A

raoaAavtaaa©

K

1

fmfctI

gf

0d

Chemical Engineering and Processing 46 (2007) 1030–1039

Evaluation of iron oxide and aluminum oxide aspotential arsenic(V) adsorbents

Youngran Jeong a,b, Maohong Fan a,b,∗, Shilpi Singh c, Chia-Line Chuang d,Basudeb Saha e, J. Hans van Leeuwen b

a Center for Sustainable Environmental Technologies, Iowa State University, Ames, IA 50011, USAb Department of Civil, Construction and Environmental Engineering, Iowa State University, Ames, IA 50011, USA

c Sanborn, Head & Associates, Inc., Concord, NH 03301, USAd Department of Research and Development, Allied Biotech Corp., Tao-Yuan, Taiwan

e Department of Chemistry, Iowa State University, USA

Received 18 October 2006; received in revised form 4 May 2007; accepted 7 May 2007Available online 17 May 2007

bstract

Iron (Fe2O3) and aluminum oxide (Al2O3) were found to be good and inexpensive adsorbents for As(V) removal in drinking water despite theirelatively small surface area. The experimental results for this study suggest that by careful selection of the relative concentration of arsenic, pH,nd dosages of Fe2O3 and Al2O3, As(V) removal efficiency as high as 99% can be achieved. At lower pH (<7), and also depending on the dosagesf Fe2O3 and Al2O3 and the initial concentration of As(V), over 95% of As(V) adsorption was observed within a contact time of 20–60 min. Thedsorption of As(V) on Fe2O3 and Al2O3, like that on other nonporous adsorbents, is mainly controlled by the surface area. The adsorption ofs(V) on Fe2O3 and Al2O3 was found to follow the Langmuir isotherm between the pH values of 5 and 9. The maximum As(V) uptake values

t pH 6 – the optimal pH value for adsorption – using Fe2O3 and Al2O3, were calculated as 0.66 mg/g and 0.17 mg/g, respectively. No significantariation in the uptake of As(V) on Fe2O3 as compared with Al2O3 was observed at different pH values. The initial sorption rate of Fe2O3 is higherhan that of Al O . All these factors make Fe O a better adsorbent than Al O . Fe O is a useful and effective adsorbent for POE (pint of entry)

2 3 2 3 2 3 2 3

nd POU (point of use) water treatment systems, such as small-scale commercial or individual home water treatment systems. Even though thedsorption capacities of Fe2O3 and Al2O3 for As(V) are quite low compared with those of other absorbents, their low cost makes them usefuldsorbents. They may be very useful in arsenic removal from water in endemic areas such as China, India, and Bangladesh.

2007 Elsevier B.V. All rights reserved.

saaina

eywords: Arsenic removal; Water treatment; Iron oxide; Aluminum oxide

. Introduction

Arsenic is one of the most common natural contaminantsound in water that has adverse effects on human health. Arsenicainly originates from arsenic-containing rocks and soil and

rom some anthropogenic sources including mining, glass pro-

essing, insecticides, pesticides, and landfill leaching [1,2]. It isransported into natural waters through erosion and dissolution.t occurs in natural waters in both inorganic and organic forms,

∗ Corresponding author at: Center for Sustainable Environmental Technolo-ies, Iowa State University, Ames, IA 50011, USA. Tel.: +1 515 294 3951;ax: +1 515 294 3091.

E-mail address: mfan3@mail.gatech.edu (M. Fan).

Haivc[t

[

255-2701/$ – see front matter © 2007 Elsevier B.V. All rights reserved.oi:10.1016/j.cep.2007.05.004

uch as monomethyl arsenic acid (MMAA), dimethyl arseniccid (DMAA), and arseno-sugars [3]. The inorganic form ofrsenic is most toxic to humans. Inorganic arsenic usually occursn two valence states, arsenite [As(III)] and arsenate [As(V)]. Inatural waters, As(III) species primarily consist of arseniouscid (H3AsO3), while As(V) species consists of H2AsO3

− andAsO3

2− [4]. The previous research has shown that As(V) inerobic surface water has better removal efficiency than As(III)n anaerobic groundwater [5]. Besides, As(III) is easily con-erted to As(V) by oxidizing agents such as oxygen, ozone, freehlorine, hypochlorite, permanganate, and hydrogen peroxide

6–8]. Thus, it is convenient to consider only As(V) compoundso be removed in drinking water treatment.

Long-term exposure to inorganic arsenic such as arseniteAs(III)] and arsenate [As(V)] in drinking water leads to adverse

ng an

hnetiU

a[tErfbsaesBtd

ot[asriiraaocHrni

asfgpfAateaFAw(aa

rpoidaa

ttefotas

oetoak

2

2

[utptmdsk

wt(

h

T(mt

2

Y. Jeong et al. / Chemical Engineeri

ealth effects such as cancer of the bladder, lungs, skin, kidney,asal passages, liver, and prostate [9–11]. These adverse healthffects of arsenic have mainly been reported in third world coun-ries such as Bangladesh, India, Chile, and China, but a fewncidences have been found even in affluent nations like thenited States [12–15].Based on the above-discussed adverse health effects of

rsenic, changes in WHO’s arsenic standard for drinking water16], and the need to protect citizens against the effects of long-erm, chronic exposure to arsenic in drinking water, the USnvironmental Protection Agency (US EPA) has revised the cur-

ent maximum contaminants level (MCL) of 0.05 mg/L (50 ppb)or arsenic in drinking water to 0.01 mg/L (10 ppb), which wille effective from 23 January 2006 [17]. In addition, the USEPAuggests the criteria of the best available technology (BAT) forrsenic removal, that is, technology that provides high removalfficiency, has a history of full-scale operation and a reasonableervice life, and is cost effective. The new MCL standard andAT for arsenic have led researchers to find new and improved

reatment techniques for more effective removal of arsenic fromrinking water [18].

Adsorption is a separation or purification process in whichrganic or inorganic compounds are adsorbed, for removal fromhe solution, onto porous solid media with a large surface area19]. Adsorption has a comparatively low cost and easily sep-rates a small amount of toxic elements from large volumes ofolutions. These benefits of adsorption have motivated severalesearchers to use adsorption for arsenic removal from drink-ng water. Some of the common adsorbents used for the processnclude activated alumina, manganese green sand, granular fer-ic hydroxide, soil, and mud [20–24]. Adsorption by activatedlumina (AA) was approved recently as one of the best avail-ble technologies (BAT) for arsenic removal. It has an advantagever other adsorbents in that AA used during the removal pro-ess is nonhazardous and could be safely disposed of in landfills.owever, AA is very pH sensitive and has a low regeneration

ate of 50–70% [18]. It is necessary to consider development ofew adsorbent, one that is more effective in removing arsenaten drinking water.

