843 058 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 101 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 78 374 291 665 89 - 387 383 100 047 82 This study

TiO2 25 84 3460 029 56 14 70 33 20 36 16 73 (Hao et al 2009)

TiO2 granulates 3 4150028 52

52 21 lt001 41 26767-

836(Bang et al 2011)

TiO2 granulates 3 45000 17 43 43 21 210 41 26lt00

2(Bang et al 2005)

TiO2 granulates 097 3200019

1977-

82(Gupta et al 2010)

MetsorbG 048 158 21000 06 43 43 37 15 87 30000

572 (USEPA 2008)

MetsorbG 028 15 15000 051 28 28 25 lt01 8(Hristovski et al

2007)

MetSorbG 5 28 14000 02 25 25 78(Westerhoff et al

2006)

MetsorbG 057 94 16000 02 215 215 NA 54 342 54003

877 (USEPA 2008)

8

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974

(USEPA 2008)

Adsorbia GTO 038 45 10000 22513

355 51 24 64 16lt00

2585

(USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572

(USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 3686 10500 134

9624 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 5800 135

17050 220

80-

82

(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 4300 187

32090 410

80-

82

(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 6212

623 370 79

000

375

(USEPA 2008)

9

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585

(USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572

(USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478

(USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 5110

7162 83 51

000

974 (USEPA 2008)

10

GFH 3 20900 52 52 21 lt001 41 2678-

81

(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808

(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375

(USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572

(USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478

(USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877

(USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244012

375 (Maji et al 2012)

11

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974 (USEPA 2008)

Adsorbia GTO 038 45 10000 225 13 355 51 24 64 16lt00

2585 (USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572 (USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 37 10500 134 96 24 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 37 5800 135 170 50 220

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 37 4300 187 320 90 410

80-

82(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 62 12 23 370 79 000 75 (USEPA 2008)

3

6 3

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585 (USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572 (USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478 (USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 51 10 162 83 51 000 74 (USEPA 2008)

4

7 9

GFH 3 20900 52 52 21 lt001 41 2678-

81(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375 (USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572 (USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478 (USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877 (USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244 012 75 (Maji et al 2012)

5

3

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375 (USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18 lt00 86 (USEPA 2008)

6

25

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478 (USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 2000 52 52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 7300 52 52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite 12960 gt900 01667250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72(Maiti et al 2013a)

Treated laterite 2080 2000 192 1027 1027d20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 785 998 001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

7

oxides

MediaEBCT

(min)

Media

massBV10

q(BV10)

(mgg)

As(III)

(μgL)

As(V)

(μgL)

Total As

(μgL)

Sili

ca

P(μg

L)

Alkalinity

(mgCaCO3L)Ca Mg Fe pH Reference

GTiO2 011 252 0 0 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 022 267 158 013 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 032 279 247 018 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 054 285 527 04 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 108 336 968 096 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 5 1000 843 058 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 101 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 78 374 291 665 89 - 387 383 100 047 82 This study

TiO2 25 84 3460 029 56 14 70 33 20 36 16 73 (Hao et al 2009)

TiO2 granulates 3 4150028 52

52 21 lt001 41 26767-

836(Bang et al 2011)

TiO2 granulates 3 45000 17 43 43 21 210 41 26lt00

2(Bang et al 2005)

TiO2 granulates 097 3200019

1977-

82(Gupta et al 2010)

MetsorbG 048 158 21000 06 43 43 37 15 87 30000

572 (USEPA 2008)

MetsorbG 028 15 15000 051 28 28 25 lt01 8(Hristovski et al

2007)

MetSorbG 5 28 14000 02 25 25 78(Westerhoff et al

2006)

MetsorbG 057 94 16000 02 215 215 NA 54 342 54003

877 (USEPA 2008)

8

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974

(USEPA 2008)

Adsorbia GTO 038 45 10000 22513

355 51 24 64 16lt00

2585

(USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572

(USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 3686 10500 134

9624 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 5800 135

17050 220

80-

82

(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 4300 187

32090 410

80-

82

(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 6212

623 370 79

000

375

(USEPA 2008)

9

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585

(USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572

(USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478

(USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 5110

7162 83 51

000

974 (USEPA 2008)

10

GFH 3 20900 52 52 21 lt001 41 2678-

81

(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808

(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375

(USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572

(USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478

(USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877

(USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244012

375 (Maji et al 2012)

11

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

6 3

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585 (USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572 (USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478 (USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 51 10 162 83 51 000 74 (USEPA 2008)

4

7 9

GFH 3 20900 52 52 21 lt001 41 2678-

81(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375 (USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572 (USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478 (USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877 (USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244 012 75 (Maji et al 2012)

5

3

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375 (USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18 lt00 86 (USEPA 2008)

6

25

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478 (USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 2000 52 52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 7300 52 52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite 12960 gt900 01667250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72(Maiti et al 2013a)

Treated laterite 2080 2000 192 1027 1027d20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 785 998 001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

7

oxides

MediaEBCT

(min)

Media

massBV10

q(BV10)

(mgg)

As(III)

(μgL)

As(V)

(μgL)

Total As

(μgL)

Sili

ca

P(μg

L)

Alkalinity

(mgCaCO3L)Ca Mg Fe pH Reference

GTiO2 011 252 0 0 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 022 267 158 013 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 032 279 247 018 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 054 285 527 04 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 108 336 968 096 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 5 1000 843 058 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 101 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 78 374 291 665 89 - 387 383 100 047 82 This study

TiO2 25 84 3460 029 56 14 70 33 20 36 16 73 (Hao et al 2009)

TiO2 granulates 3 4150028 52

52 21 lt001 41 26767-

836(Bang et al 2011)

TiO2 granulates 3 45000 17 43 43 21 210 41 26lt00

2(Bang et al 2005)

TiO2 granulates 097 3200019

1977-

82(Gupta et al 2010)

MetsorbG 048 158 21000 06 43 43 37 15 87 30000

572 (USEPA 2008)

MetsorbG 028 15 15000 051 28 28 25 lt01 8(Hristovski et al

2007)

MetSorbG 5 28 14000 02 25 25 78(Westerhoff et al

2006)

MetsorbG 057 94 16000 02 215 215 NA 54 342 54003

877 (USEPA 2008)

8

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974

(USEPA 2008)

Adsorbia GTO 038 45 10000 22513

355 51 24 64 16lt00

2585

(USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572

(USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 3686 10500 134

9624 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 5800 135

17050 220

80-

82

(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 4300 187

32090 410

80-

82

(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 6212

623 370 79

000

375

(USEPA 2008)

9

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585

(USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572

(USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478

(USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 5110

7162 83 51

000

974 (USEPA 2008)

10

GFH 3 20900 52 52 21 lt001 41 2678-

81

(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808

(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375

(USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572

(USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478

(USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877

(USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244012

375 (Maji et al 2012)

11

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

7 9

GFH 3 20900 52 52 21 lt001 41 2678-

81(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375 (USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572 (USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478 (USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586 (USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877 (USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244 012 75 (Maji et al 2012)

5

3

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375 (USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18 lt00 86 (USEPA 2008)

