GC

Ma

b

h

••••

a

ARRAA

KMCGRC

1

zgbtilltmr

q

0h

Journal of Hazardous Materials 265 (2014) 20– 29

Contents lists available at ScienceDirect

Journal of Hazardous Materials

jou rn al hom epage: www.elsev ier .com/ locate / jhazmat

uar gum coupled microscale ZVI for in situ treatment ofAHs: Continuous-flow column study

ilica Velimirovica,b, Queenie Simonsa, Leen Bastiaensa,∗

Flemish Institute for Technological Research (VITO), Boeretang 200, 2400, Mol, BelgiumUniversity of Antwerp, Department of Bio-Engineering, Groenenborgerlaan 171, 2020 Antwerp, Belgium

i g h l i g h t s

Guar gum slightly reduces mZVI reactivity under continuous flow conditions.Soil microbial population does degrade guar gum.Guar gum slows down mZVI corrosion and consequently prolongs mZVI lifetime.mZVI and guar gum stabilized mZVI do not significantly affect bacterial activity.

r t i c l e i n f o

rticle history:eceived 5 July 2013eceived in revised form 7 November 2013ccepted 10 November 2013vailable online 16 November 2013

eywords:icroscale zerovalent iron

a b s t r a c t

A column study was performed under in situ conditions to evaluate to which extend the inactivationof the microscale zerovalent iron (mZVI) by guar gum occurs under continuous flow conditions. Fiveaquifer containing columns were set up under different conditions. Efficient removal of trichloroethenewas observed for the column amended by mZVI. Stabilization of the mZVI with guar gum led to slightlyreduced activity. More reduced reactivity was observed in the poisoned column containing guar gumstabilized mZVI. This confirms that soil microorganisms can degrade guar gum and that subsequentremoval of the oligosaccharides by the groundwater flow (flushing effect) can reactivate the mZVI. After

hlorinated aliphatic hydrocarbonsuar gumeactive zoneolumn tests

more than six months of continuous operation the columns were dismantled. DNA-based qPCR analysisrevealed that mZVI does not significantly affect the bacterial community, while guar gum stabilized mZVIparticles can even induce bacterial growth. Overall, this study suggests that the temporarily decreasedmZVI reactivity due to guar gum, has a rather limited impact on the performance of in situ reactive zones.The presence of guar gum slightly reduced the reactivity of iron, but also slowed down the iron corrosionrate which prolongs the life time of reactive zone

. Introduction

Two decades after Gillham and O’Hannesin [1] introduced theerovalent iron (ZVI) particles as a new concept to efficiently treatroundwater contaminated with chlorinated aliphatic hydrocar-ons (CAHs), permeable reactive barriers (PRBs) are a commonechnology for in situ remediation of CAHs. PRBs are constructedn the subsurface by digging and refilling a trench using granu-ar ZVI [1,2], which is technically and economically possible forimited depths [3]. To reduce the cost of PRB installation and expand

he application area of ZVI, researchers focused more recently on

icro- and nanoscale ZVI particles. These particles have highereactivity towards different CAHs and the potential to be injected

∗ Corresponding author. Tel.: +32 14 33 5634; fax: +32 14 583 0523.E-mail addresses: milica.velimirovic@vito.be (M. Velimirovic),

ueenie.simons@vito.be (Q. Simons), leen.bastiaens@vito.be (L. Bastiaens).

304-3894/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.jhazmat.2013.11.020

© 2013 Elsevier B.V. All rights reserved.

in the deep subsurface (>10–20 m) as a slurry creating in situ reac-tive zones (RZ). However, estimated lifetimes of several weeks fornanoscale zerovalent iron (nZVI) particles [4,5] as well as fast aggre-gation and filtration in the porous medium with reduced radiusof influence [6–10], indicate that injection of nZVI slurry is morechallenging than initially expected. On the other hand, microscalezerovalent iron (mZVI) particles have longer life time [11,12], theyare less expensive and pose less risk for human health than nZVIparticles [13]. Nevertheless, sedimentation of mZVI particles in theslurry reservoir, tubing and injection wells prior to and duringinjection, which is a consequence of the high density and size ofparticles [14,15], is crucial when designing mZVI reactive zones.

To prevent sedimentation of mZVI particles during the injectionperiod, the use of shear thinning fluids may offer a solution to trans-

port mZVI particles through porous media [15–19]. Comba andBraun [20] reported guar gum as a polymer suitable to keep mZVIparticles in suspension. Guar gum (GG) is a common biopolymerfor stabilizing trenches during the soil excavation and construction

M. Velimirovic et al. / Journal of Hazardous Materials 265 (2014) 20– 29 21

Table 1Characteristics of studied mZVI.

Sample PSDa D10, D50, D90 (�m) Carbon content (%) Oxygen content (%) Sulfur content (%) Nitrogen content (%) BETb (m2 kg−1)

mZVI 24, 56, 69 0.02 1.10 0.00 0.02 57

easu

ogbtwt

udmcptgeawe

2

2

fima(tne

2

fl4r(nca(dfiomroa2ahCmw

sampling points located at every 10 cm along the column, to deter-mine CAHs concentrations and their breakdown products. pH and

Table 2Chemistry of groundwater used in the column study.

Parameters Unit

Field parametersTemperature ◦C 12 ± 0.1pH 6 ± 0.02ORP mV −56 ± 10Conductivity �s cm−1 858 ± 0Dissolved O mg L−1 0.6 ± 0.07

Redox parametersCO3

2− mg L−1 0.00HCO3

− mg L−1 82.3Dissolved Ca mg L−1 34.2Dissolved Fe mg L−1 121Dissolved Mn mg L−1 0.62Cl− mg L−1 114SO4

2− mg L−1 201

CAHsChloroethane (CA) �g L−1 30 ± 91,1-dichloroethene (1,1-DCE) �g L−1 310 ± 17Trans-dichloroethene (tDCE) �g L−1 42 ± 11,1-dichloroethane (1,1-DCA) �g L−1 6047 ± 382Cis-dichloroethene (cDCE) �g L−1 1806 ± 639

a Particle Size Distribution.b BET: Specific Surface Area according to Brunauer-Emmett-Teller (Single point m

f a treatment walls [21], but inactivation of ZVI reactivity by guarum has been reported [21]. Recently, we published a detailedatch degradation study, proving that guar gum negatively impactshe reactivity of mZVI in liquid medium. However, once guar gumas degraded and breakdown products removed, full recovery of

he mZVI reactivity was obtained [22].Here we report the results of a column test that was performed

nder in situ conditions to evaluate the impact of guar gum on theegradation efficiency of mZVI particles in the presence of aquiferaterial and under continuous groundwater flow conditions. Spe-

ial attention was given to (1) the duration of the inactivationeriod for guar gum stabilized mZVI particles (GG-mZVI), and (2)he ability of naturally present microorganisms to degrade the guarum to soluble oligosaccharides, as a substitute for commercialnzymes used in our previous study. In addition, the impact of mZVInd guar gum on the naturally present soil bacterial communityas examined via microbial analyses performed at the end of the

xperiment using 16S rRNA genes based molecular techniques.

