Published in Journal of Contaminant Hydrology 192, 1-19, 2016 which should be used for any reference to this work

1

Identification of abiotic and biotic reductive dechlorination in a

by

chlorinated ethene plume after thermal source remediation

means of isotopic and molecular biology tools

Alice Badin a, Mette M. Broholm b, Carsten S. Jacobsen c, Jordi Palau a,

d a,

Philip Dennis , Daniel Hunkeler ⁎

a University of Neuchâtel, Centre for Hydrogeology & Geothermics (CHYN), Rue Emile Argand 11, CH 2000 Neuchâtel, Switzerlandb Technical University of Denmark (DTU), Department of Environmental Engineering, Miljøvej, DTU B113, DK 2800 Kgs. Lyngby, Denmarkc Geological Survey of Denmark and Greenland (GEUS), Department of Geochemistry, Ø. Voldgade 10, 1350 København K, Denmarkd SiREM, 130 Research Lane, Guelph, Ontario, N1G5G3, Canada

a b s t r a c t

to theduced

⁎ Corresponding author.

Thermal tetrachloroethene (PCE) remediation by steam injection in a sandy aquifer ledrelease of dissolved organic carbon (DOC) from aquifer sediments resulting in more re

. Thesed eightdationuctivesource

Keywords:

Chloroethenes

biotic(Dhc)gethert cDCEon (II)portedshowat thisnmoreseries

.pecificemical

redox conditions, accelerated PCE biodegradation, and changes in microbial populationschanges were documented by comparing data collected prior to the remediation event anyears later. Based on the premise that dual C-Cl isotope slopes reflect ongoing degrapathways, the slopes associated with PCE and TCE suggest the predominance of biotic reddechlorination near the source area. PCEwas the predominant chlorinated ethene near thearea prior to thermal treatment. After thermal treatment, cDCE became predominant. Thecontribution to these changes was supported by the presence of Dehalococcoides sp. DNAand Dhc targeted rRNA close to the source area. In contrast, dual C-Cl isotope analysis towith the almost absent VC 13C depletion in comparison to cDCE 13C depletion suggested thawas subject to abiotic degradation due to the presence of pyrite, possible surface-bound iror reduced iron sulphides in the downgradient part of the plume. This interpretation is supby the relative lack of Dhc in the downgradient part of the plume. The results of this studythat thermal remediation can enhance the biodegradation of chlorinated ethenes, and theffect can be traced to themobilisation of DOC due to steam injection. This, in turn, results ireduced redox conditions which favor active reductive dechlorination and/or may lead to aof redox reactions which may consecutively trigger biotically induced abiotic degradationFinally, this study illustrates the valuable complementary application of compound-sisotopic analysis combined with molecular biology tools to evaluate which biogeochprocesses are taking place in an aquifer contaminated with chlorinated ethenes.

Stable isotopesMolecular biologyThermal treatment

in

1. Introduction particularly adapted for source treatment

ethlopedtion i

subsurface sediments of relatively high permeability ).This rganic h may cesses. active

Management of sites contaminated with chlorinatedenes is known to be challenging. Among the various deveremediation methods, thermal treatment by steam injec

luence

-

s

such as sandy aquifers (von Schnakenburg, 2013remediation strategy is known to release dissolved ocarbon (DOC) (Friis et al., 2005) the increase of whictrigger a chain of microbially-mediated redox proWhen natural attenuation has been observed prior to source remediation, steam injection might thus infthe naturally occurring degradation.

http://crossmark.crossref.org/dialog/?doi=10.1016/j.jconhyd.2016.05.003&domain=pdf

Natural degradation of chlorinated ethenes might occur biotically due to the presence of adequate active microorgan-

in thbiotr conyieldvin

roces et anterehene coulerobnatiouch ang tapell. ThH2), factogy, thion oaul eurthenatingenereriumtalyzatiotia arnta e2004ted tne arr et ain thplac

19852000n-isme eveionaluctivat thresenFeIIS2 irooduchere ted t lesoundki aninatete wah cDCbounion ok anvity o redometa

turn affect the likelihood that biotically induced abiotic degradation will take place (Tobiszewski and Namieśnik,

s in the or site

nalysis ethene

extent isotopic it was elp to ic from ; Audí-., 2014; rmined rinated t years, slopes ntifica-etnik et n et al., reover, ities for t in the mRNA aracter-identify (Novais ed that od that tion in CE and lorinat-een for s been class of lorina-

encode lyze VC ' mRNA support

genes l genes ., 2004; es from essfully 08; van oreover volved Dhc, as A genes able to , 2010). lemen-y of the