Several researchers have found that Fe- and Al-baseddsorbents adsorb arsenic compounds in drinking water to aignificant extent. Of these, amorphous ferric hydroxide [25],erric oxide [26], hydrous ferric oxide [27], ferrihydrite [28],ranular ferric hydroxide [22], goethite and gibbsite [29], amor-hous aluminum oxide [30], activated alumina [31,32] are mostrequently described. However, previous studies of Fe- andl-based adsorbents have focused on adding chemicals for

ugmenting adsorption capacity (e.g., coating, heat, or acidreatment) and on reusing adsorbents to reduce cost after regen-ration (usually with sodium hydroxide). Maeda et al. [33]nd Vaughan and Reed [34] used Fe-impregnated coral ande-impregnated activated carbon to enhance arsenic removal.ltundogan et al. [24] demonstrated that the acid treatment

ith 1 M HCl solution to red mud, which removes sodalite

Na2O·Al2O3·1.68SiO2·1.73H2O), increases the adsorptionbility of red mud in industrial wastes treatment. A desorptiongent (e.g., sodium hydroxide solution) is commonly used to

aa

d Processing 46 (2007) 1030–1039 1031

ecover expensive adsorbent media such as mesoporous aluminarepared by the templating method [35]. However, pre-treatmentf the adsorbent or use of a desorption agent increases the operat-ng cost. In addition, waste solution containing HCl and NaOHiscarded from pretreatment for boosting adsorption capacitylso result in pollution. The development of new, inexpensive,nd easy-to-manage adsorbents would be quite worthwhile.

Arsenate adsorption using Fe and Al oxides without pre-reatment or regeneration processes may offer a useful newechnology, particularly in point of use (POU) or point ofntry (POE) treatment units (such as household drinking wateraucets). It would require no hazardous desorption agents, andffers easy maintenance of the adsorption system [36]. Addi-ionally, iron and alum oxides are inexpensive chemicals andre readily available at water treatment plants, where they areometimes used in other unit processes [37,38].

In this research, we studied the adsorption of arsenate [As(V)]nto Fe2O3 and Al2O3. The objectives of this study were tovaluate the use of Fe2O3 and Al2O3 as possible adsorbents forhe removal of arsenate from drinking water through a seriesf experiments conducted by changing various parameters suchs initial concentration, dosage of adsorbents, and pH that arenown as critical parameters in other researches.

. Theoretical study

.1. Adsorption kinetic equation

The pseudo-second-order rate equation was developed by Ho39] to describe the adsorption systems of divalent metal ionssing sphagnum moss peat. This equation derives the adsorp-ion capacity of solid from the solution concentration. Theseudo-second-order rate equation has recently been appliedo discussions of various reactions such as the adsorption ofetal ions or organic substances from liquid solutions and the

esign of multistage or batch adsorption facilities. The pseudo-econd-order equation was used to explain the arsenic adsorptioninetics. The pseudo-second-order equation is as follows:

t

qt

= 1

k2q2e

+ 1

qet (1)

here k2 is the rate constant of adsorption (g/(mg min)) and h ishe initial sorption rate (mg/(g min)). As time approaches zerot → 0), h can be defined as

= k2q2e (2)

he initial sorption rate (h), the equilibrium sorption capacityqe), and the pseudo-second-order rate constant (k2) can be deter-ined experimentally from the slope and intercept of the plot of

/q versus t.

.2. Adsorption isotherm

Several adsorption isotherms including Brunauer, Emmet,nd Teller (BET), Dubinin and Raduskevich (D–R), and Polanyire available to describe adsorption phenomena, but none of

1 ng and Processing 46 (2007) 1030–1039

tLs

q

wvasaf

aiaato

3

3

BaIstar

2

2

TIaoccpaa

a(XcasmSmb

Fgw

0toaASfapwas found to be 80.49 and 75.69 A, respectively, which meansthat these Fe2O3 and Al2O3 particles generally have mesopores(20–500 A) and not micropores (<20 A). Total pore volume was1.02 × 10−2 cm3/g for Fe2O3 and 1.02 × 10−3 cm3/g for Al2O3.

Table 1Properties of adsorbents

Fe2O3 Al2O3

Manufacturer Bailey-PVS oxides (USA) Praxair (USA)Purity Fe2O3 (99.1%) Al2O3 (99.5%)

032 Y. Jeong et al. / Chemical Engineeri

hem are based on or derived from the Freundlich and theangmuir [40], which are more common isotherm models foringle-solute adsorption.

The Langmuir equation can be linearized as below:

e = qmaxbCe

1 + bCeor

1

qe= 1

qmaxbCe+ 1

qmax(3)

here b and qmax are constants. qmax represents the maximumalue of qe that can be achieved. b is related to the energy ofdsorption and increases with the increase in adsorption bondtrength. The basic assumption of the Langmuir isotherm is thatdsorption of solutes happens at specific homogeneous sites andorms a monolayer.

Some important adsorbent characteristics affecting isothermsre surface area, pore size distribution, and surface chem-stry. For nonporous adsorbents, like ferrihydrite, the maximummount of adsorption is proportional to the amount of surfacerea within pores that is accessible to the adsorbate. However,he surface area of porous adsorbents is not the chief influencen adsorption capacity [41,42].

. Experimental sections

.1. Adsorbents

Ferric oxide (Fe2O3-PVS; Physical Vapor Synthesis) byailey-PVS and aluminum oxide (Al2O3-ALO101) by Praxairre the two metal oxides used in our experiments as adsorbents.ron oxide is produced from the decomposition of iron chlorideolution within the spray roasting reactor. This reaction requireshe presence of water vapor and oxygen at a temperature betweenpproximately 600 and 1600 ◦F (589 and 1144 K). The basiceactions are as follows:

FeCl2 + 2H2O + (1/2)O2 → Fe2O3 + 4HCl (R1)

FeCl3 + 2H2O → Fe2O3 + 6HCl (R2)

his also makes the recovery of hydrochloric acid possible.n order to elucidate the possibility of using of iron oxides an appropriate adsorbent for removal of As(V), aluminumxide, having characteristics similar to those of iron oxide washosen as a comparable adsorbent. Aluminum oxide (Al2O3)omes from Praxair (Wisconsin, USA), a manufacturing com-any that produces thermal spray powders. This Al2O3 is coatedt high temperature to resist to abrasion, erosion, alkali, andcids.