. Materials and methods

.1. Micro-iron particles

The reactive mZVI particles used in this study were obtainedrom Höganäs (Sweden). Characteristics of the particles are givenn Table 1. Guar gum HV7000 (Rantec Corporation, USA) was used as

ZVI stabilizer. For the column test, mZVI and guar gum were useds received. Previous batch degradation tests with selected mZVIdata not shown) using standardized test procedure [23] indicatehat CAHs were degraded predominantly by �-elimination witho vinyl chloride (VC) as an intermediate product, and ethene andthane as the final degradation products.

.2. Column test

The one dimensional column test setup is shown in Fig. 1. Fiveow-through plexiglas columns (height, 50 cm; inside diameter,

cm) were wet-filled in an anaerobic glove box (nitrogen) with rep-esentative CAHs-contaminated groundwater and aquifer materialcollected from 8 m below ground surface) from a real contami-ated site in Belgium. Detailed groundwater chemistry and aquiferontamination are presented in Table 2 and SI1 respectively. Thequifer material was used as received and mixed with filter sand1–2 mm, considered nonreactive, 25%, w/w) to increase the con-uctivity of the medium. The first column (C-1) was completelylled with the homogenized mixture of aquifer and sand with-ut mZVI to investigate the biodegradation potential of aquiferaterial from site. Column 2 (C-2) contained the same filling mate-

ial amended with 25 g kg−1 of mZVI to investigate the reactivityf non-stabilized mZVI. To compare reactivity of non-stabilizednd guar gum stabilized mZVI, Column 3 (C-3) was amended with5 g kg−1 of mZVI as well as 2 g kg−1 of guar gum. Column 4 (C-4)nd 5 (C-5) were poisoned by addition of 0.25% (v/v) of formalde-

yde to the feed [24] and served as abiotic controls (AC) of C-1 and-3. To obtain uniform distributions of filling materials, approxi-ately 10 portions of 100 g of aquifer/sand mixture were preparedith and without mZVI/GG-mZVI, from which were added all in

rement)—analysis was conducted by Höganäs

aliquots of 20 g to the columns which contained autoclaved deoxy-genized MilliQ water.

The columns were fed with the real groundwater, mainlypolluted by 6.0 ± 0.4 mg L−1 of 1,1-dichloroethane (1,1-DCA),0.3 ± 0.1 mg L−1 of trichloroethene (TCE), 1.8 ± 0.6 mg L−1 of cis-dichloroethene (cDCE) and 0.3 ± 0.0 mg L−1 of 1,1-dichloroethene(1,1-DCE), that are frequently detected groundwater contaminantsin industrialized countries. It was continuously pumped (WatsonMarlov 205s peristaltic pump) in an upward flow through thecolumns with an average total flow rate of 18 ± 3 mL day−1, cor-responding to an initial pore water velocity of 4 ± 2 cm day−1 andhydraulic retention time of 13 ± 5 days. Characteristics of everycolumn individually are given in SI2. The highest flow rate wasobserved for C-1 and C2, while the lowest for C-5. The highest porewater velocity was observed for C-2 (7.31 cm day−1) providing theshortest contact time. The longest contact time was observed for C-4 (2.83 cm day−1) and C-5 (2.91 cm day−1). The pore water velocityfor C-1 and C-3 was 4.64 cm day−1 and 4.23 cm day−1, respectively.The columns were operated at temperature of 12 ◦C. To avoid acci-dental microbial contamination, columns and columns parts weresterilized by autoclaving before start of experiment.

During operation, influent and effluent of the columns weresampled every two weeks for measurements of CAHs concentra-tions and their breakdown products. 5 mL of liquid samples wasdirectly collected by connecting 12 mL vials (previously capped andflushed with nitrogen) to a sampling port allowing the samplesto spontaneously flow into the vials. In addition, after 4, 5 and 6months of operation, liquid samples were taken at intermediate

1,1,1-trichloroethane (1,1,1-TCA) �g L−1 15 ± 3Trichloroethene (TCE) �g L−1 344 ± 109Tetrachloroethene (PCE) �g L−1 17 ± 13Vinyl chloride (VC) �g L−1 7 ± 2

22 M. Velimirovic et al. / Journal of Hazardous Materials 265 (2014) 20– 29

Fig. 1. Schematic overview of the laboratory continuous flow column test set up. All columns were operated under anaerobic conditions. C-4 and C-5 were included as abioticc

ra

wi

2

pgio

ontrols (AC). GG: guar gum.

edox potential (ORP) were measured in effluent samples after 4, 5nd 6 months of column operation.

After 5 months of column operation 100 �L of effluent samplesere withdrawn for adenosine tri-phosphate (ATP) analysis as an

ndicator for viable biomass estimation in groundwater [25–27].

.3. Column dismantling

After more than 6 months of operation (corresponding to 15 ± 5

ore volumes), the columns were dismantled in an anaerobiclove box (Jacomex, France). Column filling was vertically dividednto ten sections of 5 cm length, beginning at the influent sidef the column. For selected column sections (0–5 cm, 20–25 cm

and 40–45 cm) composite subsamples of approximately 2 g weretaken for DNA based microbial analyses. For guar gum analysis,approximately 5 g of homogenized sample was taken from selectedselections (0–10 cm, 10–20 cm, 20–30 cm, 30–40 cm and 40–50 cm)and extracted with 15 mL of anaerobic deionized water for oneweek. To quantify the remaining iron in solid samples, approx-imately 5 g of homogenized sample was taken from the samecompartments.