6

25

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478 (USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 2000 52 52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 7300 52 52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite 12960 gt900 01667250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72(Maiti et al 2013a)

Treated laterite 2080 2000 192 1027 1027d20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 785 998 001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

7

oxides

MediaEBCT

(min)

Media

massBV10

q(BV10)

(mgg)

As(III)

(μgL)

As(V)

(μgL)

Total As

(μgL)

Sili

ca

P(μg

L)

Alkalinity

(mgCaCO3L)Ca Mg Fe pH Reference

GTiO2 011 252 0 0 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 022 267 158 013 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 032 279 247 018 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 054 285 527 04 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 108 336 968 096 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 5 1000 843 058 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 101 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 78 374 291 665 89 - 387 383 100 047 82 This study

TiO2 25 84 3460 029 56 14 70 33 20 36 16 73 (Hao et al 2009)

TiO2 granulates 3 4150028 52

52 21 lt001 41 26767-

836(Bang et al 2011)

TiO2 granulates 3 45000 17 43 43 21 210 41 26lt00

2(Bang et al 2005)

TiO2 granulates 097 3200019

1977-

82(Gupta et al 2010)

MetsorbG 048 158 21000 06 43 43 37 15 87 30000

572 (USEPA 2008)

MetsorbG 028 15 15000 051 28 28 25 lt01 8(Hristovski et al

2007)

MetSorbG 5 28 14000 02 25 25 78(Westerhoff et al

2006)

MetsorbG 057 94 16000 02 215 215 NA 54 342 54003

877 (USEPA 2008)

8

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974

(USEPA 2008)

Adsorbia GTO 038 45 10000 22513

355 51 24 64 16lt00

2585

(USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572

(USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 3686 10500 134

9624 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 5800 135

17050 220

80-

82

(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 4300 187

32090 410

80-

82

(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 6212

623 370 79

000

375

(USEPA 2008)

9

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585

(USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572

(USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478

(USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 5110

7162 83 51

000

974 (USEPA 2008)

10

GFH 3 20900 52 52 21 lt001 41 2678-

81

(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808

(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375

(USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572

(USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478

(USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877

(USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244012

375 (Maji et al 2012)

11

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

3

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375 (USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18 lt00 86 (USEPA 2008)

6

25

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478 (USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 2000 52 52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 7300 52 52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite 12960 gt900 01667250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72(Maiti et al 2013a)

Treated laterite 2080 2000 192 1027 1027d20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 785 998 001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

7

oxides

MediaEBCT

(min)

Media

massBV10

q(BV10)

(mgg)

As(III)

(μgL)

As(V)

(μgL)

Total As

(μgL)

Sili

ca

P(μg

L)

Alkalinity

(mgCaCO3L)Ca Mg Fe pH Reference

GTiO2 011 252 0 0 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 022 267 158 013 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 032 279 247 018 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 054 285 527 04 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 108 336 968 096 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 5 1000 843 058 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 101 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 78 374 291 665 89 - 387 383 100 047 82 This study

TiO2 25 84 3460 029 56 14 70 33 20 36 16 73 (Hao et al 2009)

TiO2 granulates 3 4150028 52

52 21 lt001 41 26767-

836(Bang et al 2011)

TiO2 granulates 3 45000 17 43 43 21 210 41 26lt00

2(Bang et al 2005)

TiO2 granulates 097 3200019

1977-

82(Gupta et al 2010)

MetsorbG 048 158 21000 06 43 43 37 15 87 30000

572 (USEPA 2008)

MetsorbG 028 15 15000 051 28 28 25 lt01 8(Hristovski et al

2007)

MetSorbG 5 28 14000 02 25 25 78(Westerhoff et al

2006)

MetsorbG 057 94 16000 02 215 215 NA 54 342 54003

877 (USEPA 2008)

8

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974

(USEPA 2008)

Adsorbia GTO 038 45 10000 22513

355 51 24 64 16lt00

2585

(USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572

(USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 3686 10500 134

9624 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 5800 135

17050 220

80-

82

(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 4300 187

32090 410

80-

82

(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 6212

623 370 79

000

375

(USEPA 2008)

9

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585

(USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572

(USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478

(USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 5110

7162 83 51

000

974 (USEPA 2008)

10

GFH 3 20900 52 52 21 lt001 41 2678-

81

(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808

(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375

(USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572

(USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478

(USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877

(USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244012

375 (Maji et al 2012)

11

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

25

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478 (USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 2000 52 52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 7300 52 52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite 12960 gt900 01667250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72(Maiti et al 2013a)

Treated laterite 2080 2000 192 1027 1027d20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 785 998 001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

7

oxides

MediaEBCT

(min)

Media

massBV10

q(BV10)

(mgg)

As(III)

(μgL)

As(V)

(μgL)

Total As

(μgL)

Sili

ca

P(μg

L)

Alkalinity

(mgCaCO3L)Ca Mg Fe pH Reference

GTiO2 011 252 0 0 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 022 267 158 013 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 032 279 247 018 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 054 285 527 04 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 108 336 968 096 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 5 1000 843 058 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 101 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 78 374 291 665 89 - 387 383 100 047 82 This study

TiO2 25 84 3460 029 56 14 70 33 20 36 16 73 (Hao et al 2009)

TiO2 granulates 3 4150028 52

52 21 lt001 41 26767-

836(Bang et al 2011)

TiO2 granulates 3 45000 17 43 43 21 210 41 26lt00

2(Bang et al 2005)

TiO2 granulates 097 3200019

1977-

82(Gupta et al 2010)

MetsorbG 048 158 21000 06 43 43 37 15 87 30000

572 (USEPA 2008)

MetsorbG 028 15 15000 051 28 28 25 lt01 8(Hristovski et al

2007)

MetSorbG 5 28 14000 02 25 25 78(Westerhoff et al

2006)

MetsorbG 057 94 16000 02 215 215 NA 54 342 54003

877 (USEPA 2008)

8

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974

(USEPA 2008)

Adsorbia GTO 038 45 10000 22513

355 51 24 64 16lt00

2585

(USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572

(USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 3686 10500 134

9624 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 5800 135

17050 220

80-

82

(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 4300 187

32090 410

80-

82

(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 6212

623 370 79

000

375

(USEPA 2008)

9

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585

(USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572

(USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478

(USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 5110

7162 83 51

000

974 (USEPA 2008)

10

GFH 3 20900 52 52 21 lt001 41 2678-

81

(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808

(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375

(USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572

(USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478

(USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877

(USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244012

375 (Maji et al 2012)

11

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

oxides

MediaEBCT

(min)

Media

massBV10

q(BV10)

(mgg)

As(III)

(μgL)

As(V)

(μgL)

Total As

(μgL)

Sili

ca

P(μg

L)

Alkalinity

(mgCaCO3L)Ca Mg Fe pH Reference

GTiO2 011 252 0 0 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 022 267 158 013 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 032 279 247 018 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 054 285 527 04 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 108 336 968 096 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 5 1000 843 058 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 101 374 291 665 89 - 387 383 100 047 82 This study

GTiO2 78 374 291 665 89 - 387 383 100 047 82 This study

TiO2 25 84 3460 029 56 14 70 33 20 36 16 73 (Hao et al 2009)