The colors of iron and aluminum oxide are reddish brownnd grayish white, respectively. As shown in Fig. 1, the SEMscanning electron microscope) images obtained with a PhilipsL-30 (25 kV LaB6 filament) show that Fe2O3 particles exist as

lustered and aggregated shapes while Al2O3 particles occurss acicular forms with a smooth surface. The actual particleizes of Fe2O3 cannot be determined, but those of Al2O3 are

uch larger. The length and width of the Al2O3 particle in theEM figure seem to be 10–50 and 5–20 �m, respectively. Thisatches 5–45 �m of average particle size of Al2O3 provided

y the manufacture. The average particle size of Fe2O3 was

PAPS

ig. 1. (a) SEM figure of Fe2O3 at 1100× magnification, clustered and aggre-ated shapes; (b) SEM figure of Al2O3 at 1100× magnification, acicular formsith a smooth surface.

.7 �m according to the company and was 7–70 times smallerhan that of Al2O3. As provided by manufacturers, the ironxide consists of 99.1% Fe2O3 and a small portion of other met-ls such as Cu and Zn, while aluminum oxide contains 99.5%l2O3 and a small portion of other elements such as Ca andi. The effective pore size, total pore volume, and specific sur-ace area of the two adsorbents were measured using BET gasdsorption methods (ASAP2010, Micromeritics, USA) and arerovided in Table 1. The effective pore size of Fe2O3 and Al2O3

article size (�m) 0.7 5–45verage pore size (A) 80.49 75.69ore volume (cm3/g) 1.02 × 10−2 1.02 × 10−3

pecific surface area (m2/g) 5.05 0.55

ng an

F5

bf8[3aflhsatesetrbp

3

7wr((pTmT((aac[AasmfT(ci

3

t(t

ti4a

3

uMdstatotoe

4

4i

somitosits

AtAooAT20tbdwaoa

Y. Jeong et al. / Chemical Engineeri

inally, the specific surface area of Fe2O3 was found to be ca..05 m2/g, and that of Al2O3 was ca. 0.55 m2/g.

The specific surface areas of some of the adsorbents haveeen reported by several investigators; for example, 200 m2/gor ferrihydrite [43], 290 m2/g for amorphorous ferric oxide [44],40 m2/g for iron oxide impregnated activated carbon (FeAC)34], and from 15 m2/g for aluminum-loaded Shirasu-zeolite to00 m2/g for highly porous activated alumina [35,45]. Fe2O3nd Al2O3 have very small surface areas compared with the sur-ace areas of the above absorbents. However, Fe2O3 has a mucharger surface area than Al2O3. In general, metal oxides or metalydroxides do not dissolve in neutral solutions but are easilyoluble in acidic and strongly basic solutions because of theirmphoteric characteristics. Mesoporous alumina was reportedo dissolve in both acids (pH < 3) and bases (pH > 8) [35]. How-ver, Fe2O3 and Al2O3 are almost insoluble in acid and alkalineolutions, and once filtered, they do not have significant adverseffects on water quality, such as imparting color. Because ofhese properties, Fe2O3 and Al2O3 were found to be stable andeliable adsorbents for drinking water treatment, and they haveeen used in our As(V) adsorption experiments without furtherurification.

.2. Batch experiment

Batch experiments were conducted in a jar tester (PB-00TM, Phipps & Bird, USA) to study the removal of As(V)ith Fe2O3 and Al2O3 in drinking water. All chemicals were

eagent grade from Fisher Chemicals, while sodium arsenateNa2HAsO4·7H2O) was from Matheson Coleman and BellNorwood, Ohio, USA). Arsenate solutions were freshlyrepared by dissolving sodium arsenate in deionized water.he pH values were measured using a Corning model 320 pHeter, calibrated using commercial pH 4.01 and 7.0 buffers.he pH values of the test solutions were adjusted to 5–9±0.1 pH unit) using either diluted 0.1 M hydrochloric acidHCl) or 0.1 M sodium hydroxide (NaOH) solutions beforedsorption. As(V) concentrations were controlled at 200, 400,nd 600 �g/L (2.67–8 × 10−6 M) considered reasonable con-entrations encountered in the United States and Bangladesh46]. Approximately 0.05–1 g/L of Fe2O3 or 0.5–6 g/L ofl2O3 was added to a solution prepared at a predetermined

rsenate concentration using deionized water, followed bytirring at 23 ± 0.5 ◦C for 1 and 2 h, respectively. The pH waseasured, and for analysis, supernatant was collected directly

rom the jar after reaction using a 10 mL disposable syringe.he samples were filtered through 0.45 �m syringe filters

Millipore Millex) and analyzed for arsenic. The adsorptionapacities were calculated from the difference between thenitial and the equilibrium concentrations.

.3. Kinetics

Kinetic studies were also conducted at different intervals ofime and concentrations in a 500 mL jacketed reactor vesselChemglass, New Jersey, USA) equipped with a constant-emperature circulating bath (Cole-Parmer, USA) to determine

pttt

d Processing 46 (2007) 1030–1039 1033

he rate of arsenate removal by Fe2O3 and Al2O3. Kinetic stud-es were done at the three different temperatures of 5, 25, and5 ± 0.5 ◦C. The same procedure was used for analysis as indsorption experiments.

.4. Arsenate analysis

Arsenate concentrations in all the samples were analyzedsing Inductively Coupled Plasma–Mass Spectrometry (ICP-S, 4500 Series, HP) following Standard Methods [47]. The

etection limit of ICP-MS was 0.1 �g/L for arsenic. The mea-urements were accepted as reasonable data in cases of lesshan 5 or 10% relative standard deviation (R.S.D.) when thersenic concentration in the samples was greater than or lesshan 50 �g/L, respectively. In order to amplify the consistencyf results, the experiments were performed in triplicate andhe mean values considered. No detectable As(V) adsorbedn the walls of the jar was ascertained through the blankxperiments.