2.4. Chemical analysis

The concentrations of CAHs, intermediate- and end-productswere determined via direct headspace measurements using a

M. Velimirovic et al. / Journal of Hazardous Materials 265 (2014) 20– 29 23

Fig. 2. Measured concentrations of TCE, cDCE, 1,1-DCE and 1,1-DCA in effluent samples of columns containing aquifer material (C-1), non-stabilized mZVI (C-2) and guarg 4) andi

VRlaetwOT2wassm

mqmdpgG(at

and probes were synthesized by Operon Biotechnologies GmbH

um stabilized mZVI (C-3). Two abiotic controls were included: aquifer material (C-ncluded for comparison.

arian GC-FID (CP-3800 with CTC-autosampler) equipped with at-U plot column for the detection of ethene, ethane and acety-

ene or a split-splitless injector followed by a Rt-X column (Restek)nd a DB-1 column (J&W Scientific) for analysis of CAHs [23]. Anxternal calibration standard was prepared in the 12 mL vial withhe same water/headspace volume ratio as the samples. The pHas measured with a HI1330 electrode (Hanna instruments). TheRP was measured with a PT5900A electrode (Schott Instruments).he reference system was a Ag/AgCl electrode with a potential of10 mV (20 ◦C). Guar gum concentration in effluent samples, asell as column fillings, were determined by the phenol–sulfuric

cid tests using colorimetric method [28]. The absorbance mea-urements were performed using a Spectronic Genesys 6 UV–vispectrophotometer (Thermo Electron Corporation) with the maxi-um absorbance for guar gum occurring at �max = 490 nm.To quantify remaining ZVI in solid samples obtained after dis-

antling of columns, acid digestion followed by hydrogen (H2)uantification was used [29,30]. Approximately 5 g of sample wasixed with 5 mL of 1 M HCl during one month for complete acid

igestion of the highest possible concentration of ZVI (25 g kg−1)resent in the sample. After one month of incubation, the hydro-en gas produced was quantified in headspace using a gas Trace

C MPT-10286 (Interscience) equipped with HayeSepQ column

Alltech Associates, Inc.) and a thermal conductivity detector. Themount of iron retained in each column was calculated assuminghat 1 mol of H2 is produced for 1 mol of zerovalent iron [31].

guar gum stabilized mZVI (C-5). Average influent concentrations (n = 2) were also

2.5. Calculation of rate coefficients

The pseudo-first-order model was applied to describe the reduc-tive dechlorination of CAHs by mZVI and guar gum stabilized mZVI[22]. The degradation rates were calculated by fitting the exper-imental breakthrough curves using the linear regression method.Residence times were calculated using porosity and water flow,while dispersion was neglected.

2.6. Microbial characterization

Total ATP in 100 �L of effluent samples was measured using the100 �L of BacTiter-GloTM reagent (G8231; Promega Corporation,Dübendorf, CH) and Luminoskan Ascent Microplate Luminometer(Thermo Scientific) as described in SI3.

DNA was extracted from the aquifer/sand/(mZVI)samples asdescribed by Hendrickx et al. [32]. The population dynamics wasquantified using real time quantitative PCR (qPCR) with differ-ent group specific primer sets, targeting total bacteria [33,34],Dehalococcoides spp. [35], SRB [36] and Geobacter [37]. Primers

(Cologne, Germany). All qPCR runs were performed on the West-burg Rotor-Gene 3000 (Corbett Research, Sydney, Australia) andthe data were analyzed using the quantification analysis with thebuilt-in software.

2 Hazardous Materials 265 (2014) 20– 29

3

3g

icDottatslsCcacw

rcirccercdshu−paam(6idmpnir

3c

pC

cwrraiC

0

0.5

1

1.5

2

2.5

876543210

guar

gum

(g L-1

)

# pore volumes

mZVI + GG + aqui fer (C-3)mZVI + GG + aqui fer AC (C-5)

4 M. Velimirovic et al. / Journal of

. Results and discussion

.1. CAH degradation in continuous flow column systems withoutuar gum

The timely evolutions of TCE, cDCE, 1,1-DCE and 1,1-DCAnfluent and effluent concentrations (Fig. 2) indicate during theonditioning phase (week 1–7) some retardation of TCE, cDCE, 1,1-CE and 1,1-DCA in columns (C-1 and C-4), explicable by sorptionnto the aquifer solids. Once 3 pore volumes (PV) were pumpedhrough the column (week 5–7, depending on the column sys-em as described in SI2), retardation of contaminants decreasednd the columns were considered close to steady-state condi-ion (60–82% of contaminants were recovered from the influentample). Nevertheless, ongoing sorption onto organic matter andosses due to diffusive-dispersive processes during for instanceampling events cannot be excluded completely afterwards. NoAH-biodegradation in the aquifer was observed as C-4 (abioticontrol) showed a similar pattern as C-1 (only “aquifer”). Addition-lly, no significant production of VC (indication of biodegradation),hloroethane (CA), ethene and ethane was observed in comparisonith influent concentrations.

In the presence of non-stabilized mZVI (C-2), the averageemoval of TCE was 100%. cDCE, 1,1-DCE and 1,1-DCA effluent con-entrations decreased 37, 76 and 8%, respectively, compared to thenfluent. In contrast to cDCE, high efficiencies in TCE and 1,1-DCEemoval by non-stabilized mZVI might be explained by signifi-antly lower concentrations present in the groundwater. DecreasedDCE disappearance rates in the presence of TCE have also beenxplained before by interspecies competition [38]. The very limitedemoval of 1,1-DCA is in line with very slow degradation of theseontaminant after the contact with mZVI [39]. The main degra-ation products observed were ethene and ethane (SI4), whileignificant concentrations of VC (<40 �g L−1) and CA (<60 �g L−1)ave not been observed. During passage of the water through col-mn C-2, the average ORP decreased slightly from −56 ± 10 mV to85 ± 28 mV, while no significant pH change was observed (fromH 6 ± 0.02 to pH 6 ± 0.6). These trends in ORP are explicable bynaerobic corrosion of the present mZVI [2,31,40]. In addition, ORPnd pH values in the effluent of the columns measured after 6onths of continuous operation were −44 (C-1), −105 (C-2), −112

C-3), −24 (C-4) and −126 (C-5) mV and 5.19 (C-1), 5.82 (C-2),.80 (C-3), 5.09 (C-4) and 5.18 (C-5), respectively. The pH change

s less pronounced than previously observed in the batch degra-ation experiments [22]. The buffering properties of the aquiferaterial in the columns and the continuous operation mode are

ossible explanations. The presented results show that the testedon-stabilized mZVI is highly efficient in reducing TCE and 1,1-DCE

n aquifers under continuous flow conditions and can be used as aeactive material for in situ applications.