ess the on a biogeo-. More

2

2

isms in specific redox conditions, as well as abioticallypresence of reduced iron (Fe) minerals. Sequential reductive dechlorination of the ubiquitous groundwatetaminant tetrachloroethene (PCE) consecutively trichlo-roethene (TCE), cis-dichloroethene (cDCE), chloride (VC) and eventually non-toxic ethene. This ptakes place in strictly anaerobic systems (Wiedemeier1999; Bradley, 2000) and is the most commonly encounaturally occurring biotic degradation of chlorinated etAccording to laboratory and field observations, PCEundergo reductive dechlorination in virtually all anaconditions while reductive TCE, cDCE and VC dechloriwould generally occur in more reduced conditions, siron-reducing for TCE and ideally sulfate-reducimethanogenic for cDCE and VC (Vogel et al., 1987; Ch1996; Bradley, 2000; Tiehm and Schmidt, 2011)presence of and competition for molecular hydrogen (key electron donor, can also be a determining (Ballapragada et al., 1997). Depending on its mineralopresence of iron may also induce competitive inhibitchlorinated ethene biotic reductive dechlorination (Pal., 2013). The occurrence of reductive dechlorination fdepends on the presence and activity of specific dechlorimicroorganisms. Mem-bers from various bacterial such as Sulfurospirillum, Dehalobacter, DesulfitobactDesulfuromonas or Geobacter have been reported to casome steps of chlorinated ethene reductive dechlorinHowever, while some enrich-ment cultures and consorable to dechlorinate to ethene (Flynn et al., 2000; Auleal., 2002; Duhamel et al., 2002; Hoelen and Reinhard, to date, the only organisms which have been reporcatalyze complete reductive dechlori-nation to ethesome species of the genus Dehalococcoides (Dhc) (Löffle2013). cDCE is hence often found to accumulate subsurface. Microbial oxidation might also take particularly in the case of cDCE and VC (Hartmans et al.,Bradley and Chapelle, 1998, Bradley and Chapelle, Despite their presence in the subsurface, microorgamay display a low activity, are sometimes inactive or ardormant (Meckenstock et al., 2015), which may addi-thinder reductive dechlorination. Abiotic reddechlorination can also take place naturally, provided thadequate minerals and geochemical conditions are pIron sulphides such as mackinawite (FeIIS) or pyrite (iron oxides such as magnetite (FeIIO·FeIII2O3), andhydroxides such as green rusts, which are corrosion prof iron or steel ([FeII6-xFexIII(OH)12)]X+[(A)x/nyH2O]X− wis an anion, typically SO4− or Cl−), have been reporcatalyze abiotic reductive degradation yieldingchlorinated com-pounds and other non-toxic compsuch as acetylene in various proportions (TobiszewsNamieśnik, 2012). Pyrite is known to reduce all chlorethenes (Lee and Batchelor, 2002a) while mackinawishown to reduce PCE and TCE but was non-reactive wit(Butler and Hayes, 1999, Jeong et al., 2011). Surface-Fe(II) is also known to catalyze abiotic degradatreducible contaminants (Elsner et al., 2004; McCormicAdriaens, 2004, Han et al., 2012). Furthermore, the activarious bacteria in the subsurface may change the localconditions. This might affect the redox potential of contained in minerals, which might in

e ic -s yl s l., d s. d ic n s o e, e a r e f t r g a , e n. e t ), o e l., e e, ; ). s n ly e e t. ), n ts A o s s d d s E d f d f x ls

2012).In order to explore the occurrence of such processe

subsurface and thus evaluate the effect of remediationmanagement, various tools may be employed.

In recent decades, compound specific isotopic ahas been increasingly used to explore chlorinated degradation processes. It was demonstrated that theof biodegradation could be determined based on measurements (Hunkeler et al., 2010). Additionally,suggested that dual C\\Cl isotopic analysis may hdifferentiate degradation pathways, for example biotabiotic degradation (Elsner et al., 2005; Abe et al., 2009Miró et al., 2013) despite some limitations (Badin et alRenpenning et al., 2014). The range of laboratory detedual C\\Cl isotope slopes associated with various chloethene degradation processes has increased in recenthus enriching the database to which dual C\\Cl isotopemeasured in the field can be compared for process idetion (Fig. 1) (Abe et al., 2009; Audí-Miró et al., 2013; Cral., 2013; Kuder et al., 2013; Wiegert et al., 2013; Badi2014; Cretnik et al., 2014; Renpenning et al., 2014). Morapid advances in molecular biology open new possibilthe characterization of microbial communities presensubsurface and the assessment of their activity based onanalysis. For example bacterial 16S rRNA gene pool chization via amplicon pyrosequencing can be used to the present bacterial communities in high detail and Thorstenson, 2011). It was moreover demonstratpyrotag sequencing is a robust and reproducible methcan be used for reliable microbial community exploranatural systems (Pilloni et al., 2012). Additionally, as cDVC degradation usually represents the bottle neck of ched ethene natural attenuation, it is essential to scrmarkers of their degradation. Dhc screening has thucarried out in numerous studies since it is the only microorganisms reported to perform cDCE and VC dechtion. Furthermore, assessing the presence of genes thatfor VC reductive dehalogenases (rdhA) known to catareduction to ethene, such as vcrA, and measuring geneslevel constitutes a stronger line of evidence to complete reductive dechlorination. The vcrA and bvcAidentified in Dhc are so far the only two functionadescribed to encode VC rdhA (Krajmalnik-Brown et alMüller et al., 2004) and their presence in field samplsites contaminated with chlorinated ethenes was succrelated to complete dechlorination (Scheutz et al., 20der Zaan et al., 2010; Damgaard et al., 2013a). It was mshown based on field samples that rdhA genes directly inin dechlorination should be targeted in addition to different microbial species might harbour vcrA and bvcdue to horizontal gene transfer and are therefore alsodechlorinate VC down to ethene (van der Zaan et al.Finally, targeting mRNA constitutes an essential comptary analysis as it will additionally inform on the activitcorresponding degrader (Bælum et al., 2013).