. Results and discussion

.1. The effects of contact time with adsorbent dosages ornitial As(V) concentrations

The adsorption rate is influenced by many factors, includingolubility and molecular size of the adsorbate, characteristicsf the adsorbent, and agitation. In addition, temperature com-only affects the adsorption rate, solute hydrolysis status, and

onization constant concurrently [19]. The rates of As(V) adsorp-ion on Fe2O3 were not dependent on temperature, while thosen Al2O3 were slightly dependent on temperature (figure nothown). The effect of temperature may not have a significantmpact on As(V) adsorption using Al2O3 because of the rela-ively stable variation of water temperature in water treatmentystems.

Figs. 2 and 3 show the rates of As(V) adsorption on Fe2O3 andl2O3 at different dosages of adsorbents and different As(V) ini-

ial concentrations. The rates of As(V) adsorption on Fe2O3 andl2O3 were found to be significantly time dependent. The ratef As(V) adsorption was found to be higher with high dosagesf Fe2O3 and Al2O3 to As(V) and to be higher with lower initials(V) concentrations to the same dosages of Fe2O3 and Al2O3.o achieve the less than 10 �g/L of the new arsenic standard from00 �g/L of initial As(V) concentration at pH 6, the dosages are.5 g/L of Fe2O3 or 3 g/L of Al2O3. That means that to achievehe same equilibrium As(V) concentration for both the adsor-ents, Al2O3 was required at a dosage of more than six times theosage of Fe2O3. Additionally, the As(V) adsorption onto Fe2O3as rapid in the first 20 min and then slowed down considerably

s the reaction approached equilibrium. However, the majorityf adsorption onto Al2O3 was achieved in the first 60 min. Piercend Moore found that the 99% of As(V) adsorption onto amor-

hous Fe hydroxide was achieved after 4 h of stirring at each ofhree final pH values (4.0, 8.0, and 9.9). They also observed thathe removal rate was somewhat faster for higher As(V) concen-rations (1 mg/L) than for lower As(V) concentrations (50 �g/L).

1034 Y. Jeong et al. / Chemical Engineering and Processing 46 (2007) 1030–1039

Fig. 2. Time courses of As(V) adsorption for different dosage of (a) Fe O and(s

Tawvhu[aAFaFt

4

eo

Fd8

oco

fprsts(∼5.5 m2/g), is more dynamic than Al2O3, with a larger averageparticle size (5–45 �m) and smaller surface area (∼0.5 m2/g), inadsorption of As(V).

Table 2Pseudo-second-order rate constants

2 3

b) Al2O3. Initial As(V) concentration, 200 �g/L; sample volume, 0.5 L; stirringpeed, 83 ± 5 rpm; pH 5 ± 0.1; 25 ◦C; error bars = standard error.

hey suggested that an adequate time for As(V) equilibrium onmorphous Fe hydroxide was 24 h [25]. Iron oxide-coated sandas observed to adsorb As(V) completely in 50 min [48]. Acti-ated alumina grains (100 mesh, Macherey-Nagel, Germany)ave been reported to typically have low adsorption rates, andp to 2 days are required to reach half of the equilibrium value42]. Mesoporous alumina prepared by a templating method asn adsorbent may take approximately 5 h to meet the equilibriums(V) concentration [35]. Compared with those of amorphouse hydroxide, activated alumina, and mesoporous alumina, thedsorption rates of Al2O3 and Fe2O3 were found to be higher.urthermore, the adsorption rates of Fe2O3 were faster than

hose of Al2O3.

.2. Adsorption kinetics

The rate of sorption is one of the most important factors invaluating the efficiency of sorption and in determining the sizef water treatment unit processes. In order to estimate the rates

FA

ig. 3. Time courses of As(V) adsorption onto (a) Fe2O3 and (b) Al2O3 forifferent As(V) initial concentrations. Sample volume, 0.5 L; stirring speed,3 ± 5 rpm; pH 6 ± 0.1; 25 ◦C: (a) dosage, 0.8 g/L, (b) dosage, 3.0 g/L.

f adsorption and to identify the behavior of the adsorptive, weonducted experiments related to the kinetics of As(V) removaln Fe2O3 and Al2O3.

The adsorption rates of As(V) using Fe2O3 and Al2O3 areound to fit this pseudo-second-order kinetics equation well. Theseudo-second-order rate constant (k2) and the initial sorptionate (h) are estimated in Fig. 4 and are listed in Table 2. It ishown that the h value for Fe2O3 (0.26 mg/(g min)) is higher thanhat for Al2O3 (0.04 mg/(g min)), which means Fe2O3, having amaller average particle size (0.7 �m) and a larger surface area

h (mg/(g min)) k2 (g/(mg min))

e2O3 0.26 1.68l2O3 0.04 2.64

Y. Jeong et al. / Chemical Engineering and Processing 46 (2007) 1030–1039 1035

FIA

4

ooaatw

AauldiuteftaataaFrTFaoounta

Fig. 5. Arsenate adsorption onto Fe2O3 as a function of pH at (a) differentdosages and (b) As(V) initial concentrations. Sample volume, 1 L; stirring speed,130 ± 5 rpm; 25 ◦C; stirring time, 1 h. (a) Dosage of Fe O , 0.05–1.0 g/L; As(V)id

i[rdsicomtFFF

ig. 4. Pseudo-second-order sorption kinetics of As(V) onto Fe2O3 and Al2O3.nitial As(V) concentration, 200 �g/L; pH 5 ± 0.1; 25 ◦C; dosage of Fe2O3 andl2O3, 0.5 g/L, respectively.

.3. The effect of pH

The uptake of As(V) per mass unit of adsorbent as a functionf pH was studied to find the optimum pH value for adsorptionf As(V) using iron or aluminum oxides. The stirring time forrsenate adsorption was kept at 1 h for iron oxide and 2 h forluminum oxide. Fe2O3 and Al2O3 were able to remove morehan 95% of As(V) from 200 �g/L of arsenate-contaminatedater and met the new 10 �g/L drinking water standard.Figs. 5 and 6 show the arsenic(V) uptakes of Fe2O3 and