.2. CAH reduction by guar gum stabilized mZVI (GG-mZVI) inontinuous flow column systems

The effect of guar gum on mZVI reactivity was examined by com-aring the results of C-2 (non-stabilized mZVI), C-3 (GG-mZVI) and-5 (GG-mZVI, abiotic control).

Guar gum concentrations measured in effluent samples of theolumns revealed guar gum removal from both C-3 and C-5 (Fig. 3),hich is a crucial step for reactivity recovery of GG-mZVI. Indeed,

emoval of guar gum fragments from the GG-mZVI is expected to

esult in more efficient treatment of CAHs. Four weeks (1.5 PVs)fter start-up, the guar gum concentration decreased significantlyn the effluent of C-3 (88%), as well as in the effluent of the poisoned-5 (87%), indicating that guar gum was flushed out of the columns

Fig. 3. Guar gum (GG) concentration in effluent samples of the continuous columnflow systems amended with guar gum stabilized mZVI (C-3) and its abiotic control(C-5) as a function of number of pore volumes pumped through the column.

by the groundwater flow. At that time, the effluent samplings forCAH-analyses were started.

The columns simulate a 50 cm wide in situ reactive system witha groundwater flow velocity of approximately 10 m year−1. Accord-ing to data from a previously reported batch degradation study,where minimal 3 pore volumes were estimated to be requiredfor full reactivity recovery of GG-mZVI [22], a reduced reactivityperiod of approximately 2 months can be predicted for the col-umn test. Data from the first CAH-measurements (1.5 months afterstart) showed already that the TCE removal efficiency of GG-mZVI(C-3) was comparable to the non-stabilized mZVI (C-2) after condi-tioning phase (Fig. 2). The average efficiency of TCE removal in thecolumns was the highest (99–100%), followed by 1,1-DCE (67%).On the other hand, for 1,1-DCA and especially cDCE, a reducedremoval was observed in C-3. At the end of the conditioning phase(2 months), the difference between C-2 and C-3 nearly disappeared.For cDCE, and 1,1-DCA, an average concentration decrease of 20%and 8% respectively was observed. An increase of final degradationproducts such as ethene and ethane was observed (SI5). However,significantly lower concentrations were observed for GG-mZVIand GG-mZVI (AC) compared to mZVI. The same was previouslyreported in batch tests [22]. ORP (−115 ± 33 mV) and pH (7 ± 0.08)values in the effluent of GG-mZVI amended column were similarto the results obtained for the column amended by non-stabilizedmZVI, pointing again at anaerobic corrosion of the mZVI generatingFe2+ on the iron grain surface [2,31,40]. However, further formationand precipitation of ferrous hydroxide is expected to be slow dueto the high buffer capacity of the aquifer material and no significantchanges in pH [41].

To compare the reactivity of non-stabilized mZVI and GG-mZVIin more detail, TCE, cDCE, 1,1-DCE and 1,1-DCA concentration pro-files along the columns were recorded. The results obtained after4 months of operation (10 PV) are presented in Fig. 4. Guar gumhas a slightly negative impact on the TCE removal by mZVI. Sim-ilar results were obtained after 5 months of operation (SI5). After6 months, TCE-removal by GG-mZVI and mZVI was similar again(SI5), and the reactivity of GG-mZVI (abiotic control) was improvedcompared to the TCE-removal obtained after 4 and 5 months. Thiscan be explained by the intensive rinsing process resulting in lowerconcentration of the guar gum present as a coupling layer on themZVI [22]. Similar trends were observed for 1,1-DCE. In contrastGG-mZVI and GG-mZVI (AC) were not capable to remove cDCE and

1,1-DCA.

Pseudo-first-order rate disappearance constants (kobs) with cal-culated half life time (t1/2) were used to compare reactivity of mZVIin C-2 and C-3, taking into account residence times calculated for

M. Velimirovic et al. / Journal of Hazardous Materials 265 (2014) 20– 29 25

0

0.2

0.4

0.6

0.8

1

1.2

50403020100

C/C 0

distance (cm)

TCE

0

0.2

0.4

0.6

0.8

1

1.2

50403020100

C/C 0

distanc e (cm )

cDCE

0

0.2

0.4

0.6

0.8

1

1.2

50403020100

C/C 0

distanc e (cm )

1,1 -DCE

0

0.2

0.4

0.6

0.8

1

1.2

50403020100

C/C 0

distanc e (cm )

1,1 -DCA

mZVI + aqui fer (C-2) mZVI + GG + aqui fer (C-3) mZV I + GG + aqui fer AC (C-5)

Fig. 4. Breakthrough curves for TCE, cDCE, 1,1-DCE and 1,1-DCA dechlorination using non-stabilized mZVI (C-2), GG-mZVI (C-3) and GG-mZVI—abiotic control (C-5) after 4months of operation.

Table 3Summary of the results for TCE, cDCE, 1,1-DCE and 1,1-DCA disappearance rate constants with half life time of the selected chlorinated compounds in continuous flow columnsystems after 4, 5 and 6 months of operation.

TCE (0.5 ± 0.1 mg L−1)a cDCE (0.9 ± 0.4 mg L−1) 1,1-DCE (0.7 ± 0.4 mg L−1) 1,1-DCA (5.4 ± 0.9 mg L−1)

kobb t1/2

c kob t1/2 kob t1/2 kob t1/2

After 4 months of operationmZVI + aquifer (C-2) 2.75 × 100 0.25 4.84 × 10−2 14.3 1.15 × 10−1 6.02 3.62 × 10−2 19.2mZVI + GG + aquifer (C-3) 9.62 × 10−1 0.72 1.08 × 10−2 64.2 3.64 × 10−2 19.0 3.80 × 10−3 183d

mZVI + GG + aquifer AC (C-5) 3.89 × 10−1 1.78 1.57 × 10−3 442d 1.82 × 10−2 38.1 3.87 × 10−3 179d

After 5 months of operationmZVI + aquifer (C-2) 2.60 × 100 0.27 5.69 × 10−2 12.2 2.25 × 10−1 3.08 3.02 × 10−2 22.9mZVI + GG + aquifer (C-3) 8.20 × 10−1 0.85 1.02 × 10−2 68.2 1.10 × 10−1 6.30 4.36 × 10−2 15.9mZVI + GG + aquifer AC (C-5) 2.38 × 10−1 2.92 1.63 × 10−3 426d 3.46 × 10−2 20.1 2.90 × 10−3 239d