Here we combine such innovative methods to assimpact of source remediation by steam injectionchlorinated ethene plume that occurrs in a complex chemical system where iron minerals are present

specifically, the aim was to evaluate if a plume detachment occurred due to steam injection as reported by a

ion, erm

e fac anarisosourcut tredootop DN genom 2e flow

et awn ol PCanin

operated between 1964 and 2001. A sandy aquifer N50 m thick occasionally containing less permeable silt and clay

site me of Fig. 2) 1 km l cover l lines is less rganic quifers verage viously y vary sulting

to the entire some cted by viously nsist of nd are rposes. n were

Fig. 1. Literature values for dual C\\Cl isotope slopes associated with various degradation pathways of chlorinated ethenes (Abe et al., 2009; Audí-Miró et al., 2013; Cretnik et al., 2013; Kuder et al., 2013; Badin et al., 2014; Cretnik et al., 2014; Renpenning et al., 2014). Slopes are given for each of the substrates. Green arrows represent dihaloelimination, i.e. the main pathway for abiotic reduction, which typically occurrs in presence of zero-valent Fe. Blue arrows represent aerobic oxidation of chlorinated ethenes. Red arrows correspond to hydrogenolysis occurring during reductive dechlorination mediated by bacteria via their specific corrinoid-containing reductive dehalogenase enzymes. Dual C\\Cl isotope slopes associated with hydrogenolysis by these enzymes and corrinoids as well as by chemical models mimicking corrinoids (cobalamin, cobaloxime) were also recently reported (Renpenning et al., 2014). However, due to the lack of correlations between these slopes and slopes associated with bacterially mediated dechlorination, these values are not reported here.

3

previous study (Sleep and Ma, 1997), and in additnatural degradation was stimulated by the thtreatment.

A major advantage of this field site resides in ththat the plume was formerly well characterizedstudied (Hunkeler et al., 2011), which allows for compbetween before and after (7–8 years later) remediation. An extensive campaign was carried oevaluate the impact of steam injection where parameters, chlorinated ethene concentrations and iscompositions thereof, 454 pyrotag sequencing, Dhcand rRNA, bvcA and vcrA functional genes, andtranscripts (mRNA) were analysed in samples taken frwells at different screening depths along the plumline.

2. Materials and methods

2.1. Study site description

The studied site previously described by Hunkeleris located in southern Jutland, Denmark, in the toRødekro (Hunkeler et al., 2011). Briefly, the originagroundwater contamination comes from a dry-cle

facility which

if al

t d n e o x ic A e 0

l. f E g

lenses was characterized based on formercharacterization campaigns. A chlorinated ethene plu~2 km length that follows the groundwater flow (southward from the site and turns southeast after was identified based on extensive monitoring wel(55 multilevels). The differ-ence in equipotentiabetween 2006 and 2014 indicates that the gradientsteep in 2014 than in 2006 (Fig. 2). Low amounts of omatter as well as high levels of iron are expected in ain this part of Jutland (Postma et al., 1991). An agroundwater velocity of 0.24 m·day-1 was preestimated (Hunkeler et al., 2011), though this malocally due to different hydraulic conductivities refrom the large grain size distribution.

Thermal remediation by steam injection was appliedsource zone between October and December 2006. Theplume was sampled in 2006 before remediation andpoints further out in the plume that were not yet imparemediation were sampled in 2007. The data prediscussed by Hunkeler et al. (Hunkeler et al., 2011) codata collected during these two sampling campaigns areferred to as data from 2006, for simplification puThe main conclusions drawn from the previous campaigthat

(i) PCE and TCE were likely biotically degraded by reductivedechlorination in the first 400 m downgradient of the source,

ic noadien0 anbiotiurthecDCC\\Cion ocoul

Concentrations of redox species and chlorinated thenes were regularly measured between 2004 and 2014 s part

pecies pling

source 50 m)

city, it ported 6 and h an ly as

Fig. 2. Groundwater equipotential lines from 2006 (blue) and 2014 (red), approximate plume flow line (blue dashed arrow) and monitoring wells (grey an orangetargets). Wells sampled in 2014 appear in orange.

4

(ii) cDCE was not affected by degradation (neither biotbiotically induced abiotic) in the first 1050 m downgrfrom the source, but it was degraded between 1051900 m downgradient, likely at least partially byreductive dechlorination, and (iii) VC was transformed fby undetermined processes. The process responsible fordegradation remained uncertain due to the lack in dualstudies associated with abiotic reductive dechlorinatcDCE to which the dual C\\Cl slope observed in Rødekrobe compared.

ea

d

rtdcrElfd

of the monitoring process. The evolution of redox sbetween 2004 and 2014 is given in Fig. 3 B for sampoints from the plume centerline located close to the (B16-1; 100 m), in the middle of the plume (B34-4; 10and at the front of the plume (B61-2; 1700 m).

Based on the assessed average groundwater velocan be estimated that species might have been transabout 600 m downgradient between the end of 200the new campaign carried out in 2014. Sucestimation should however be treated cautiousheterogeneities might change

2

Fig. 3. A: Redox sensitive species concentrations in 2006 (Hunkeler et al., 2011) and 2014 (this study). B: Evolution of redox species concentrations between 2004 and 2014 along the plume centre line at 100 m (well B16–1), 1050 m (B34–4), and 1700 m (B61–2) downgradient from the source. These wells are “circled” in A. Dissolved oxygen (DO), nitrate (NO3−), Fe(II), dissolved organic carbon (DOC) and methane (CH4) concentrations are given by the left y-axis while sulfate (SO4−) concentrations are given by the right y-axis. DOC concentrations were measured from the time of source remediation in B16-1 and only in 2014 in B34-4 and B61-2. The red dashed line corresponds to the source remediation event. Note that the left graph has a different left y-axis than the other two graphs.