TiO2 granulates 3 4150028 52

52 21 lt001 41 26767-

836(Bang et al 2011)

TiO2 granulates 3 45000 17 43 43 21 210 41 26lt00

2(Bang et al 2005)

TiO2 granulates 097 3200019

1977-

82(Gupta et al 2010)

MetsorbG 048 158 21000 06 43 43 37 15 87 30000

572 (USEPA 2008)

MetsorbG 028 15 15000 051 28 28 25 lt01 8(Hristovski et al

2007)

MetSorbG 5 28 14000 02 25 25 78(Westerhoff et al

2006)

MetsorbG 057 94 16000 02 215 215 NA 54 342 54003

877 (USEPA 2008)

8

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974

(USEPA 2008)

Adsorbia GTO 038 45 10000 22513

355 51 24 64 16lt00

2585

(USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572

(USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 3686 10500 134

9624 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 5800 135

17050 220

80-

82

(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 4300 187

32090 410

80-

82

(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 6212

623 370 79

000

375

(USEPA 2008)

9

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585

(USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572

(USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478

(USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 5110

7162 83 51

000

974 (USEPA 2008)

10

GFH 3 20900 52 52 21 lt001 41 2678-

81

(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808

(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375

(USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572

(USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478

(USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877

(USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244012

375 (Maji et al 2012)

11

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Adsorbia GTO 057 7 4000 02 51 5110

7162 83 51

000

974

(USEPA 2008)

Adsorbia GTO 038 45 10000 22513

355 51 24 64 16lt00

2585

(USEPA 2008)

Adsorbia GTO 048 79 16000 05 43 43 37 15 87 30000

572

(USEPA 2008)

Adsorbia GTO 038 45 12500 04 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

Adsorbsia GTO 01 16 5288 012 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 025 38 7755 007 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 028 22 29000 034 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbsia GTO 05 4 10575 009 28 28 25 lt01 8(Hristovski et al

2007)

Adsorbia GTO 057 94 22000 03 15 15 NA 54 342 54003

877 (USEPA 2008)

TiO2 pillared

montmorillonite2 3686 10500 134

9624 120

80-

82(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 5800 135

17050 220

80-

82

(Li et al 2012)

TiO2 pillared

montmorillonite2 3686 4300 187

32090 410

80-

82

(Li et al 2012)

GFH 055 188 gt23000 gt1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

GFH 05 161 52000 2 1 61 6212

623 370 79

000

375

(USEPA 2008)

9

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585

(USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572

(USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478

(USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 5110

7162 83 51

000

974 (USEPA 2008)

10

GFH 3 20900 52 52 21 lt001 41 2678-

81

(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808

(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375

(USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572

(USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478

(USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877

(USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244012

375 (Maji et al 2012)

11

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

GFH 068 198 11000 04 51 5110

7162 83 51

000

974 (USEPA 2008)

GFH 25 38 3300 028 50 50 22 156 88(Westerhoff et al

2005)

GFH 25 38 24000 202 50 50 22 156 76(Westerhoff et al

2005)

GFH 022 64 23000 061 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

GFH 022 64 36000 225 13 355 51 24 64 16lt00

2585

(USEPA 2008)

GFH 058 194 50000 14 43 43 37 15 87 30000

572

(USEPA 2008)

GFH 05 169 48000 12 05 395 40 84 19 160 40000

478

(USEPA 2008)

GFH 5 723 30000 168 33 33 39 128 77(Westerhoff et al

2005)

GFH 05 732 28000 039 25 25 78(Westerhoff et al

2006)

GFH 5 100000 18000 02 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

GFH 5 91314 70000 085 13 13 34 195 75(Westerhoff et al

2005)

GFH 25 31399 1500 007 50 50 22 156 88(Westerhoff et al

2005)

GFH 62 8000 51 5110

7162 83 51

000

974 (USEPA 2008)

10

GFH 3 20900 52 52 21 lt001 41 2678-

81

(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808

(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375

(USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572

(USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478

(USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877

(USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244012

375 (Maji et al 2012)

11

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

GFH 3 20900 52 52 21 lt001 41 2678-

81

(Bang et al 2011)

GFO 3 58000 52 52 21 lt001 41 26748-

808

(Bang et al 2011)

E33 055 179 11000 1 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 05 161 40000 18 1 61 6212

623 370 79

000

375

(USEPA 2008)

E33 058 194 44000 14 43 43 37 15 87 30000

572

(USEPA 2008)

E33 05 161 44000 11 05 395 40 84 19 160 40000

478

(USEPA 2008)

E33 05 57 gt34000 071 25 25 78(Westerhoff et al

2006)

E33 022 57 20000 062 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

E33 033 8 25000 04 215 215 NA 54 342 54003

877

(USEPA 2008)

E33 5 4700 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

E33 5 100000 54000 059 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

E33 45 40000 12 1 61 6212

623 370 79

000

375 (USEPA 2008)

Fe-sand 1 332 2400 03 1703 1703 - - - - 12 21 74(Thirunavukkarasu

et al 2003)

Fe-loaded rock 41 48 474 001 40 40 - 1540 30 356 244012

375 (Maji et al 2012)

11

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Fe coated sponge 6 336 062 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 8 263 049 156 156 15 15 15 62(Nguyen et al

2006)

Fe coated sponge 6 168 011 56 56 NA NA 12 21(Nguyen et al

2006)

Fe coated sponge 8 189 013 56 56 NA NA 12 21(Nguyen et al

2006)

HFO-coated GAC 1 12000 139 45 15 6012

5703 59 113

76-

80(Jang et al 2008)

Mn-HFO 105 105 3131979

6167 008 74 (Gupta et al 2010)

Fe-Ce adsorbent 11500 6 71 71 15 07 78 (Dou et al 2006)

Fe residual solids 125 43 43 249 81(Gibbons and

Gagnon 2010)

Z33 05 149 10000 005 25 25 78(Westerhoff et al

2006)

Z33 5 100000 10000 016 33 33 39 14 175 56 12007

877

(Westerhoff et al

2006)

FS50 (Fe-Al) 05 28 6000 004 25 25 78(Westerhoff et al

2006)

AAFS50 075 256 2500 005 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AAFS50 068 231a few

thousand013 1 61 62

12

623 370 79

000

375

(USEPA 2008)

AAFS50 03 87 6700 019 08 402 41 51 33 69 18lt00

2586

(USEPA 2008)

12

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

AAFS50 45 7000 012 05 395 40 84 19 160 40000

478

(USEPA 2008)

AAFS50 5 650 64 15 655 28 51 505 57 1420 81 (USEPA 2008)

AA 25 86 800 007 56 14 70 33 20 36 16 73 (Hao et al 2009)

AA 3 200052

52 21 lt001 41 26(Westerhoff et al

2005)

AA-SH 25 96 3100 023 56 14 70 33 20 36 16 73 (Hao et al 2009)

MAA 3 730052

52 21 lt001 41 2678-

83(Bang et al 2011)

Treated laterite 3 325 3000 117 170 215 385 40 700 54369-

74(Maji et al 2012)