After 6 months of operationmZVI + aquifer (C-2) 2.51 × 100 0.28 7.37 × 10−2 9.41 2.44 × 10−1 2.84 5.52 × 10−2 12.5mZVI + GG + aquifer (C-3) 1.15 × 100 0.60 4.57 × 10−3 152 5.77 × 10−3 120 NDd NDd

mZVI + GG + aquifer AC (C-5) 7.82 × 10−1 0.89 1.63 × 10−3 426d NDd NDd NDd NDd

a Initial concentration or concentration range.

e41m

b kob the first order decay constant (day−1).c t1/2 half-life time (day).d No degradation.

ach column separately (Table 3). Data reflecting the condition after months of operation, provide evidence that kobs of TCE, cDCE and,1-DCE were 3–4 times lower for GG-mZVI than for non-stabilizedZVI. Even lower kobs were obtained for the abiotic control.

Comparable data were obtained after 5 months of the columnsoperation. Finally, after 6 months of operation, TCE disappear-ance rate was in the same order of magnitude for non-stabilizedand GG-mZVI. Calculated t1/2 < 1 day−1 for mZVI, GG-mZVI and

2 Hazardous Materials 265 (2014) 20– 29

GfigTbtpaioipctdmrpeiod

m4somsgtptiteaowr

0

10

20

30

40

50

60

40-50 cm30-40 cm20-30 cm10-20 cm0-10 cm

mg

GG

/ k

g of

col

umn

fillin

g

column lenght (cm)

mZVI + GG + aquifer (C-3) mZVI + GG + aquifer AC (C-5)

Fig. 5. Guar gum (GG) concentration in columns amended with GG- mZVI (C-3) and

Fo

6 M. Velimirovic et al. / Journal of

G-mZVI (AC) indicate that mZVI reactivity towards TCE in theeld application of the reactive zone is not significantly affected byuar gum. Similarly, Tratnyek et al. [42] concluded that decreasedCE reduction rates, due to competition for reactive surface sitesy adsorbed natural organic matter, does not significantly impacthe performance of iron permeable reactive barriers. However, theerformance of mZVI towards cDCE and 1,1-DCA is significantlyffected by guar gum and might be explained by the direct block-ng of reactive surface sites [9] which possibly has a larger impactn the lower degrading CAH-compounds like cDCE and 1,1-DCAn comparison with TCE or by significantly higher concentrationsresent in the groundwater. As previously explained, interspeciesompetition between cDCE and TCE, where TCE is more attractedo the reactive surface of ZVI [38] might also lead to the reducedegradation rate of cDCE in column systems. Interestingly, after 6onths of operation with non-stabilized mZVI, the disappearance

ate constants for cDCE, 1,1-DCE and 1,1-DCA were faster in com-arison to data collected after 4 and 5 months of operation. Possiblexplanations may be Fe(II)–Fe(III) salts (green rust) formation dur-ng the anaerobic corrosion of zerovalent iron [43] or formationf reactive FeS precipitates [44], which both can result in fasterisappearance of CAHs.

In the abiotic control C-5 a clearly reduced reactivity of GG-ZVI was observed (Figs. 2 and 4). kobs of TCE and 1,1-DCE after

and 5 months of operation were approximately 6–11 timeslower than for the non-stabilized particles, while no degradationf cDCE and 1,1-DCA was observed. These findings indicate that soilicroorganisms do have a function in degrading guar gum and the

ubsequent removal of the guar gum degradation products by theroundwater flow (intensive rinsing), both required to reactivatehe guar gum stabilized mZVI particles. This hypothesis is sup-orted by the guar gum concentrations analyzed at the effluent ofhe columns (Fig. 3). After 2 PVs, guar gum was no longer detectedn the effluent of guar gum stabilized mZVI, while the effluent ofhe corresponding abiotic control still contained guar gum. This isxplicable by the absence of active soil microorganisms. However,

fter 6 PVs almost no guar gum was detected in the effluent samplesf the abiotic control, while no full recovery of the mZVI reactivityas realized at that moment. One explanation might be that the

insing process as described earlier for batch degradation tests [22]

0

2

4

6

8

10

12

14

16

18

mZVI + aqu ife r(control sample)

aquifer AC (C-4)aquifer (C-1)

g ZV

I / k

g of

col

umn

fillin

g

Backgr oun dvalues

Backgroundvalues

ig. 6. ATP concentrations as an indicator for viable biomass in effluent samples. Abiotic con no viable biomass.

GG-mZVI—abiotic control (C-5).

was still insufficient, leaving guar gum remains in the system as acoupling layer on the mZVI. Quantification of the remaining guargum concentrations on the aquifer samples at the end of the exper-iment (Fig. 5) supports this hypothesis, as 89% of guar gum wasremoved for C-5 (GG-mZVI, abiotic control) as compared to 98%for C-3 (GG-mZVI). These data prove that the soil microbial com-munity degrades guar gum and therefore, indirectly, improves thereactivity of the GG-mZVI by enabling pollutants migration towardsthe mZVI [22]. However, the rinsing process is not fully effectiveunder simulated in situ conditions as it was in the batch exper-iments where even slightly faster disappearance rate constantswere observed.

3.3. Effect of mZVI and guar gum stabilized mZVI on the soilmicrobial population

The column test was also conducted to evaluate the effect ofmZVI and guar gum stabilized mZVI on the naturally present micro-bial population.

mZVI + GG + aquifermZVI + aquifer (C-2)(C-3)

mZVI + GG + a quife rAC (C-5)

ntrols were below method detection limit (MDL = 1 × 10−2 ng ATP mL−1) indicating

M. Velimirovic et al. / Journal of Hazardous Materials 265 (2014) 20– 29 27

1.00E+00

1.00E+01

1.00E+02

1.00E+03

1.00E+04

1.00E+05

1.00E+06

1.00E+07

1.00E+08

1.00E+09

copi

es /

g o

f soi

l

EUB

GEO

DSR

DHC

aquife r AC (C-4) aquifer (C-1) mZVI + aquifer (C-2) mZVI + GG + aquifer (C- 3) mZVI + GG + aqui fer AC (C-5 )

cter an

mpowatmti((b

atsbwpwdttacitciwoost

as

Fig. 7. Dynamic of the total bacterial community, Dehalococcoides (DHC), Geoba

ATP analysis was performed on effluent samples taken after 5onths of operation, to determine the activity of total bacterial

opulation under the different conditions (Fig. 6). In the effluentsf the abiotic controls (C-4 and C-5), the total ATP concentrationsere below the detection limit, showing the absence of bacterial

ctivity. Interestingly, similar microbial activity was observed forhe C-1 with aquifer material (1.66 × 10−1 ng mL−1) and C-2 with

ZVI (1.98 × 10−1 ng mL−1). These data imply that there is no nega-ive impact of the selected mZVI on total microbial activity presentn aquifer material. Finally, the highest concentration of total ATP4.86 × 100 ng mL−1) was observed in the effluent samples for C-3GG-mZVI), suggesting a positive impact of GG-mZVI on the micro-ial activity in the column.