5

the hydraulic conductivity by a few orders of magnitude in areas where different subsurface materials such as gravel/

urtherce, tergen deep

ns wume ssolv3−) abottlltratiamp

in ha spikamp 40 mted cles 40 mith lodspaO3 wh pHctivie weth 3 md wt no les. ith orederati0 mLillapoSIREples A), ascribquid (GEU

weANAry Aenma−1, 0etheap GtectiEthenat tnmarimadck B/

1500) with detection limits of 0.43 μg·L−1, 0.94 μg·L−1, 0.47 μg·L−1 and 0.38 μg·L−1, respectively.

at the system rinated 2014), B-VRX

by N2(VICI), 120 °

he gas m, 10

(35 °C .1 min t for 5 ustion ™ 100 the C andard ured in e 0.6‰ 26) for rtainty to ISO , 0.7‰ s that ratios ded in for VC

ethene anually

a gas-, 2012) r. The min−1

solved Henry ter) for

ed as d to an (Santa m, 0.25 of 1.2 ration. carried ., 2010. 0.3 ‰ ‰ and rmerly at the nce to SMOC) Isotope ethod inuous

e 454 h, ON, d and moved g DNA

6

2

2

coarse sand or fine sand/silt/clay lenses are present. Fmore, with increasing distance and depth from the souclay lenses progressively disappear. This results in a divof the flow direction which causes the plume to diveinto the aquifer.

2.2. Groundwater sampling

Groundwater in 42 screens from 20 different locatiosampled in May 2014 after purging 3 times the volwells and checking the stability of pH, temperature, dioxygen (O2) and conductivity. Samples for nitrate (NOsulfate (SO4−) analysis were collected in hard plastic samples for DOC were collected in glass bottles after fiand spiked with H3PO4 upon arrival in the laboratory; sfor dissolved iron (Fe(II)) concentration were sampledplastic bottles after filtration (sterile filter, 0.45 μm) andwith HNO3 (to pH 2) upon arrival in the laboratory; sfor chlorinated ethene concentration were collected inglass vials closed without headspace with a Teflon-coaand analysed upon arrival in the laboratory. Sampchlorinated ethene isotope analysis were collected inglass vials or 1 L Schott bottles (for isotope analysis wchlorinated ethene concentrations) closed without heawith a cap containing a Teflon-coated septum. HNadded previ-ously to the containers in order to reacwhen filled with the sample and to stop any microbial aSamples for methane, acetylene, ethene and ethancollected in 6 mL Exetainer® glass vials (LabCo, UK) wiheadspace which were previously evacuated and fille100 μL concentrated H2SO4. Caution was taken so thabubbles were injected with the groundwater sampaqueous samples were stored in ice boxes topped wpacks until arrival at the laboratory where they were st4 °C until analysis. Samples for 454 pyrotag next gensequencing analysis were collected by passing 300–40groundwater through Sterivex™ Filters (EMD MBillerica, MA, USA) and were then shipped to Guelph (commercial laboratory, Canada) with ice packs. SamDhc DNA and rRNA, bvcA and vcrA functional genes (DNgene transcript (mRNA) analysis were collected as depreviously in Bælum et al., 2013 snap-shot frozen in liand kept at −80 °C until analysis in Copenhagen Denmark).

2.3. Analyses

Chemical analyses: NO3−, Fe(II), SO4− and DOCmeasured by accredited methods by the accredited (DInternational Standards Organization 17,025) laborato(ALS Denmark A/S, alsglobal.dk, alsglobal.com) in Dwith quantification limits of 0.03 mg·L−1, 0.01 mg·Lmg·L−1 and 0.1 mg·L−1 respectively. Chlorinated concentrations were analysed at ALS by Purge & TrChromatograph-Mass Spectrometer (GC–MS) with a delimit of 0.02 μg·L−1 and standard deviations of 10%. methane, ethane and acetylene were determined Technical University of Denmark (DTU, Kgs. Lyngby, Deby headspace GC-Flame Ionisation Detector (FID) (ShGC.14 A with a packed column with 80/120 CarbopaSP-

r-he ce er

as of ed nd es; on les rd ed les L

ap for L w ce as 2 ty. re L

ith air All ice at on of re, M, for nd ed N2 S,

re K, LS rk .5 ne as on e, he k) zu 3%

Isotopic analysis: C isotopic analysis was performedUniversity of Neuchâtel (CHYN, Switzerland) by the previously described for samples containing chloethene concentrations exceeding 5 μg·L−1 (Badin et al.,except that a QS-PLOT column was used instead of a Dto improve VC separation. The compounds degassedpurging were retained on a Vocarb 3000 trap transferred to a cryogenic trap (Tekmar Dohrmann) at −C to enable compound concentration, and sent to tchromatograph (GC) column (QS-PLOT, 30 m, 0.32 mμm) of an Agilent™ 7890a GC for compound separationfor 6 min, ramp of 15 °C·min−1 until 130 °C kept for 0followed by a ramp of 20 °C·min−1 until 240 °C kepmin). After combustion via an Isoprime GC5 combinterface, the resulting CO2 gas was sent to an Isoprimeisotope ratio mass spectrometer (IRMS) to measureisotope ratio. Samples were measured in duplicate. Stdeviations σ of the in-house reference materials measthe same sequences as samples from the field site wer(n = 32), 0.3‰ (n = 24), 0.5‰ (n = 24), 1.0‰ (n = PCE, TCE, cDCE, and VC, respectively. The standard unceof duplicate measurements was deter-mined accordingguidelines (BIPM, 1993) as σ/√2, i.e. 0.4‰, 0.2‰, 0.3‰for PCE, TCE, cDCE, and VC, respective-ly. Samplecontained reference compounds with known isotope(Elemental Analyser-IRMS measurement) were inclueach sequence to verify the method accuracy except which was not characterized by EA.