Acid-acitivated

laterite133 20 001 5200 52 22 700 054 78

(Maiti et al

2010b)

Treated laterite12960 gt900 01667

250000

0250

645-

650270-290

30-

70

30-

50

03-

101

71-

72

(Maiti et al

2013b)

Treated laterite2080 2000 192 1027 1027d

20-

30

600-

1000610-650

200-

250

lt0

03

25-

45

74-

78(Maiti et al 2010a)

Fe-mineral and

limestone870 1594 03 4000 400 (Shan et al 2013)

Modified granular

natural siderite45 110 580 854 6654

26

3ND 838 (Zhao et al 2014)

Natural siderite 7859984076

4001 33000 330 2420 213 72 (Maji et al 2008)

Iron-Mn binary

oxides15 40 129 214 343 323

173

7025 786 (Kong et al 2013)

Iron-Mn binary 15 40 145 176 321 22 253 347 754 (Maiti et al 2010a)

13

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

oxides

aUnit for ion concentration is mgL except as notedbestimated from bed volume breakthrough curves and influent As concentration if not explicitly expressed in the reference cThe species is As(V) if only one concentration is given

d

The groundwater As concentration was 385-440 μgL and additional ~600 μgL (As(III)As(V)=11) arsenic was added

14

222324252627

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Fig S1 Boxplot of groundwater As levels in columns and filters in the literature in Table S1 Individual samples are shown as diamonds The range of each box represents the 25th

and 75th

percentile whereas the whiskers represent the 10th

and 90th

percentile

Synthesis of granular TiO2

Granular TiO2 (GTiO2) was prepared by hydrolysis of titanyl sulfate (TiOSO4)

Generally 300 g TiOSO4 was mixed with 1800 mL DI water in a 10 L jar reactor in an

ice bath at 4 oC Then 10 M NaOH was slowly added to adjust the pH to 6 The

precursor of TiO2 was obtained by washing the suspension with DI water several

times till the conductivity of the supernatant was less than 100 μScm Then 800 g

TiO2 precursor was mixed with 40 mL polyvinyl alcohol solution at 80 oC in a water

bath The product was crushed into 60-80 mesh and dried in an oven at 60 oC for 12 h

15

282930313233343536373839404142434445464748

49

50

51

52

53

54

55

56

57585960

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Fig S2 Characterization of GTiO2 SEM (a) and XRD pattern with crystal face for each peak (b)

Fig S1 SEM images for GTiO2 in this study

Table S2 Average concentrations of As coexisting ions (mgL) and water chemistry for groundwater 1 and 2 from two wells

ParametersGroundwater 1

Groundwater 2

As(III) 0374 plusmn 0056 0165 plusmn 0022As(V) 0291 plusmn 0067 0052 plusmn 0010Si 89 plusmn 03 98 plusmn 01Ca 391 plusmn 21 112 plusmn 03Mg 1043 plusmn 42 296 plusmn 05Na 3294 plusmn 186 928 plusmn 97

Cl- 6143 plusmn 392 139 plusmn 12

PO43- lt002 008 plusmn 004

SO42- 1176 plusmn 37 04 plusmn 01

NO3- 34 plusmn 16 12 plusmn 05

F- 12 plusmn 01 07 plusmn 01Br- 23 plusmn 30 26 plusmn 09K 14 plusmn 02 06 plusmn 02Al 006 plusmn 003 014 plusmn 005Fe 047 plusmn 019 004 plusmn 003Mn 012 plusmn 004 006 plusmn 001Alk (mg CaCO3 L) 629 plusmn 20 98 plusmn 12

T (oC) 116 plusmn 03 116 plusmn 03pH 820 plusmn 006 811 plusmn 007

16

61626364656667686970717273

7475

76

7778

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

DO (mgL) 042 plusmn 061 012 plusmn 026ORP -40 plusmn 22 -78 plusmn 23TDS (gL) 175 plusmn 002 039 plusmn 003Conductivity (mScm) 266 plusmn 018 062 plusmn 027

Table S3 Water parameters in the synthetic water used in the studyParameters Ca2+ Mg2+ HCO3

- Na+ Cl- pHConc (mgL) 401 1043 388 2460 3928 82

Table S4 PHREEQC input for a representative column with EBCT=054 min SOLUTION 0

This keyword is to define the influent groundwater chemistry parameters which are included in the adsorption reactions like ldquoArseniterdquo and ldquoArsenaterdquo and the experimental conditions including ldquopHrdquo The information defined in SOLUTION is available for subsequent transport and adsorptive reaction calculations

temp 23 pH 82 pe 4 redox pe units mgL density 1 Al 006 Arsenite 0374 Arsenate 0291 CO4

2-

4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 NO3

-

34 Na 16872

17

79

80

8182

83

84858687888990919293949596979899

100101102103104105106107108109110111112

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

SO42-

1176 Si 89 -water 1 kg

SOLUTION 1-4 The small column was divided into four cells SOLUTION 1-4

are the groundwater chemistry parameters for the four cells temp 23 pH 83 pe 4 redox pe units ppm density 1 Al 006 C(4) 4401 Ca 194 Cl 1456 F 115 Fe 047 K 137 Mg 395 Mn 012 N(3) 34 Na 16872 S(6) 1176 Si 89 -water 1 kg

SOLUTION_SPECIES This keyword is to define the dissociation constant logK

and diffusion constant (dw) in solution for arsenite and arsenate

H3AsO4 = H2AsO4- + H+ log_k -2243-dw 96e-13H3AsO4 = HAsO4-2 + 2H+ log_k -9001-dw 96e-13H3AsO4 = AsO4-3 + 3H+ log_k -20597-dw 96e-13H3ArseniteO3 = H3ArseniteO3 log_k 0-dw 96e-13

18

113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

H3ArseniteO3 = H2ArseniteO3- + H+ log_k -9228 delta_h 656 kcal-dw 96e-13H3ArseniteO3 = HArseniteO3-2 + 2H+ log_k -2133 delta_h 142 kcal-dw 96e-13H3ArseniteO3 = ArseniteO3-3 + 3H+ log_k -34744 delta_h 2025 kcal-dw 96e-13

SURFACE_MASTER_SPECIES GTiO2 Surface species and their inherent surface

charge Surf_s Surf_sOH-033Surf_w Surf_wO-067

SURFACE 1-4 This keyword is to define the amount and composition

of each GTiO2 surface in columns -equilibrate with solution 1-sites DENSITYSurf_sOH-033 1544 196 262

surface sites for Surf_sOH-033 (154 sitesnm2

) surface area (196 m2

g) and mass (262 g GTiO2 in EBCT of 054 min) used in columns

-capacitance 236 5 Capacitance for the 0-1 plane in the CD-MUSIC

formulation of 236 Fm2

and for the 1-2 plane in the CD-MUSIC formulation of 5 Fm2

Surf_wO-067 3

surface sites for Surf_wO-067 (154 sitesnm2

) -cd_music

CD-MUSIC model is used in PHREEQC integrating the parameters from the batch experiments modeling

-donnan 1e-008 The composition of the diffuse layer is calculated

using donnan in CD-MUSIC model

SURFACE_SPECIES

19

157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

This keyword is to define an adsorptive reaction and logK for each composition with surface species