l2O3 as a function of pH for different dosages of adsorbentsnd different initial As(V) concentrations. The maximum As(V)ptakes of both Fe2O3 and Al2O3 were achieved at a pH 6. At pHess than 8, increasing As(V) uptake rates were observed withecreasing dosages of Fe2O3 and Al2O3, and increasing As(V)nitial concentrations; however, for pH greater than 8, As(V)ptakes were little affected by adsorbent dosages or As(V) ini-ial concentrations. Besides, regardless of varying pH values,qual As(V) uptake rates for Fe2O3 and Al2O3 were observedor 1 g/L of Fe2O3 and 4 g/L of Al2O3. Thus, it can be concludedhat the As(V) uptake rates of Fe2O3 and Al2O3 are significantlyffected by pH in conditions of comparatively lower dosages ofdsorbent and higher As(V) initial concentrations, which fur-her means comparatively lower As(V) concentrations or higherdsorbent dosages are better for As(V) adsorption on Fe2O3nd Al2O3 at varying pH conditions. The uptake of As(V) one2O3 does not change significantly from pH 5 to 7 but decreasesapidly at pH 8, while that of Al2O3 decreases rapidly at pH 6.his result with Fe2O3 is comparable to that with amorphouse(OH)3, where As(V) adsorption was not affected significantlyt lower pH (<7) but decreased rapidly at pH > 7 [25,49]. Similarbservations were made in another study in which FeAC (ironxide impregnated activated carbon) was used. The adsorption

ptake rate of Fe oxides on the surface of FeAC decreases sig-ificantly at high pH values [34]. According to some reports,he decrease in adsorption rate occurs because the surface of thedsorbent becomes negatively charged and columbic repulsion

uiFH

2 3

nitial concentration, 200 �g/L. (b) Initial As(V) concentration, 200–600 �g/L;osage of Fe2O3, 0.5 g/L.

s enhanced [45,48,50]. According to Stumm and Sulzberger51], the decrease is because of ligand exchange, where anionseact with surface hydroxides on the adsorbents. Others haveescribed these observation as the formation of an inner-sphereurface complex on adsorbents such as goethite (�-FeOOH) orron oxide-coated sand (IOCS) [52,53]. Thus, it can be con-luded that the form of H2AsO4

− may be primarily adsorbedn Fe2O3 at a lower pH. However, with increasing pH, theonovalent arsenate anion does not get adsorbed on Fe2O3 due

o a negative surface charge and coulombic repulsion [54,55].inally, the arsenate anions interact with the surface sites ofe2O3 that are occupied by hydrogen ions and transform toe(H2AsO4)0, Fe(HAsO4)−, and Fe(AsO4)2−.

The As(V) uptake of Al2O3 decreases sharply at pH val-es differing from 6 when compared with the Fe O as shown

2 3n Fig. 6. Thus, it is concluded that Al2O3 is more likely thane2O3 to absorb H2AsO4

− at lower pH but may prefer OH− to2AsO4

− at higher pH (>6). According to some researchers, Fe

1036 Y. Jeong et al. / Chemical Engineering and Processing 46 (2007) 1030–1039

Fig. 6. Arsenate adsorption onto Al2O3 as a function of pH at (a) differentdosage and (b) As(V) initial concentrations. Sample volume, 1 L; stirring speed,1id

iswcsprozatAowaocrof

FA(

rpcFdmTFfaA

4

30 ± 5 rpm; 25 ◦C; stirring time, 2 h. (a) Dosage of Al2O3, 0.5–6.0 g/L; As(V)nitial concentration, 200 �g/L. (b) Initial As(V) concentration, 200–600 �g/L;osage of Al2O3, 2.0 g/L.

s superior to Al in ligand exchange with arsenate anions on theurface of Fe and Al oxide, that is, arsenic removal efficienciesith different electrode materials in an electrocoagulation pro-

ess follow the sequence: iron > aluminum [56] and thereforeupports the above-noted observations. This result is also sup-orted by several studies where activated alumina used in the pHange of 5.5–8.5 preferred OH− to H2AsO4

− [20,41,57]. Basedn studies of activated alumina and aluminum-loaded Shirasu-eolite [41,45], the As(V) adsorption mechanism of Al2O3 canlso be considered a ligand exchange process between As(V) andhe hydroxide groups. According to these results for Fe2O3 andl2O3, we can conclude that the favorable arsenate adsorptionf both adsorbents takes place at pH 6. This experimental result,here As(V) removal uptake on Fe2O3 and Al2O3 increased

t low pH and decreased at high pH, was the same as thatbserved with other absorbents of Fe and Al oxides such as ferric

hloride [58], ferrihydrite [43], ferric oxide [26], hydrous fer-ic oxide [27], goethite and gibbsite [29], amorphous aluminumxide [30], and activated alumina [32] as well as amorphouserric hydroxide and ferric-impregnated activated carbon. It was

ic

c

ig. 7. Adsorption isotherm plots for As(V) onto (a) Fe2O3 and (b) Al2O3.s(V) initial concentration, 600 �g/L; pH 5–9 ± 0.1: (a) dosage, 0.05–1 g/L,

b) dosage, 0.5–6 g/L.

evealed that surface or groundwater should be controlled in theH range of less than 7 to achieve better arsenate removal effi-iency when Fe2O3 or Al2O3 is used as an absorbent, and thate2O3 is less sensitive to pH than Al2O3. The observed slightecreases of As(V) removal efficiencies in the pH range of 5–6ay result from the dissolutions of Fe2O3 and Al2O3 [59,60].he pH values of the solution did not change significantly whene2O3 and Al2O3 were used and were almost between 4 and 7or Fe2O3 and 5.3 and 7.4 for Al2O3 (note that pH values weredjusted from 5 to 9 ± 0.1 before adsorption with Fe2O3 andl2O3).

.4. Adsorption isotherm

To design an appropriate sorption system for removing As(V)

n drinking water, it is important to find the well-fitted isothermurves of Fe2O3 and Al2O3.

Fig. 7 shows the relationship between As(V) equilibrium con-entration and the adsorption capacity of Fe2O3 and Al2O3 at

Y. Jeong et al. / Chemical Engineering and Processing 46 (2007) 1030–1039 1037

Table 3Arsenate adsorption isotherm parameters of Fe2O3 and Al2O3

Adsorbents Adsorption pH

Isotherms Parameters 5 6 7 8 9

Fe2O3 Langmuirb (L/mg) 42.01 47.12 32.83 23.36 5.09qmax (mg/g) 0.65 0.66 0.56 0.49 0.47R2 0.90 0.92 0.87 0.86 0.85

A.08.16.86

titaatFT

F(0

Fcai(

l2O3 Langmuirb (L/mg) 10qmax (mg/g) 0R2 0

he range of pH 5–9. With increasing pH, a decreasing trendn the amount of arsenate taken up by Al2O3 was observed. Inhe isotherm studies, it was found that the experimental data fordsorption on Fe2O3 and Al2O3 fitted well with the Langmuirdsorption isotherm in the range of pH 5–9. Plots of adsorp-

ion capacity (mg/g) versus equilibrium As(V) concentration inig. 7 yielded straight lines for each of five different pH values.he Langmuir curves and parameters for As(V) removal using

ig. 8. Linearized Langmuir isotherm plots for As(V) onto (a) Fe2O3 andb) Al2O3. As(V) initial concentration, 600 �g/L; pH 5–9 ± 0.1: (a) dosage,.05–1 g/L, (b) dosage, 0.5–6 g/L.