For the samples showing chlorinated concentrations below 5 μg·L−1, 1 L bottles were mconnected to a purge system that consisted of a frit fromwashing bottle as described formerly (Hunkeler et al.instead of passing 40 mL glass vials by autosamplebottles were purged for 30 min at a rate of 150 mL·which led to a removal of 90, 75, 50 and 100% of the disPCE, TCE, cDCE and VC, respectively, considering coefficients at 20 °C of 0.533, 0.314, 0.14, 0.891 (gas/waPCE, TCE, cDCE and VC, respectively.

Cl isotopic analysis of PCE and TCE was performpreviously described with an Agilent 7890 GC coupleAgilent 5975C quadrupole mass selective detector Clara, CA, USA) (Badin et al., 2014). A DB-5 column (30 mm, 0.25 μm, Agilent) with a constant helium flowmL·min−1 was used to perform chromatographic sepaMolecular ions were targeted and calculations were out according to the method developed by Aeppli et alCalibration with two external standards (δ37ClEIL1 = +and δ37ClEIL2 =−2.5‰ for PCE and δ37ClEIL1 = +3.05δ37ClEIL2 =−2.70‰ for TCE) which were focharacterized by the Holt method (Holt et al., 1997)University of Waterloo was completed for each sequeobtain δ values on the Standard Mean Ocean Chloride (scale. Cl isotopic analysis of cDCE was performed at Tracer Technologies Inc. (Waterloo) according to the mdeveloped by Shouakar-Stash et al., 2006, using a ContFlow (CF) IRMS.

Molecular biology analysis: DNA extraction for thpyrotag analysis was carried out at SiREM (GuelpCanada) as follows: Sterivex™ filters were openethe filter membrane with attached biomass was reand placed into the Bead Solution of a PowerMaIsolation Kit

http://alsglobal.com

(MoBio, Carlsbad, CA, USA) and pulverized using a sterile pipet tip. Cell lyses were performed using a MiniBeadbeater-8

f thing ) anoDro) an wer6f (5G GGrimeaniumowin, 52 °0 minit (Lif witRochersitysis o et ared btus.preseny. Therencimeret anomscripang eed awas a

crooring i wernt (i.d tha slighnominateing inismera ouctivand/oitionucin

counum ople t eacta arortin

vc

RNA RNanic prioliquirk) t

clay particles. Extracted DNA and RNA was purified usingNucleoSpin RNA Clean-up XS kit (Macherey-Nagel, Duren,

cDNAlimit

in the

rly toountne oration,point

� δ13CVC

entra-13CPCE, otopic prop-

ertain lated:

initialinatednds incDCE

otopicrada-tion isand

re ased as

c and with ion on m 27 plume

red in . 3A. In g·L−1

wn to mg·L

Fe(II)

7

(Biospec Products Bartlesville, OK, USA) at 50% omaximum setting for 30 s. DNA was purified usKingFisher™Duo (ThermoFisher Waltham, MA, USAeluted in 150 μL. DNA was quantified using a Nanspectrophotometer (NanoDrop Inc. Wilmington, DEstored at −80 °C after extraction. 16S rRNA genesamplified from DNA extracts with universal primers 92AAA CTY AAA KGA ATT GAC GG-3′) and 1392r (5′-ACGGT GTG TRC-3′) for 454 pyrotag analysis. The reverse palso contained a 10 nucleotide barcode and 454 FLX TitLib-L ‘B’ adapter. PCR was performed under the follconditions: 94 °C for 3 min; 25 cycles of 94 °C for 30 sfor 30 s, and 72 °C for 1 min; and finally 72 °C for 1Amplicons were purified with GeneJET PCR Purification KTechnologies, Burlington, ON, Canada) and sequencedRoche GS-FLX Titanium series kits and system (Branford, CT, USA) at Genome Quebec and McGill UnivInnovation Centre (Montreal, PQ, Canada). Finally, analthe reads was performed using QIIME v.1.8 (Caporaso2010). Initially, raw reads were demultiplexed and filtequality (NQ20) and length (N250 nt) using the pick_oscript with usearch61 (Edgar et al., 2011) option and reptative sequences were selected using the pick_rep_set.psequences were aligned to the Greengenes Core refalignment by PyNAST (Caporaso et al., 2010). Putative chsequences were removed using ChimeraSlayer (Haas 2011). Taxonomic assignment of the operational taxounits (OTU) was performed by assign_taxonomy.py with the Ribosomal Database Project (RDP) method (Wal., 2007; Martins et al., 2013). A sequence was definbelonging to a particular OTU when the similarity level least 97%.