Results of qPCR analysis targeting different bacterial groups inquifer samples collected from different horizontal column sec-ions are presented in Fig. 7. Eubacteria could be detected in allections of the aquifer column (C-1), as well as in columns amendedy mZVI (C-2) and GG-mZVI (C-3). Lower numbers of Eubacteriaere detected for both abiotic controls (C-4, C-5), as expected. Com-ared to the column (C-1), a clear increase in Eubacteria communityas observed in C-2, suggesting that mZVI after in situ applicationoes not negatively affect the total amount of bacteria on the longerm. Even a slightly higher number of Eubacteria was detected inhe downstream part C-3, confirming the conclusions based on ATPnalyses, namely a positive impact of guar gum on the bacterialommunity. Geobacter and Desulforeducers species were detectedn all aquifer samples with the highest values in C-2 and C-3, andhe lowest ones in the abiotic controls (C-4, C-5). These results areonsistent with the study of Gu et al. [45] where biomass increasesn and in the surroundings of the ZVI barrier were reported and

ere explained by microbial consumption of residual byproductsf the guar gum which was used during trench excavation. More-ver, Van Nooten et al. [46] reported that the presence of carbonources can make the aquifer material a favorable environment for

he growth of microorganisms.

Finally, Dehalococcoides spp. were found to be present in thequifer samples, but in very low numbers that are insufficient forupporting the biodegradation of high concentrations of CAHs. In

d Desulfo reducers on DNA level for aquifer samples collected in column study.

the effluent, the ethene and ethane concentrations were not highercompared to the influent (SI5). No significant biodegradation wasobserved in batch reactors with aquifer material collected from thesite either, nor in reactors supplied by guar gum as an extra electrondonors (results not shown.). Additionally, low pH values measuredin the influent samples might also be a reason for the low numberof Dehalococcoides detected in samples. Optimal pH range of nearneutral was reported for dechlorination activity of Dehalococcoides[47,48].

3.4. Impact of guar gum on the life-time of mZVI particles underin situ conditions

To evaluate the impact of guar gum on the iron consumptionand iron corrosion, both influencing the life-time of mZVI, theremaining concentrations of zerovalent iron in the column fillingswere determined after 6 months of continuous operation (Fig. 8).Before start-up, columns contained 13 ± 3 g of ZVI per kg of col-umn filling. After more than 6 months of operation the averageZVI concentration in C-2 (mZVI) was 11 ± 1 g of ZVI per kg of col-umn filling. When assuming a worst case initial iron concentration(C0 = 15.7 g ZVI/kg), on average, approximately 70% of ZVI remainedin the column suggesting that mZVI will possibly be reactive at least1.7 years. This is 3 times longer than initially presented for nZVIsused for in situ applications [30]. A significantly higher remainingconcentration of ZVI was detected in C-3 (GG-mZVI) compared toC-2 (mZVI), predicting a possible life-time of 11 years. The presenceof guar gum slightly reduced the reactivity of iron by inhibiting sur-face reaction at specific active sites [9], which slows down the ironcorrosion rate and prolongs the life-time of the reactive zone. Thereduced corrosion rate can be explained by the reduced accessi-bility of guar gum coupled mZVI particles. Finally, mass transferresistance, porosity loss and decreased access of contaminants to

iron particles due to the precipitation of insoluble Fe-oxides andFe-(oxi)-hydroxides formed on particle surface as suggested forgranular ZVIs in column systems [49] should be considered tofinally deduce on life-time of mZVIs.

28 M. Velimirovic et al. / Journal of Hazardous Materials 265 (2014) 20– 29

0

2

4

6

8

10

12

14

16

18

mZVI + aqu ife r mZVI + GG + aquifermZVI + aquifer (C-2)aquifer AC (C-4)aquifer (C-1) mZVI + GG + a quife r

g ZV

I / k

g of

colu

mn

fillin

g

Backgr oun dvalues

Backgroundvalues

n the c

4

ttfedcwciatwttol

aic

p

A

Usas

A

i2

[

[

[

[

[

[

[

[

[

(control sa mple)

Fig. 8. Average ZVI concentration i

. Conclusions and implications for RZ design

Results of our column study demonstrate that under conditionshat are closer to the real in situ environment, guar gum may affecthe performance of mZVI in the reactive zone, but in a slightly dif-erent way than was previously observed in our batch degradationxperiments [22]. The ability of the soil microbial population toegrade guar gum without addition of commercial enzymes, wasonfirmed. The reduced reactivity of GG-mZVI as compared to mZVIas also confirmed, but was found to be less pronounced under

ontinuous flow conditions. On the other hand, to gain full reactiv-ty recovery of guar gum stabilized mZVIs, more time was neededs the guar gum degradation by microorganisms requires someime and the removal of guar gum and its breakdown productsas slower under in situ conditions than predicted via the batch

est. The coupling of guar gum with mZVI which is responsible forhe latter finding, was hypothesized to be also responsible for thebserved reduced corrosion rate of GG-mZVI, and consequently theonger life-time of the particles.

Importantly, mZVI and GG-mZVI did not significantly affect thectivity and/or the amount of the total bacterial community, evennduced bacterial growth and activity was observed, which may beonsidered a strong point for in situ applications.

In conclusion, GG-mZVI particles were shown to have goodotential for the in situ treatment of CAHs.

cknowledgments

This research was conducted in the framework of the Europeannion project AQUAREHAB (FP7-Grant Agreement Nr. 226565). The

upport of Höganäs and Prof. Dr. Piet Seuntjens was appreciatednd we thank Dr. Pieter Jan Haest for his valuable comments to thistudy.

ppendix A. Supplementary data

Supplementary data associated with this article can be found,n the online version, at http://dx.doi.org/10.1016/j.jhazmat.013.11.020.