AamaatbvTetcacaaevctts

ahcwi[fvF[bo

5

t

10.31 9.76 9.13 8.990.17 0.14 0.13 0.130.90 0.87 0.87 0.80

e2O3 and Al2O3 as shown in Fig. 8 and in Table 3. The cal-ulated parameters of the Langmuir isotherm model for Fe2O3nd Al2O3, as well as the correlation coefficients (R2) are listedn Table 3. It is found that, at pH 6, the correlation coefficientR2) values for the Langmuir isotherm for As(V) on Fe2O3 andl2O3 are 0.92 and 0.90, respectively. Therefore, the maximum

dsorption capacities of Fe2O3 and Al2O3 at pH 6 were esti-ated from the Langmuir isotherm, and found to be 0.66 mg/g

nd 0.17 mg/g, respectively. These results show that Fe2O3 isbetter absorbent than Al2O3. As noted above, the parame-

er b is a function of the strength of adsorption. The largermeans that the adsorption bond is stronger. With a larger b

alue, Langmuir isotherm curves approach a saturation plateau.he isotherm shows that no more adsorbate can be removedven at low loading ratios of adsorbate to adsorbent. In turn,he adsorption capacity is independent of the equilibrium As(V)oncentration. As shown in Table 3, the value of b for Fe2O3nd Al2O3 increases with decreasing pH. Thus, we can con-lude that at lower pH, the As(V) adsorption bond with Fe2O3nd Al2O3 is stronger and more irreversible, and that the As(V)dsorption bond with Fe2O3 is stronger than with Al2O3. How-ver, at pH 9, As(V) adsorption bonds with Fe2O3 and Al2O3 areery weak, and the qe values change remarkably with even smallhanges in the equilibrium As(V) concentration, Ce. Based onhe assumptions of the Langmuir isotherm, it can be estimatedhat both Fe2O3 and Al2O3 should have mainly homogenousites.

Compared with some adsorbents that having high As(V)dsorption capacity such as ferrihydrite, amorphous aluminumydroxide, and mesoporous alumina [30,35,61], the adsorptionapacities of Fe2O3 and Al2O3 are low. However, comparedith other adsorbents that have lower As(V) adsorption capac-

ties such as IOCS, activated alumina, and activated red mud24,42,61], Fe2O3 can specifically be considered as a very use-ul adsorbent for As(V) removal in drinking water due to itsery low cost. According to the As(V) removal capacity ofe2O3 obtained in this research and that of ferrihydrite reported62–64], and their market prices, it is anticipated that Fe2O3ased POU As(V) removal cost is at least 50% lower than thatf ferrihydrite.

. Conclusions

Iron oxide (Fe2O3) and aluminum oxide (Al2O3) were foundo be good and inexpensive adsorbents for lowering As(V) ini-

1 ng an

tapaawrFtiilonAteipatIpSama

sPothfaris

A

Ib

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

038 Y. Jeong et al. / Chemical Engineeri

ial concentration in drinking water due to the fast adsorption ofrsenate anions and their high As(V) removal efficiency at lowerH (<7). As(V) adsorption on Fe2O3 and Al2O3, both of whichre nonporous adsorbents, is mainly controlled by the surfacerea of the adsorbents. It was observed that over 95% of As(V)as adsorbed within short contact times using Fe2O3 and Al2O3,

espectively. The maximum As(V) removal efficiency for bothe2O3 and Al2O3 was observed at a pH value of 6. In addition,

he removal efficiency for Fe2O3 did not vary much with vary-ng pH compared with that of Al2O3. In the isotherm studies,t was found that the observed data on Fe2O3 and Al2O3 fol-owed the Langmuir adsorption isotherm between the pH valuesf 5 and 9, which means both adsorbents have mainly homoge-ous sites on their surfaces. Fe2O3 is a better absorbent thanl2O3 for achieving higher adsorption capacity and initial sorp-

ion rate. Al2O3 and especially Fe2O3 have many advantagesven though the As(V) adsorption capacity of Fe2O3 and Al2O3s low compared with that of other absorbents such as amor-hous ferric oxide, granular ferric hydroxide, granular activatedlumina, and mesoporous alumina. Because Fe2O3 comes fromhe regeneration process of a by-product, it may be economical.n addition, these materials do not need any other pretreatmentrocess, such as pore increasing agents or desorption agents.ignificant water quality deterioration did not occur after As(V)dsorption using Fe2O3 and Al2O3. Therefore, these adsorbentsay be easily used at water treatment facilities due to their many

dvantages.Fe2O3 seems to be good for POE and POU water treatment

ystems because of the advantages outlined above. POE andOU arsenate removal systems utilizing iron oxide or aluminumxide can be considered for either small-scale commercial waterreatment systems for community water supplies or individualome treatment systems. These absorbents (Fe2O3 and Al2O3)or removing arsenate should be very useful in most endemicreas where arsenate-contaminated wells are used as wateresources (e.g., China, India, and Bangladesh) as simplicityn application and low cost are often the essential factors foruccessful performance of adsorbents.

cknowledgements

We thank Chester Lo for SEM analysis in Ames Laboratory,owa State University, Iowa. This research was supported in party the Busan Metropolitan Water Works in Korea.

eferences

[1] D.K. Nordstrom, C.N. Alpers, Negative pH, efflorescent mineralogy, andconsequences for environmental restoration at the Iron Mountain Super-fund site California, Proc. Natl. Acad. Sci. U.S.A. 96 (7) (1999) 3455–3462.

[2] US EPA (US Environmental Protection Agency), 1999. A review ofcontaminant occurrence in public water systems [EPA 816-R-99-006].http://www.epa.gov/safewater/occur/occur.html, last visited August 4,

2005.

[3] NAS (National Academy of Sciences), 2001, Arsenic in drinking water.http://www.epa.gov/safewater/arsenic.html, accessed September 3, 2005.