In order to evaluate which and to which extent miganisms relevant to redox processes potentially occurrthe subsurface were present, OTU reads per sampletransformed to cells·L−1 based on the total bacteria coutotal DNA extracted from the samples). It was assumeall the extracted DNA was prokaryotic, which leads to aoverestimation, and that the average microbial gecontains 4 ⋅10−6 ng DNA ⋅cell−1 (Paul, 1996). Since chlorethene degradation as well as redox processes occurrthe subsurface were of interest, detected microorgawere grouped in taxonomic categories such as genwhich some strains are known to perform complete reddechlorina-tion, partial reductive dechlorination of PCE TCE, oxidation of cDCE and or VC under aerobic condbacteria reported to be found in iron- and sulfate-redconditions as well as during pyrite oxidation. Bacteria were then summed in each group and divided by the sbacteria counts of all targeted groups in each samevaluate the proportion of each bacteria group withinsample relative to the groups of interest. These dasummarised in Table S 1 and Table S 2 of the SuppInformation (SI).

Dehalococcoides DNA and rRNA, bvcA andfunctional genes (DNA) and genes transcripts (manalysis was performed on co-extracted DNA andusing a combined phenol-chloroform and mechbeadbeating method (Bælum et al., 2013). In brief:to cell lysis, the samples were mixed with 0.5 mL G2 DNA/RNA enhancer (Ampliqon, Odense, Denma

cover binding sites of the

e a d p d e ′-C r

g C . e h e, y f l., y y -e e ic l., ic t t s t

-n e e. t t e d n s f e r s, g ts f o h e g

rA

) A al r d o

Germany). RNA was converted to cDNA and DNA andPCR amplified using standard protocols with a detectionof 104 copies · L−1. The detailed protocols can be foundSI.

2.4. Calculations for isotopic data interpretation

2.4.1. C isotope balanceIn order to evaluate isotopic data and more particula

determine whether degradation released a significant amof compounds which were not detected, such as etheethane during complete sequential reductive dechlorinthe C isotope balance was determined for each samplingaccording to:

δ13Csum ¼PCE½ � � δ13CPCE þ TCE½ � � δ13CTCE þ cDCE½ � � δ13CcDCE þ VC½ �

PCE½ � þ TCE½ � þ cDCE½ � þ VC½ �:

where [PCE], [TCE], [cDCE] and [VC] are the molar conctions of PCE, TCE, cDCE and VC, respectively, and δδ13CTCE, δ13CcDCE, and δ13CVC their corresponding C iscomposition. The uncertainty was determined by erroragation (Reddy et al., 2002).

2.4.2. Extent of degradationIn order to estimate the extent of degradation in cparts of the plume, the following coefficient was calcu

D ¼ 1− exp Δδ13Cε

!

where Δδ13C corresponds to the difference between theand final C isotopic composition of the considered chlorethene. Such calculationwas performed only for compousampling points where no precursor was present (e.g. forwhen no PCE or TCE was detected) to ensure that the iscomposition was merely affected by the compound degtion and not by its production. In the case of PCE, such cauunnecessary as it can only be degraded. Minimummaximum enrichment factors reported in the literatuwell as a field determined enrichment factor were ussummarised in Table S3, SI.

3. Results and discussion

In this section, results from chemical, isotopimolec-ular biology analyses are given and discussedthe aim to understand the effect of thermal remediatthe chlorinat-ed ethene plume. Only the data froscreens of 13 wells which are located along the centreline are discussed.

3.1. Redox conditions

Concentrations of redox sensitive species measu2006 (Hunkeler et al., 2011) and 2014 are shown in Figgeneral, O2 and/or NO3− concentrations larger than 1 mare detected in the shallow top part of the aquifer (do10–15 m depth) whereas very low concentrations (b0.1−1) are found in deeper parts. Here, the presence ofand/or

2methane indicates the occurrence of reducing conditions. Below 15 m depth, Concentrations of SO4− range from 23 to

part oompot thII)/SO pyritich ers i

ns ararkefounin thurfaclayin 4 mm thgicalnge 4 anin thO2 arden

Fe(Ievenduce of thermient o(Fi

rt anutlan higdiateerm1 (F3 201ue treate2005h fielarbohieverganThis ormemg·

muc welwher for Ovalue NOatte

d B2ourc 201by th2 wa 201whicduce

2

2014 in the upper part of the aquifer from the source zone to 750 m downgradient could also be attributed to the oxidation

conse-lack of Fe(II)

part ns are itions. ions is SO4−

ue to served re less DOC entra-nother Fe(II) adient wells, ncen-pling uction ide or e DOC owline e area ons in le for , espe- At the urce), o 2006 icates

.

s was plume eight source m the r part t 20 m dives duced marily thene

cDCE wells Fig. 2) re also m the ion of .g. 1.1 urther nd for E) and ations 2014 6. The

8

2

2

3

2

59 mg·L−1 and are above 40 mg·L−1, i.e. in the higher the concentration range. Based on the sulfur isotopic csition of SO4−, Hunkeler et al. suggested thadisappearance of O2/NO3− and the increase of Fe(−concentrations with depth could be associated withoxidation processes (Hunkeler et al., 2011), whsupported by the presence of pyrite in sandy aquifJutland (Postma et al., 1991).