H+ + Surf_sOH-033 = Surf_sOH2+067 log_k 58 -cd_music 1 0 0 0 0Surf_sOH-033 = Surf_sOH-033 log_k 0H+ + Surf_wO-067 = Surf_wOH+033 log_k 58 -cd_music 1 0 0 0 0Surf_wO-067 = Surf_wO-067 log_k 0Na+ + Surf_sOH-033 = Surf_sOHNa+067 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_sOH-033 = Surf_sOH2Cl-033 log_k 48 -cd_music 1 0 -1 0 0Na+ + Surf_wO-067 = Surf_wONa+033 log_k -1 -cd_music 0 0 1 0 0Cl- + H+ + Surf_wO-067 = Surf_wOHCl-067 log_k 48 -cd_music 1 0 -1 0 0Ca+2 + Surf_sOH-033 = Surf_sOHCa+167 log_k 45 -cd_music 0 0 0 01 2Ca+2 + Surf_wO-067 = Surf_wOCa+133 log_k 1 -cd_music 0 0 2 0 0 Ca+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHCaAsO4H-033 log_k 3 -cd_music 0 -2 0 01 2 Ca+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHCaArseniteO3H2+067 log_k 3

-cd_music 0 -1 0 01 2 Mg+2 + Surf_sOH-033 = Surf_sOHMg+167 log_k 4 -cd_music 0 0 0 01 2Mg+2 + Surf_wO-067 = Surf_wOMg+133 log_k 1 -cd_music 0 0 2 0 0Mg+2 + Surf_sOH-033 + HAsO4-2 = Surf_sOHMgAsO4H-033

20

201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

log_k 3 -cd_music 0 -2 0 01 2Mg+2 + Surf_sOH-033 + H2ArseniteO3-1 = Surf_sOHMgArseniteO3H2+067 log_k 2 -cd_music 0 -1 0 01 2H4SiO4 + 2Surf_sOH-033 = (Surf_sO)2Si(OH)2-066 + 2H2O log_k 1508 -cd_music -2 -2 0 05 4CO3-2 + 2H+ + 2Surf_sOH-033 = (Surf_sO)2CO-066 + 2H2O log_k 22 -cd_music -2 -2 0 033 4H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

Arsenite adsorption equation on GTiO2 surface H3ArseniteO3 is the primary species of As(III) at the groundwater pH 82 (Surf_sO)2ArseniteOH-066 is the adsorptive bidentate binuclear structure of As(III) on GTiO2

log_k 1635 The adsorptive reaction constant for As(III) on GTiO2 in

the current condition which could be changed with different experimental conditions

-cd_music -2 -1 0 066 3 CD-MUSIC model is integrated in the PHREEQC

modeling The numbers represent -2 the change in charge at the plane of Surf_sOH due to loss of two hydrogens -1 the change in charge at 1 plane due to gain of one hydrogen and one oxygen on Arsenite 0 the change in charge at the 2 plane usually caused by the outer sphere adsorption 066 the fraction of the central ion (As) charge that is associated with plane 0 which reflecting the charge distribution 3 the charge on the central ion (As)

H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 18 -cd_music -2 -4 0 05 5H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsO2-166 + 2H2O log_k 173 -cd_music -2 -4 0 05 52H+ + HAsO4-2 + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 29 -cd_music -2 -3 0 05 5H+ + H2AsO4- + 2Surf_sOH-033 = (Surf_sO)2AsOOH-066 + 2H2O log_k 24 -cd_music -2 -3 0 05 5

21

245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

TRANSPORT This keyword is to simulate one-dimensional

transport of groundwater composition in the process of dispersion diffusion and diffusion into GTiO2 porosity

-cells 4 The number indicates four cells in the small

column -shifts 10403

The number indicates the shifts numbers in the transport simulation

-time_step 48 seconds Time step defines the time period for each shift

The total column performance time is shifts times time step

-boundary_conditions constant flux This word is to define the concentration of each

composition during transport -lengths 40017

This word is to define the length of each cell (m) -dispersivities 40001

This word is to define dispersivity of each cell for transport simulations

-correct_disp true This word indicates that dispersivity is corrected

for flux-boundary end cells -diffusion_coefficient 3e-010

This word is to define the diffusion coefficents for the aqueous species

-thermal_diffusion 2 3e-010 This word is to calculate the diffusion part of

heat transport This is the default value -print_cells 4

This word is to identify that the results of the fourth cell is written into the output file

SELECTED_OUTPUT-file EBCT-054xls

Output file name-totals As(5) Arsenite Ca Mg Si

The composition in the output file -step true

The step is shown in the output file -ph true

22

289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

The pH is shown in the output file -pe true

The pe is shown in the output file

Table S54 Parameters of Weber-Morris intraparticle model for As(IIIV) adsorption kinetics on GTiO2 using synthetic water Initial As concentration = 800 μgL adsorbent dose = 02 gL pH = 82 in synthetic groundwater

　 External film diffusion

Intraparticle diffusion 　

　 aRex2 bkp1

cb1aRin

2 bkp2cb 2

As(V) 0984 0497 0022 0976 01331878

As(III) 0970 0583

-0088 0981 0218

1105

aRex is WM external proportion fitting parameter while Rin is WM intraparticle proportion fitting parameterbk The unit of kp is mggh05 cb The unit of b is mgg

23

333334335336337338

339

340

341342343

344345346347348349350351352353354355356357358359360361362363364365366

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

24

367368369370371372373

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Table S6 Comparison of the adsorption capacities of As(III) and As(V) on GTiO2 and the representative granular adsorbents in similar experimental condition from the literature

AdsorbentParticle size (mm) As

qm

(mgg)Dose (gL) Experimental conditions Literature

GTiO2 018~025 As(III) 980 1 0~500 mgL at pH 82 01 M NaCl This studyGTiO2 018~025 As(V) 358 1 0~500 mgL at pH 82 01 M NaCl This study

Granular TiO2 015~06 As(III) 392 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

Granular TiO2 015~06 As(V) 400 10~60 mgL As at pH 70 in simulated groundwater (100 mgL Ca 20 mgL Mg 50 mgL Na 200 mgL Cl-

)(Bang et al 2005)

E33 (FeOOH)0044~0075 As(V) 202 1 0~15 mgL at pH 70 in 001 M NaCl

(Kanematsu et al 2010)

E33 (FeOOH) 05~2 As(V) 180 1 0~50 mgL at pH 70 in 004 M NaClO4 (Jing et al 2012)

GFH 02~06 As(V) 650 10 0~2000 mgL at pH 8 in 01 M NaNO3

(Guan et al 2008)

GFH0009~0125 As(V) 23 025 0~07 μgL at pH 65 in DI water

(Banerjee et al 2008)

GFH0009~0125 As(V) 200 - 0~01 μgL at pH 70 in DI water

(Badruzzaman et al 2004)

GFH 050~065 As(V) 155 - 0~06 μgL at pH 70 in DI water (Saha et al 2005)

GHFO-NN 015~030 As(V) 330 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

GHFO-VL 015~030 As(V) 360 0001 0~225 mgL at pH 8 in DI water(Streat et al 2008)