[4] J.F. Ferguson, J. Gavis, Review of the arsenic cycle in natural waters, WaterRes. 6 (1972) 1259–1274.

[

d Processing 46 (2007) 1030–1039

[5] W.R. Penrose, Arsenic in the marine and aquatic environments; analysis,occurrence and significance, CRC Crit. Rev. Environ. Control 4 (1974)465–482.

[6] P. Frank, D. Clifford, As(III) oxidation and removal from drinking water.USEPA Proj. Summary 600/S2-86/021, 1986.

[7] D.W. Oscarson, P.M. Huang, W.K. Liaw, U.T. Hammer, Kinetics of oxida-tion of arsenite by various manganese dioxides, Soil Sci. Soc. Am. J. 47(4) (1983) 644–648.

[8] G.F. Lauf, M.A. Waer, Arsenic removal using potassium permanganate,in: Proceedings of the Water Quality Technology Conference (PT.2), 1993,pp. 1025–1038.

[9] C.J. Chen, C.W. Chen, M.M. Wu, T.L. Kuo, Cancer potential in liver, lung,bladder and kidney due to ingested inorganic arsenic in drinking water, Br.J. Cancer 66 (5) (1992) 888–892.

10] M.M. Wu, T.L. Kuo, Y.H. Hwang, C.J. Chen, Dose–response relationbetween arsenic concentration in well water and mortality from cancersand vascular diseases, Am. J. Epidemiol. 130 (6) (1989) 1123–1132.

11] A.H. Smith, C.B. Hopenhayn-Rich, N. Michael, H.M. Goeden, I. Hertz-Picciotto, H.M. Duggan, R. Wood, M.J. Kosnett, M.T. Smith, Cancerrisks from arsenic in drinking water, Environ. Health Perspect. 97 (1992)259–267.

12] T. Murphy, J. Guo, Aquatic Arsenic Toxicity and Treatment, BackhuysPublishers, Leiden, The Netherlands, 2003.

13] US EPA, 2000. Arsenic occurrence in public drinking water supplies.[EPA-815-R-00-023]. http://www.epa.gov/safewater/ars/occurrence.pdf,last visited August 5, 2005.

14] D.N Mazumder, R. Haque, N. Ghosh, B.K. De, A. Santra, D. Chakraborti,A.H. Smith, Arsenic in drinking water and the prevalence of respiratoryeffects in West Bengal India, Int. J. Epidemiol. 29 (6) (2000) 1047–1052.

15] A.H. Smith, E.O. Lingas, M. Rahman, Contamination of drinking-waterby arsenic in Bangladesh: a public health emergency, Bull. World HealthOrgan. 78 (9) (2000) 1093–1103.

16] WHO (World Health Organization), 2001, Arsenic in drinking water.http://www.who.int/mediacentre/factsheets/fs210/en/, last visited August12, 2005.

17] US EPA, 2003. The arsenic rule: background and rule provisions.http://www.epa.gov/safewater/ars/presentations/train1-background.pdf,last visited August 16, 2005.

18] US EPA, 2001. National primary drinking water regulations. 66(14).http://www.epa.gov/safewater/mcl.html#mcls, last visited August 6, 2005.

19] D.D. Do, Adsorption Analysis: Equilibria and Kinetics, Imperial CollegePress, London, UK, 1998.

20] E. Rosenblum, D. Clifford, Equilibrium arsenic capacity of activated alu-mina [EPA-600/2-83-107], 1983, p. 89.

21] K.S. Subramanian, T. Viraraghavan, T. Phommavong, S. Tanjore, Man-ganese greensand for removal of arsenic in drinking water, Water Qual.Res. J. Canada 32 (3) (1997) 551–561.

22] O.S. Thirunavukkarasu, T. Viraraghavan, K.S. Subramanian, Arsenicremoval from drinking water using granular ferric hydroxide, Water SA29 (2) (2003) 161–170.

23] N.T. Livesey, P.M. Huang, Adsorption of arsenate by soils and its relationto selected chemical properties and anions, Soil Sci. 131 (2) (1981) 88–94.

24] H.S. Altundogan, S. Altundogan, F. Tumen, M. Bildik, Arsenic adsorptionfrom aqueous solutions by activated red mud, Waste Manage. 22 (3) (2002)357–363.

25] M.L. Pierce, C.B. Moore, Adsorption of arsenite and arsenate on amor-phous iron hydroxide, Water Res. 16 (7) (1982) 1247–1253.

26] D.A. Dzombak, F.M. Morrel, Surface Complexation Modelling-HydrousFerric Oxide, John Wiley, New York, USA, 1990.

27] J.A. Wilkie, J.G. Hering, Adsorption of arsenic onto hydrous ferric oxide:effects of adsorbate/adsorbent ratios and co-occurring solutes, ColloidsSurf. A: Physicochem. Eng. Aspects 107 (1996) 97–110.

28] A. Jain, K.P. Raven, R.H. Loeppert, Arsenite and arsenate adsorption on−

ferrihydrite: surface charge reduction and net OH release stoichiometry,

Environ. Sci. Technol. 33 (8) (1999) 1179–1184.29] B.A. Manning, S. Goldberg, Modeling arsenate competitive adsorption

on kaolinite, montmorillonite and illite, Clay Clay Miner. 44 (5) (1996)609–623.

ng an

[

[

[

[

[

[

[

[

[

[[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

Y. Jeong et al. / Chemical Engineeri

30] M.A. Anderson, J.F. Ferguson, J. Gavis, Arsenate adsorption on amorphousaluminum hydroxide, J. Colloid Interf. Sci. 54 (3) (1976) 391–399.

31] S.K. Gupta, K.Y. Chen, Arsenic removal by adsorption, J. Water Pollut.Control Fed. 50 (3) (1978) 493–506.

32] P. Singh, T.S. Singh, K.K. Pant, Removal of arsenic from drinking waterusing activated alumina, Res. J. Chem. Environ. 5 (3) (2001) 25–28.

33] S. Maeda, A. Ohki, S. Saikoji, K. Naka, Iron(III) hydroxide-loaded corallimestone as an adsorbent for arsenic(III) and arsenic(V), Sep. Sci. Technol.27 (5) (1992) 681–689.

34] R.L. Vaughan, B.E. Reed, Modeling As(V) removal by a iron oxide impreg-nated activated carbon using the surface complexation approach, Water Res.39 (6) (2005) 1005–1014.