[

(C-3) AC (C-5)

olumns after column dismantling.

References

[1] R.W. Gillham, S.F. O’Hannesin, Metal-catalysed abiotic degradation of halo-genated organic compounds, in: Proceedings of IAH Conference Modern Trendsin Hydrogeology, Hamilton, Ontario, Canada, May 10–14, 1992.

[2] R.W. Gillham, S.F. O’Hannesin, Enhanced degradation of halogenated aliphaticsby zerovalent iron, Ground. Water. 32 (1994) 958–971.

[3] S.R. Day, S.F. O’Hannesin, L. Marsden, Geotechnical techniques for the construc-tion of reactive barriers, J. Hazard. Mater. 67 (1999) 285–297.

[4] F. He, D. Zhao, Manipulating the size and dispersibility of zerovalent ironnanoparticles by use of carboxymethyl cellulose stabilizers, Environ. Sci. Tech-nol. 41 (2007) 6216–6221.

[5] C. Noubactep, S. Caré, R.A. Crane, Nanoscale metallic iron for environmen-tal remediation: prospects and limitations, Water Air Soil Pollut. 223 (2012)1363–1382.

[6] B. Schrick, B.W. Hydutsky, J.L. Blough, T.E. Mallouk, Delivery vehicles for zerova-lent metal nanoparticles in soil and groundwater, Chem. Mater. 16 (2004)2187–2193.

[7] S.R. Kanel, D. Nepal, B. Manning, H. Choi, Transport of surface-modified ironnanoparticle in porous media and application to arsenic(III) remediation, J.Nanopart. Res. 9 (2007) 725–735.

[8] N. Saleh, K. Sirk, Y. Liu, T. Phenrat, B. Dufour, K. Matyjaszewski, R.D. Tilton, G.V.Lowry, Surface modification enhance nanoiron transport and NAPL targetingin saturated porous media, Environ. Eng. Sci. 24 (2007) 45–57.

[9] T. Phenrat, Y. Liu, R.D. Tilton, G.V. Lowry, Adsorbed polyelectrolyte coatingsdecrease Fe0 nanoparticle reactivity with TCE in water: conceptual model andmechanisms, Environ. Sci. Technol. 43 (2009) 1507–1514.

10] A. Tiraferri, R. Sethi, Enhanced transport of zerovalent iron nanoparticles insaturated porous media by guar gum, J. Nanopart. Res. 11 (2009) 635–645.

11] E.J. Reardon, Styles of corrosion and inorganic control on hydrogen pressurebuildup, Environ. Sci. Technol. 39 (2005) 7311–7317.

12] K.-F. Chen, S. Li, W.-X. Zhang, Renewable hydrogen generation by bimetalliczerovalent iron nanoparticles, Chem. Eng. J. 170 (2011) 562–567.

13] C. Lee, J.Y. Kim, W.I. Lee, K.L. Nelson, J. Yoon, D.L. Sedlak, Bactericidal effectof zerovalent iron nanoparticles on Escherichia coli, Environ. Sci. Technol. 42(2008) 4927–4933.

14] M. Elimelech, J. Gregory, X. Jia, R.A. Williams, Particle Deposition andAggregation: Measurement, Modeling and Simulation, first ed., Butterworth-Heinemann, Oxford, 1995.

15] E.D. Vecchia, M. Luna, R. Sethi, Transport in porous media of highly concentratediron micro- and nanoparticles in the presence of xanthan gum, Environ. Sci.Technol. 43 (2009) 8942–8947.

16] K.J. Cantrell, D.I. Kaplan, T.J. Gilmore, Injection of colloidal Fe-0 particles in sandwith shear-thinning fluids, J. Environ. Eng. ASCE 123 (1997) 786–791.

17] M. Oostrom, T.W. Wietsma, M.A. Covert, V.R. Vermeul, Zero-valent ironemplacement in permeable porous media using polymer additions, GroundWater. Monit. R. 27 (2007) 122–130.

18] M.J. Truex, V.R. Vermeul, D.P. Mendoza, B.G. Fritz, R.D. Mackley, M. Oostrom,

T.W. Wietsma, T.W. Macbeth, Injection of zero-valent iron into an unconfinedaquifer using shear-thinning fluids, Ground Water. Monit. R. 31 (2011) 50–58.

19] D. Xue, R. Sethi, Viscoelastic gels of guar and xanthan gum mixtures providelong-term stabilization of iron micro- and nanoparticles, J. Nanopart. Res.(2012), DOI: 10.1007/s11051-012-1239-0.

Hazard

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

(2001) 1474–1481.

M. Velimirovic et al. / Journal of

20] S. Comba, J. Braunn, An empirical model to predict the distribution of ironmicro-particles around an injection well in a sandy aquifer, J. Contam. Hydrol.132 (2012) 1–11.

21] D. Navon, R.C. Loehr, H.M. Liljestrand, D.E. Daniel, Impact of biodegradabletrenching slurry on iron treatment wall performance, Water Sci. Technol. 38(1998) 49–53.

22] M. Velimirovic, H. Chen, Q. Simons, L. Bastiaens, Reactivity recovery of guargum coupled mZVI by means of enzymatic breakdown and rinsing, J. Contam.Hydrol. 142–143 (2012) 1–10.

23] M. Velimirovic, P.-O. Larsson, Q. Simons, L. Bastiaens, Reactivity screeningof microscale zerovalent irons and iron sulfides towards different CAHsunder standardized experimental conditions, J. Hazard. Mater. 252–253 (2013)204–212.

24] T. Van Nooten, L. Diels, L. Bastiaens, Design of a multifunctional permeablereactive barrier for the treatment of landfill leachate contamination: laboratorycolumn evaluation, Environ. Sci. Technol. 42 (2008) 8890–8895.

25] H.S.C. Eydal, K. Pedersen, Use of an ATP assay to determine viable microbialbiomass in Fennoscandian Shield groundwater from depths of 3–1000 m, J.Microbiol. Methods 70 (2007) 363–373.

26] F. Hammes, F. Goldschmidt, M. Vital, Y. Wang, T. Egli, Measurement and inter-pretation of microbial adenosine triphosphate (ATP) in aquatic environments,Water Res. 44 (2010) 3915–3923.