25

374375

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Spherical schwertmannite 1~15 As(V) 36 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Cylindrical schwertmannite 1~25 As(V) 223 05

0~5 mgL at pH 70 in DI water(Dou et al 2013)

Irregular schwertmannite 10~16 As(V) 317 05 0~5 mgL at pH 70 in DI water (Dou et al 2013)Fe impregnated chitosan beads (MICB) 25 As(III) 117 1 0~05 mgL at pH 60 in DI water

(Wang et al 2014)

MICB 25 As(V) 191 1 0~1 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(III) 353 1 0~15 mgL at pH 60 in DI water(Wang et al 2014)

MICB 25 As(V) 357 1 0~16 mgL at pH 60 in DI water(Wang et al 2014)

Iron hydroxide granules (GIH) 3~4 As(V) 23 4

5~100 mgL at pH 70 in 10 mgL SO42-

14 mgL NH4Cl buffer solution (Daus et al 2004)

Activated Al2O3

0074~0149 As(V) 159 - 285~115 mgL at pH 52 in DI water

(Lin and Wu 2001)

Activated Al2O3

0074~0149 As(III) 35 - 079~490 mgL at pH 70 in DI water

(Lin and Wu 2001)

Activated Al2O3 ~20 As(V) 73 at pH 70 in 004 M NaClO4 (Jing et al 2012)

Activated natural siderite 05~10 As(V) 22 10 0 1~20 mgL at pH 7 in DI water(Zhao and Guo 2014)

Polyaluminum Granulate 1~3 As(III) 705 10 0~150 mgL at pH 51-56 in DI water(Mertens et al 2012)

Polyaluminum Granulate 1~3 As(V) 180 10 0~150 mgL at pH 51-56 in DI water (Mertens et al

26

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

2012)

Acidified laterite 1~12 As(III) 19 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Acidified laterite 1~12 As(V) 13 1 025~5 mgL at pH 7 in 100 mgL NaHCO3

(Glocheux et al 2013)

Laterite 1~12 As(III) 07 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Laterite 1~12 As(V) 04 1025~5 mgL at pH 7 in 100 mgL NaHCO3 (Glocheux et al

2013)

Treated laterite 03-05 As(V) 216 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite 03-05 As(III) 94 05 02-20 mgL at pH 70 in DI water(Maiti et al 2012)

Treated laterite03-05

As(V) 241 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Treated laterite03-05

As(III) 81 05 02-20 mgL at pH 70 in DI water(Maiti et al 2010a)

Acid-activated laterite-103-07

As(III) 0633 50 02-10 mgL at pH 66-70 in DI water(Maiti et al 2010b)

Acid-activated laterite-203-07

As(III) 087650 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(III) 063450 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-103-07

As(V) 059850 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

27

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Acid-activated laterite-203-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Acid-activated laterite-303-07

As(V) 07550 02-10 mgL at pH 66-70 in DI water (Maiti et al

2010b)

Treated laterite-1 03-06 As(V) 019 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-1 03-06 As(III) 018 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(V) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-2 03-06 As(III) 003 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(V) 007 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

Treated laterite-3 03-06 As(III) 004 20 0-20 mgL As at pH 72 in DI water(Maiti et al 2013b)

28

376377

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Charge distribution multi-site surface complexation (CD-MUSIC) modeling

The CD-MUSIC model with the 2-pK TPM adsorption option was used to

describe adsorption behaviors of As cations and anions on GTiO2 The CD-MUSIC

model incorporates the structural information of surfaces adsorbed species and the

double layers by distributing the charge of an adsorbate between the surface and

interfacial water (Hiemstra and Van Riemsdijk 2006) The CD-MUSIC model has

been successfully used in describing As adsorption in batch experiments (Pena et al

2006 Stachowicz et al 2008)

The adsorption constants of counter ions were set to -1 and the proton affinity

constant was set to the PZC of 58 for TiOH-13 (Luo et al 2010) The charge

distribution (CD) value shows the fraction (f) of the charge of the central As ion

attributed to the surface plane and the remaining part (1-f) was attributed to the other

ligands of the complex which were positioned toward the diffuse layer The CD value

for As surface complexes was obtained by fitting the experimental data The

calculation was performed using the chemical equilibrium program MINTEQ to

simulate the adsorption and the aqueous reactions

29

378379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Fig S32 Adsorption of silicate (a) calcium and magnesium (b) as a function of GTiO2

dose in groundwater 1 The solid lines represent the CD-MUSIC modeling result

30

400

401

402

403

404

405

406

407

408409410

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Table S75 Surface parameters and species used in the CD-MUSIC modeling in As(IIIV) removal under different GTiO2 dose in 01 M NaCl and groundwater 1 in Fig 2 and Fig S32

Species P0 P1

P2 TiOH Ti2O H AsO4 H3AsO3 Na Cl Ca Mg H4SiO4 CO3

2- logKTiOH-13 1TiOH2

+23 1 1 1 58TiOHNa+23 1 1 1 -1TiOH2ClO4

-13 1 -1 1 1 1 48Ti2O-23 1Ti2OH+13 1 1 1 58Ti2ONa+13 1 1 1 -1Ti2OHCl-23 1 -1 1 1 1 48Ti2O2AsO2

-53 -125 - 2 2 1 233a 260b 218c

Ti2O2AsOH-53 -11 11 2 1 40a 51b 66c

TiOCa067 02 08 1 -1 1 -37Ti2OHCa133 2 1 1 3TiOMg067 02 08 1 -1 1 -45Ti2OHMg133 2 1 1 3Ti2O2SiO2H2

-067 -14 14 2 1 28Ti2O2SiOOH-167 -12 02 2 -1 1 -38Ti2O2CO-067 068 068 2 2 1 22Surface SOH site density (mmolg) 6Surface area (m2g) 196Inner-sphere capacitance C1 (F m-2) 236Outer-sphere capacitance C2 (F m-2) 5

P0= exp(-FΨ0RT) P1

= exp(-FΨ1RT) P2 = exp(-FΨ2RT) F the Faraday constant (C mol-1) R the gas constant (J mol-1 K-1) T the absolute temperature (K)

Ψ0 Ψ1 Ψ2 the electrostatic potential (V) of 0- 1- and 2-plane respectively alogK for 01 M NaCl at 6 hours blogK for 01 M NaCl at 6 days clogK for groundwater 1 at 6 days

31

411412

413414415

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

32

416

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Table S86 The charge distribution (f) of central As atom to TiO2 surface in simulating the effect of cations and anions on As adsorption in 01 M NaCl at pH 82 in Fig 4　 BK Si HCO3 All

As(V) 022 017 012 016As(III) 040 020 058 056

Fig S43 Molar percentage () of As Ca Mg and Si on pristine and spent GTiO2

surfaces using EDX analysis The spent GTiO2 samples after field column experiments using groundwater 1 were identified with EBCT in min under x-axis The value of molar percentage of each element was averaged from randomly-chosen positions on the adsorbent surface