35] Y. Kim, C. Kim, I. Choi, S. Rengaraj, J. Yi, Arsenic removal using meso-porous alumina prepared via a templating method, Environ. Sci. Technol.38 (3) (2004) 924–931.

36] MDE (Maryland Department of the Environment Water ManagementAdministration), 2004, Arsenic water treatment for individual wellsin Maryland. http://www.mde.state.md.us/assets/document/FactSheets/arsenic%20fact%20sheet%20final.pdf, accessed September 1, 2005.

37] K.N. Scott, J.F. Green, H.D. Do, S.J. McLean, Arsenic removal by coagu-lation, J. Am. Water Works Assoc. 87 (4) (1995) 114–126.

38] L.S. McNeill, M. Edwards, Predicting As removal during metal hydroxideprecipitation, J. Am. Water Works Assoc. 89 (1) (1997) 75–86.

39] Y.S. Ho, Ph.D. Thesis, University of Birmingham, Birmingham, UK, 1995.40] M.D. LeVan, Fundamentals of Adsorption, Kluwer Academic, Boston,

USA, 1996.41] AWWA, Water Quality and Treatment: A Handbook of Community Water

Supplies, 5th ed., Americam Water Works Association, 1999, p. 13.2.42] T.-F. Lin, J.-K. Wu, Adsorption of arsenite and arsenate within activated alu-

mina grains: equilibrium and kinetics, Water Res. 35 (8) (2001) 2049–2057.43] K.P. Raven, A. Jain, R.H. Loeppert, Arsenite and asenate adsorption on

ferrihydrite: kinetics, equilibrium, and adsorption envelopes, Environ. Sci.Technol. 32 (3) (1998) 344–349.

44] S. Goldberg, C.T. Johnston, Mechanisms of arsenic adsorption on amor-phous oxides evaluated using macroscopic measurements, vibrationalspectroscopy, and surface complexation modeling, J. Colloid Interf. Sci.234 (1) (2001) 204–216.

45] Y. Xu, T. Nakajima, A. Ohki, Adsorption and removal of arsenic(V) fromdrinking water by aluminum-loaded Shirasu-zeolite, J. Hazard. Mater. 92(3) (2002) 275–287.

46] S.R. Wickramasinghe, B. Han, J. Zimbron, Z. Shen, M.N. Karim, Arsenicremoval by coagulation and filtration: comparison of groundwaters fromthe United States and Bangladesh, Desalination 169 (3) (2004) 231–244.

47] APHA, AWWA, WEF, Standard Methods for the Examination of Waterand Wastewater, 20th ed., APHA, Washington, DC, USA, 1998.

48] T. Yuan, J.Y. Hu, S.L. Ong, Q.F. Luo, W.J. Ng, Arsenic removal from house-hold drinking water by adsorption, J. Environ. Sci. Health A: Toxic/Hazard.Subst. Environ. Eng. 37 (9) (2002) 1721–1736.

[

[

d Processing 46 (2007) 1030–1039 1039

49] P. Singh, W. Zhang, D.M. Muir, R.G. Robins, The effect of silicate on theadsorption of arsenate on coprecipitated ferrihydrite arsenic metallurgy,in: Proceedings of the Symposium Held During the TMS Annual Meet-ing, San Francisco, CA, United States, February 13–17, 2005, pp. 129–135.

50] T.H. Hsia, S.L. Lo, C.F. Lin, Arsenic(V) adsorption on amorphousiron oxide: triple layer modeling, Chemosphere 25 (12) (1992) 1825–1837.

51] W. Stumm, B. Sulzberger, The cycling of iron in natural environments: con-siderations based on laboratory studies of heterogeneous redox processes,Geochim. Cosmochim. Acta 56 (8) (1992) 3233–3257.

52] M.M. Benjamin, R.S. Sletten, R.P. Bailey, T. Bennett, Sorption and filtra-tion of metals using iron-oxide-coated sand, Water Res. 30 (11) (1996)2609–2620.

53] P.R. Grossl, M. Eick, D.L. Sparks, S. Goldberg, C.C. Ainsworth, Arsenateand chromate retention mechanisms on goethite. Part 2. Kinetic evaluationusing a pressure-jump relaxation technique, Environ. Sci. Technol. 31 (2)(1997) 321–326.

54] D.G. Brookins, Eh-pH Diagrams for Geochemistry, Springer-Verlag,Berlin, Germany, 1988.

55] A.I. Zouboulis, K.A. Kydros, K.A. Matis, Arsenic(III) and arsenic(V)removal from solutions by pyrite fines, Sep. Sci. Technol. 28 (15–16) (1993)2449–2463.

56] P. Ratna Kumar, S. Chaudhari, K.C. Khilar, S.P. Mahajan, Removal ofarsenic from water by electrocoagulation, Chemosphere 55 (9) (2004)1245–1252.

57] R.R. Trussell, A. Trussell et al. Selenium removal from groundwaterusing activated alumina [EPA 600280153], 1980, http://yosemite.epa.gov/water/owrccatalog.nsf/, last visited August 4, 2005.

58] J.G. Hering, P.Y. Chen, J.A. Wilkie, M. Elimelech, S. Liang, Arsenicremoval by ferric chloride, J. Am. Water Works Assoc. 88 (4) (1996)155–167.

59] S.L. Lo, T.Y. Chen, Adsorption of Se(IV) and Se(VI) on an iron-coatedsand from water, Chemosphere 35 (5) (1997) 919–930.

60] N.K. Pandey, Effect of pH on adsorption of U(VI) ions on TiO2, �-Al2O3,and SiO2 surface from aqueous solution, Oxid. Commun. 29 (3) (2006)698–707.

61] O.S. Thirunavukkarasu, T. Viraraghavan, K.S. Subramanian, Removal ofarsenic in drinking water by iron oxide-coated sand and ferrihydrite—batchstudies, Water Qual. Res. J. Canada 36 (1) (2001) 55–70.

62] M. Badruzzaman, P. Westerhoff, D.R.U. Knappe, Intraparticle diffusionand adsorption of As(V) onto granular ferric hydroxide (GFH), Water Res.38 (18) (2004) 4002–4012.

63] C. Namasivayam, K. Ranganathan, Removal of Cd(II) from wastewater byadsorption on waste Fe(III)/Cr(III) hydroxide, Water Res. 29 (7) (1995)1737–1744.

64] Q. Sun, L. Yang, The adsorption of basic dyes from aqueous solution onmodified peat-resin particle, Water Res. 37 (7) (2003) 1535–1544.