Unlike in 2006, in 2014, mixed redox conditiogenerally observed at the local scale. Indeed, in 2014, mfor different redox conditions are simultaneously within some screens (e.g. presence of O2 and Fe(II) same well). This might be explained by a certain subsheterogeneity that results in various redox zones overeach other within screen intervals ranging from 1 toSuch a spatial change in redox conditions may result frotemporal change in redox conditions affecting geolodifferent layers at different speed. This temporal chasupported by the redox species evolution between 2002014 as depicted in Fig. 3 and is especially apparent first 750 m after the source where both Fe(II) and present. More particularly, the NO3− concentration suddrops 100 m downgradient from the source whereconcentrations increase right after the remediation (Fig. 3), which indicates a shift toward more reconditions in 2014 compared to 2006 in the first 750 mplume (Fig. 3). A striking change directly following thremediation is the appearance of DOC in and downgradthe source area that decreases over time 3A and B16-1 in Fig. 3), presumably due to its transpoconsumption. Although aquifers in this part of Jgenerally contain low levels of organic matter,concentra-tions of DOC are measured immedowngradient of the source area after the thremediation treatment, reaching values of 6.1 mg·L−

and 3.1 mg·L−1 (B16, located 100 m downgradient) inSuch a release of sediment-bound organic matter dthermal treatment has been previously reported near tsource areas (Newmark and Aines, 1995, Friis et al., Friis et al. confirmed via experiments performed witmaterial that up to 8% of sediment-bound organic ccould be released in temperature conditions usually acwith thermal treatments (Friis et al., 2005). Releasing omatter can be expected to affect redox conditions. indeed observed immediately downgradient of the fsource area (wells F2 and F3) where O2 (b1 −1) and NO− (up to 12.1 mg·L−1) concentrations arelower in 2014 than those measured in 2006 in nearby(B5, in the source and B11, 100 m downgradient), concentrations of up to 5.2 mg·L−1 and 28.0 mg·L−1

and NO3− were measured, respectively. The lower measured in 2014 could be related to O2 andconsumption during oxidation of organic mAdditionally, the DOC content in B16, B20, B22, anwhich are located up to 400 m downgradient of the swas higher in 2010 (Westergaard et al., 2011) than insupporting the gradual consumption of DOC released thermal treatment. Concomitantly, depletion in Ogenerally observed in B16 B20 B22 B23 and B28 in(Westergaard et al., 2011) compared to 2014, coincides with DOC consumption leading to more re

conditions. The significant change in Fe(II) and SO4concentrations observed between 2006 and

f -e 4

e is n

e rs d e e g . e ly is d e e ly I) t d e al f g. d d h ly al ) 4. o d ). d n d ic is r Lh ls e 2

s 3−

r. 3, e, 4, e s 0 h d −

2

2

2

of organic matter released from remediation, which quently led to more reduced conditions and temporal pyrite oxidation due to lack of oxygen. Indeed, higherconcentrations (N1 mg⋅L−1), especially in the shallowand in B28, as well as slightly lower SO4− concentratioobserved in this part, indicating more reduced condSuch a DOC impact leading to more reduced conditparticu-larly reflected by B16-1 (Fig. 3). High DOC andconcentrations (probably due to pyrite oxidation dsteam injection saturated with air) were indeed obright after remediation in this sampling point, wheconcentrated NO3− and O2 concentrations (due tooxidation) were concomitantly observed. The DOC conction then gradually decreased (B16-1 in Fig. 3). Astriking change between 2006 and 2014 is the lack ofdetected in 2014 between 1000 and 1500 m downgrfrom the source (wells B34, B47 and B58). In theseconcentrations of methane increased whereas SO4− cotrations showed up to 20% decrease in 5 out of 8 sampoints, which suggests the occurrence of SO4− redfollowed by the precipitation of metastable iron sulphFe(II) binding to other minerals. This suggests that threlease affected this part of the aquifer too as the fldescends where the clay layer observed under the sourcdisappears. The occurrence of sulfate-reducing conditithis part of the plume would be additionally favorabchlorinated ethene microbial reductive dechlorinationcially of cDCE and VC (Vogel et al., 1987; Chapelle, 1996).fringe of the plume (i.e. N1800 m from the soconcentra-tions of Fe(II) increased in 2014 compared t(N1 mg·L−1) and methane is detected in B64, which indmore strongly reduced conditions in 2014 than in 2006

3.2. Evolution of chlorinated ethene concentrations

The distribution of individual chlorinated etheneevaluated in a vertical section along the centerline before (2006, Hunkeler et al., 2011) andyears after (2014) performing the thermal remediation (Fig. 4). In both cases, in the first 350 m frosource, the contaminant plume is confined in the uppeof the aquifer due to the presence of a clayey layer adepth. Further downgradient, the plume widens andtoward deeper zones while exhibiting more reconditions, as discussed above. The remediation priresulted in a dramatic drop in chlorinat-ed econcentration immediately downgradient.

In 2014, a strong decrease in PCE, TCE andconcentra-tions of more than 85% is observed inimmediately downgradient (i.e. wells F2, F3 and F4, between 2006 and 2014. Much lower concentrations agenerally measured in wells situated within 750 m frosource in 2014 compared to 2006, with the excepthigher values obtained for TCE at the bottom of B23 (eμmol·L−1 in 2014 instead of 0.43 μmol·L−1 in 2006). Fdowngradient, lower concentrations are generally fouPCE and TCE, which eventually disappear 1050 m (PC1450 m (TCE) downgradient. Similarly, cDCE concentrdecreased 1050 m downgradient of the source incompared to 1050 m downgradient of the source in 200

Fig. 4. Chlorinated ethene concentration in the subsurface in 2006 (Hunkeler et al., 2011) and 2014 (this study).

9

Fig. 5. Mole fractions and C isotopic composition of chlorinated ethenes before (2006, Hunkeler et al., 2011) and after (2014) source thermal treatment.