33

417

418

419420

421422

423

424

425

426

427

428

429

430

431

432

433434435436437438

439

440

441

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

As speciation analysis on spent GTiO2 using XANES

The spent GTiO2 samples were ground using a mortar and pestle and then the

ground powder was spread on Scotch tape The As k-edge XANES spectra were

collected at beamline 01C1 at the National Synchrotron Radiation Research Center

(NSRRC) Taiwan Spectra were acquired from -150 to 300 eV relative to the As K-

edge of 11867 eV at cryogenic temperature (77 K) using a cryostat to prevent the

oxidation of As(III) by X-rays The fluorescence signals were collected using a Lytle

detector Two to four scans were collected for each sample and averaged to improve

the signalnoise ratio The XANES analysis was performed with the Athena program

in the IFEFFIT computer package for linear combination fit (LCF) (Cui et al 2013

Ravel and Newville 2005)

34

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

PHREEQC modeling for As breakthrough curves in field column experiments

PHREEQC a geochemical model integrates macroscopic mass transfer

processes with a microscopic surface complexation model (Parkhurst and Appelo

2013) It can be applied in calculating solution complex speciation batch-reaction

and one-dimensional reactive-transport and inverse modeling In our study the one-

dimensional column was defined by four cells with the same length The cell length

was calculated to be one-fourth of the actual column length Time step (s) was

calculated according to equation Time step = Lcell(Q60A) where Lcell is the cell

length (cm) Q is flow velocity (mLmin) A is the effective column cross section area

(cm2) which is the column cross-sectional area multiplied by porosity (060) Total

shift number was calculated by dividing actual column operation time by the time

step The transport block was modeled by shifting the solute content from one cell to

the next one In each cell the aqueous solute was mixed and adsorbed on GTiO2

according to the equations In each cell the aqueous solute was mixed and assumed to

be in equilibration with GTiO2 The sequence of shifting mixing and equilibrium

adsorption was repeated for each cell until the total number of shifts was completed

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2

(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

For CD-MUSIC modeling in the PHREEQC take As(III) as an example to

explain the adsorption equation The As(III) adsorption is considered as a ligand

35

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

exchange reaction with the hydroxide group on GTiO2 (Surf_sOH-033) forming a

surface complex

H3ArseniteO3 + 2Surf_sOH-033 = (Surf_sO)2ArseniteOH-066 + 2H2O

log_k 1635

-cd_music -2 -1 0 066 3

H3ArseniteO3 is the master species in groundwater at pH 82

(Surf_sO)2ArseniteOH-066 is the common adsorbed bidentate binuclear complex

evidenced by our previous reports(Pena et al 2006 Yan et al 2015) LogK is the

adsorption equilibrium constants Generally every adsorption equation under certain

experimental condition has a unique value of logK In the adsorption equation ldquo-2rdquo

represents the lost two hydrogens by Surf_sOH-033 after it adsorbs H3ArseniteO3 ldquo-

1rdquo represents the change in charge on Arsenite at 1 plane due to gaining one hydrogen

(+1) and one oxygen (-2) ldquo0rdquo represents no change in charge at the 2 plane ldquo066rdquo is

the fraction of the central ion (As) on GTiO2 surface that is associated with plane 0

ldquo3rdquo is the charge on the central ion As(III) Following this equation and the relevant

parameters As(III) adsorbed on GTiO2

The diffusion coefficient Dw in PHREEQC accounting for retardation was

calculated as Dw=x2(36t) where x stands for the diffusion length (estimated as the

radius of the particle size) and t represents equilibrium time adopted from the batch

kinetic data (Mertens et al 2012)

To well simulate the As breakthrough curve in columns four As(V) adsorption

reactions (Table 1) were employed Adsorption site density of 60 mmolg was

36

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

adopted from our previous study (Pena et al 2006) and the total adsorption sites were

calculated to be 1844 sitesnm2 The final optimized ratio of strong to weak sites was

51 The surface complexation constants were reasonably adjusted according to the

batch modeling results as the column reaction may not reach equilibrium in local

micro-interfacial areas due to much less contact time between As and GTiO2 than that

in batch experiments (Appelo and Postma 1999) In the modeling a stepwise

procedure was adopted to achieve a viable result An estimation of two thirds of the

surface site density being accessible for EBCT=011 min with large flow rate of 265

mLmin was optimized which could probably result from unavailable adsorption sites

due to fast transport through the column

The adsorbed As on GTiO2 in the column was calculated by mass balance

between the influent and effluent

where Asadsorbed is the adsorbed As on GTiO2 (mgg) n is the number of collected

samples Cin and Cout are influent and effluent As concentrations (mgL) respectively

Vi is volume (L) of groundwater passed through the column between sample i-1 and i

and m is the mass of GTiO2 (g)

Even though the oxidation of adsorbed As(III) occurred during filtration as

analyzed from XANES and mass balance analysis (Fig 7) no As(III) oxidation

reactions were considered in PHREEQC modeling which would not impact the

breakthrough of As(IIIV)

37

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

38

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Fig S54 Effect of EBCT on Si Ca and Mg breakthrough curves using groundwater 1 The sample notation in panel c indicates field column experiments with EBCT in minutes

References

Appelo CAJ Postma D 1999 A consistent model for surface complexation on birnessite (MnO2) and its application to a column experiment Geochim Cosmochim Ac 63 (19-20) 3039-3048Badruzzaman M Westerhoff P Knappe DRU 2004 Intraparticle diffusion and adsorption of arsenate onto granular ferric hydroxide (GFH) Water Res 38 (18) 4002-4012Banerjee K Amy GL Prevost M Nour S Jekel M Gallagher PM Blumenschein CD 2008 Kinetic and thermodynamic aspects of adsorption of arsenic onto granular ferric hydroxide (GFH) Water Res 42 (13) 3371-3378Bang S Patel M Lippincott L Meng XG 2005 Removal of arsenic from groundwater by granular titanium dioxide adsorbent Chemosphere 60 (3) 389-397Bang S Pena ME Patel M Lippincott L Meng X Kim KW 2011 Removal of arsenate from water by adsorbents a comparative case study Environ Geochem Hlth 33 133-141Cui JL Shi JB Jiang GB Jing CY 2013 Arsenic levels and speciation from ingestion exposures to biomarkers in Shanxi China Implications for human health Environ Sci Technol 47 (10) 5419-5424Daus B Wennrich R Weiss H 2004 Sorption materials for arsenic removal from water A comparative study Water Res 38 (12) 2948-2954Dou XM Mohan D Pittman CU 2013 Arsenate adsorption on three types of granular schwertmannite Water Res 47 (9) 2938-2948Dou XM Zhang Y Yang M Pei YS Huang X Takayama T Kato S 2006 Occurrence of arsenic in groundwater in the suburbs of Beijing and its removal using an iron-cerium bimetal oxide adsorbent Water Qual Res J Can 41 (2) 140-146Gibbons MK Gagnon GA 2010 Adsorption of arsenic from a Nova Scotia groundwater onto water treatment residual solids Water Res 44 (19) 5740-5749Glocheux Y Mendez Pasarin M Albadarin AB Allen SJ Walker GM 2013 Removal of arsenic from groundwater by adsorption onto an acidified laterite by-product Chem Eng J 228 565-574Guan XH Wang JM Chusuei CC 2008 Removal of arsenic from water using granular ferric hydroxide Macroscopic and microscopic studies J Hazard Mater 156 (1-3) 178-185Gupta K Maity A Ghosh UC 2010 Manganese associated nanoparticles agglomerate of iron(III) oxide Synthesis characterization and arsenic(III) sorption behavior with mechanism J Hazard Mater 184 (1-3) 832-842Hao JM Han MJ Wang C Meng XG 2009 Enhanced removal of arsenite from water by a mesoporous hybrid material - Thiol-functionalized silica coated activated alumina Micropor Mesopor