27] M. Vital, M. Dignum, A. Magic-Knezev, P. Ross, L. Rietveld, F. Hammes, Flowcytometry and adenosine tri-phosphate analysis: Alternative possibilities toevaluate major bacteriological changes in drinking water treatment and distri-bution systems, Water Res. 44 (2012) 4665–4676.

28] M. Dubois, K.A. Gilles, J.K. Hamilton, P.A. Rebers, F. Smith, Colorimetric methodfor determination of sugars and related substances, Anal. Chem. 28 (1956)350–356.

29] Y. Liu, H. Choi, D. Dionysiou, G.V. Lowry, Trichloroethene hydrodechlorinationin water by highly disordered monometallic nanoiron, Chem. Mater. 17 (2005)5315–5322.

30] Y. Liu, G.V. Lowry, Effect of particle age (Fe0 Content) and solution pH on NZVIreactivity H2 evolution and TCE dechlorination, Environ. Sci. Technol. 40 (2006)6085–6090.

31] E.J. Reardon, Anaerobic corrosion of granular iron: Measurement and interpre-tation of hydrogen evolution rates, Environ. Sci. Technol. 29 (1995) 2936–2945.

32] B. Hendrickx, W. Dejonghe, W. Boënne, M. Brennerova, M. Cernik, T. Led-erer, M. Bucheli-Witschel, L. Bastiaens, W. Verstraete, E.M. Top, L. Diels, D.Springael, Dynamics of an oligotrophic bacterial aquifer community duringcontact with a groundwater plume contaminated with benzene, toluene, ethyl-benzene, and xylenes: An in situ mesocosm study, Appl. Environ. Microbiol. 71(2005) 3815–3825.

33] G. Muyzer, E.C. De Waal, A. Uitterlinden, Profiling of complex microbial popu-lations using denaturing gradient gel electrophoresis analysis of polymerasechain reaction-amplified genes coding for 16S rRNA, Appl. Environ. Microbiol.

59 (1993) 695–700.

34] T.H.M. Smits, C. Devenoges, K. Szynalski, J. Maillard, C. Holliger, Developmentof a real-time PCR method for quantification of the three genera Dehalobacter,Dehalococcoides, and Desulfitobacterium in microbial communities, J. Micro-biol. Methods. 57 (2004) 369–378.

[

ous Materials 265 (2014) 20– 29 29

35] J. He, K.M. Ritalahti, M.R. Aiello, F.E. Löffler, Complete detoxification of vinylchloride by an anaerobic enrichment culture and identification of the reduc-tively dechlorinating population as a Dehalococcoides species, Appl. Environ.Microbiol. 69 (2003) 996–1003.

36] J. Geets, K. Vanbroekhoven, B. Borremans, J. Vangronsveld, D. van der Lelie,L. Diels, Molecular monitoring of SRB community structure and dynamics inbatch experiments to examine the applicability of in situ precipitation of heavymetals for groundwater remediation, J. Soils. Sediments 5 (2005) 149–163.

37] D.E. Holmes, K.T. Finneran, R.A. O’Neil, D.R. Lovley, Enrichment of members ofthe family Geobacteraceae associated with stimulation of dissimilatory metalreduction in uranium contaminated aquifer sediments, Appl. Environ. Micro-biol. 68 (2002) 2300–2306.

38] D. Schäfer, R. Köber, A. Dahmke, Competing, TCE and cis-DCE degradationkinetics by zero-valent iron-experimental results and numerical simulation,J. Contam. Hydrol. 65 (2003) 183–202.

39] J.P. Fennelly, A.L. Roberts, Reaction of 1,1,1-trichloroethane with zero-valentmetals and bimetallic reductants, Environ. Sci. Technol. 32 (1998) 1980–1988.

40] Z. Shi, J.T. Nurmi, P.G. Tratnyek, Effects of nano zero-valent iron on oxidation-reduction potential, Environ. Sci. Technol. 45 (2011) 1586–1592.

41] M.S. Odziemkowski, T.T. Schuhmacher, R.W. Gillham, E.J. Reardon, Mechanismof oxide film formation on iron in simulating groundwater solutions: Ramanspectroscopic studies, Corros. Sci. 40 (1998) 371–389.

42] P.G. Tratnyek, M.M. Scherer, B. Deng, S. Hu, Effects of natural organic matter,anthropogenic surfactants, and model quinones on the reduction of contami-nants by zero-valent iron, Water Res. 35 (2001) 4435–4443.

43] W. Lee, B. Batchelor, Abiotic reductive dechlorination of chlorinated ethylenesby iron-bearing soil minerals. 2. Green rust, Environ. Sci. Technol. 36 (2002)5348–5354.

44] E.C. Butler, K.F. Hayes, Kinetics of the transformation of trichloroethyleneand tetrachloroethylene by iron sulfide, Environ. Sci. Technol. 33 (1999)2021–2027.

45] B. Gu, D.B. Watson, L. Wu, D.H. Phillips, D.C. White, J. Zhou, Microbial charac-teristics in a zero-valent iron reactive barrier, Environ. Monit. Assess. 77 (2002)293–309.

46] T. Van Nooten, F. Lieben, J. Dries, E. Pirard, D. Springael, L. Bastiaens, Impact ofmicrobial activities on the mineralogy and performance of column-scale per-meable reactive iron barriers operated under two different redox conditions,Environ. Sci. Technol. 41 (2007) 5724–5730.

47] J. Gerritse, O. Drzyzga, G. Kloetstra, M. Keijmel, L.P. Wiersum, R. Hutson, M.D.Collins, J.S. Gottschal, Influence of different electron donors and acceptors ondehalorespiration of tetrachloroethene by Desulfitobacterium frappieri TCE1,Appl. Environ. Microbiol. 65 (1999) 5212–5221.

48] A. Suyama, M. Yamashita, S. Yoshino, K. Furukawa, Isolation and characteriza-tion of Desulfitobacterium sp. strain Y51 capable of efficient dehalogenationof tetrachlorethene and polychloroethanes, Biosci. Biotechnol. Biochem. 65

49] J. Klausen, P.J. Vikesland, T. Kohn, D.R. Burris, W.P. Ball, A.L. Roberts, Longevityof granular iron in groundwater treatment processes: solution compositioneffects on reduction of organohalides and nitroaromatic compounds, Environ.Sci. Technol. 37 (2003) 1208–1218.