39

552

553

554555556557

558

559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

Mat 124 (1-3) 1-7Hiemstra T Van Riemsdijk WH 2006 On the relationship between charge distribution surface hydration and the structure of the interface of metal hydroxides J Colloid Interf Sci 301 (1) 1-18Hristovski K Baumgardner A Westerhoff P 2007 Selecting metal oxide nanomaterials for arsenic removal in fixed bed columns From nanopowders to aggregated nanoparticle media J Hazard Mater 147 (1-2) 265-274Jang M Chen WF Cannon FS 2008 Preloading hydrous ferric oxide into granular activated carbon for arsenic removal Environ Sci Technol 42 (9) 3369-3374Jing CY Cui JL Huang YY Li AG 2012 Fabrication characterization and application of a composite adsorbent for simultaneous removal of arsenic and fluoride ACS Appl Mater Interfaces 4 (2) 714-720Kanematsu M Young TM Fukushi K Green PG Darby JL 2010 Extended triple layer modeling of arsenate and phosphate adsorption on a goethite-based granular porous adsorbent Environ Sci Technol 44 (9) 3388-3394Kong S Wang Y Zhan H Yuan S Liu M Zhou C 2013 Arsenite and Arsenate Removal from Contaminated Groundwater by Nanoscale Iron-Manganese Binary Oxides Column Studies Environmental Engineering Science 30 (11) 689-696Li Y Liu JR Jia SY Guo JW Zhuo J Na P 2012 TiO2 pillared montmorillonite as a photoactive adsorbent of arsenic under UV irradiation Chem Eng J 191 66-74Lin TF Wu JK 2001 Adsorption of arsenite and arsenate within activated alumina grains Equilibrium and kinetics Water Res 35 (8) 2049-2057Luo T Cui JL Hu S Huang YY Jing CY 2010 Arsenic removal and recovery from copper smelting wastewater using TiO2 Environ Sci Technol 44 (23) 9094-9098Maiti A Basu JK De S 2010a Development of a Treated Laterite for Arsenic Adsorption Effects of Treatment Parameters Industrial amp Engineering Chemistry Research 49 (10) 4873-4886Maiti A Basu JK De S 2010b Removal of Arsenic from Synthetic and Natural Groundwater Using Acid-Activated Laterite Environmental Progress amp Sustainable Energy 29 (4) 457-470Maiti A Basu JK De S 2012 Experimental and kinetic modeling of As(V) and As(III) adsorption on treated laterite using synthetic and contaminated groundwater Effects of phosphate silicate and carbonate ions Chem Eng J 191 1-12Maiti A Thakur BK Basu JK De S 2013a Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 1176-1186Maiti A Thakur BK Basu JK De S 2013b Comparison of treated laterite as arsenic adsorbent from different locations and performance of best filter under field conditions J Hazard Mater 262 (0) 1176-1186Maji SK Pal A Pal T 2008 Arsenic removal from real-life groundwater by adsorption on laterite soil J Hazard Mater 151 (2-3) 811-820Maji SK Kao YH Wang CJ Lu GS Wu JJ Liu CW 2012 Fixed bed adsorption of As(III) on iron-oxide-coated natural rock (IOCNR) and application to real arsenic-bearing groundwater Chem Eng J 203 285-293Mertens J Rose J Kagi R Chaurand P Plotze M Wehrli B Furrer G 2012 Adsorption of arsenic on polyaluminum granulate Environ Sci Technol 46 (13) 7310-7317Nguyen TV Vigneswaran S Ngo HH Pokhre D Viraraghavan T 2006 Iron-coated sponge as

40

592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672

effective media to remove arsenic from drinking water Water Qual Res J Can 41 (2) 164-170Parkhurst DL Appelo CAJ 2013 Description of input and examples for PHREEQC version 3-A computer program for speciation batch-reaction one-dimensional transport and inverse geochemical calculations US Geological SurveyPena M Meng XG Korfiatis GP Jing CY 2006 Adsorption mechanism of arsenic on nanocrystalline titanium dioxide Environ Sci Technol 40 (4) 1257-1262Ravel B Newville M 2005 ATHENA ARTEMIS HEPHAESTUS Data analysis for X-ray absorption spectroscopy using IFEFFIT J Synchrotron Radiat 12 537-541Saha B Bains R Greenwood F 2005 Physicochemical characterization of granular ferric hydroxide (GFH) for arsenic(V) sorption from water Separation Science and Technology 40 (14) 2909-2932Shan H Ma T Wang Y Zhao J Han H Deng Y He X Dong Y 2013 A cost-effective system for in-situ geological arsenic adsorption from groundwater J Contam Hydrol 154 1-9Stachowicz M Hiemstra T van Riemsdijk WH 2008 Multi-competitive interaction of As(III) and As(V) oxyanions with Ca2+ Mg2+ PO4

3- and CO32- ions on goethite J Colloid Interf Sci 320 (2) 400-

414Streat M Hellgardt K Newton NLR 2008 Hydrous ferric oxide as an adsorbent in water treatment - Part 3 Batch and mini-column adsorption of arsenic phosphorus fluorine and cadmium ions Process Safety and Environmental Protection 86 (B1) 21-30Thirunavukkarasu OS Viraraghavan T Subramanian KS 2003 Arsenic removal from drinking water using iron oxide-coated sand Water Air Soil Poll 142 (1-4) 95-111USEPA 2008 Assessing arsenic removal by metal (hydr)oxide adsorptive media using rapid small scale column tests Govenrnment Printing Office Washington DCWang J Xu W Chen L Huang X Liu J 2014 Preparation and evaluation of magnetic nanoparticles impregnated chitosan beads for arsenic removal from water Chem Eng J 251 25-34Westerhoff P Highfield D Badruzzaman M Yoon Y 2005 Rapid small-scale column tests for arsenate removal in iron oxide packed bed columns J Environ Eng-Asce 131 (2) 262-271Westerhoff P De Haan M Martindale A Badruzzaman M 2006 Arsenic adsorptive media technology selection strategies Water Qual Res J Can 41 (2) 171-184Yan L Huang Y Cui J Jing C 2015 Simultaneous As(III) and Cd removal from copper smelting wastewater using granular TiO2 columns Water Res 68 (0) 572-579Zhao K Guo HM 2014 Behavior and mechanism of arsenate adsorption on activated natural siderite Evidences from FTIR and XANES analysis Environ Sci Pollut Res 21 (3) 1944-1953Zhao K Guo H Zhou X 2014 Adsorption and heterogeneous oxidation of arsenite on modified granular natural siderite Characterization and behaviors Appl Geochem 48 184-192

41

636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672