10

most pronounced change in chlorinated ethene composition in terms of concentrations and species proportions occurs as

ndar cDCwith 201

in thy thderin-1, thveraimila of Vquife4 thaion oa, throun

t, PC lowe mor on On thin thsourcge ie PC750 m biotce s an201

to th thth

t (i.

n b.e. PCe herC ane ratuctiv

and cDCE by pyrite which bears Fe(II) could also contribute to their attenuation (Lee and Batchelor, 2002a). In this case,

ads to via β-cted in uctive an be ounds dwater duced tion of t cDCE Abiotic e place duced 2) and e and

ble S3,

imated sotope l value m the itial Cl ta, the 13C =

100 m es are otopic which in the hibit a ce area sitions tes the 50 m that if either

11

the plume crosses the pyrite oxidation boudepicted in Fig. 11, zone 3. However, an increase in theconcentration at the plume front (N1550 m) together plume front that extends further downgradient in compared to 2006 indicates that cDCE accumulates area and is still slowly expanding, as depicted bdifference in concentration contours in Fig. 4. Consithe high overall groundwater velocity of ~0.24 m·dayalmost identical plume length indicates that the oeffect of natural attenuation is nonetheless sbetween 2006 and 2014. The highest mole fractions(up to 40% in B67) are determined in this part of the a(Fig. 5) although VC concentrations are lower in 201in 2006. Contrary to 2006, when the highest concentratchlorinated ethenes was detected in the source arehigh concentration core of the plume is now located a750 m to 1450 m downgradient.

From the source zone to 750 m downgradienis generally the predominant chlorinated ethene in thepart of the plume in 2014; its concentration reachesthan 85%of the total chlorinated ethene concentrationmolar basis at the bottom of wells B22 and B23 (Fig. 5). other hand, the mole fraction of cDCE dominates upper part of the plume within 350 m from the where mole fractions reach up to 74%. The chanchlorinated ethene distribution between 2006 wherwas the predominant chlorinated ethene in the first of the plume and 2014 suggests the occurrence ofreductive PCE/TCE dechlorination. Such an occurrenadditionally supported by higher DOC concen-trationmore strongly reduced conditions observed in compared to 2006 in this part of the plume. In contrastpart of the plume immediately downgradient fromsource where PCE dominates, cDCE is predominant chlorinated ethene further downgradienN750 m).

During biotic reductive dechlorination, PCE catransformed via sequential hydrogenolysis to ethene (i→ TCE → cDCE → VC → ethene). Yet, VC and ethene areither detected at trace levels (VC) or not detected (Vethene), which suggests that cDCE transformation is thlimiting step if it is primarily degraded by biotic red

dechlorination. Abiotic reductive dechlorination of PCE, TCE The Cl

their reflect values , Table rd the tween

350 m dient).

in ture of 50 m s 1050 ded in in the ce area either rary to

Fig. 6. Dual C\\Cl isotope slopes associated with PCE, TCE and cDCE. One C\\Cisotopic composition of cDCEwhich did not alignwith the others is representeseparately as a cross. Slopes are given with 95% confidence interval.

y E a 4 is e g e ll r C r n f e d

E r e a e e e n E

ic is d 4 e e e e.

e E e d e e

transformation of PCE, TCE and cDCE generally leformation of acetylene as the main intermediate dichloroelimination. Although acetylene was not detethe aquifer, the additional occurrence of abiotic reddechlorination cannot be overlooked as acetylene cfurther transformed to readily biodegradable compsuch as acetaldehyde, acetate and ethanol in groun(Liang et al., 2009). Finally, while biotically promackinawite may catalyze abiotic reductive dechlorinaPCE and TCE (Butler and Hayes, 1999), it was shown thawas not reactive with mackinawite (Jeong et al., 2011). reductive cDCE dechlorination is thus not likely to takin the presence of mackinawite, though biotically pronon-measurable surface-bound Fe(II) (Han et al., 201other reduced iron minerals such as green rust (LeBatchelor, 2002b) may induce such a reaction.

3.3. Chlorinated ethene C and Cl isotopic composition

Chlorinated ethene concentrations and isotopic composi-tions measured in 2014 are summarised in TaSI.

The C isotopic ratio of released PCE was formerly estto be −25.0 ‰ (Hunkeler et al., 2011). The Cl imeasurements performed in 2014 show a lowest δ37Cof −1.0‰ for PCE at 100 and 750 m downgradient frosource, which is the closest measured value to the inisotope signature of released PCE. Without further dainitial PCE isotopic signature is hence assumed to be δ−25.0‰ and δ37Cl = −1.0‰.

In 2014, apart from two sampling points located downgradient from the source, PCE C isotopic valugenerally more enriched in 13C than the initial issignature, reaching up to −19.5‰ in B34-6 (Fig. 5), clearly indicates that PCE degradation is occurring plume. On the other hand, TCE C isotopic values exhigher spatial variability. The sampling points in the sourup to 350 m downgradient show C isotopic compovarying from−27.0 to −19.3‰ (Fig. 5), which indicaoccurrence of TCE degradation. Conversely, 7downgradient, values of−30.7 ‰ to −35.0‰ indicateTCE is undergoing further degradation, this process islimited or slower than TCE production from PCE. isotopic values of PCE and TCE are coherent withrespective C isotopic values: the highest values, whichenrichment in 37Cl, are found where higher C isotopicare detected. (e.g. B23-2 for PCE and TCE, B22-3 for TCES3, SI). This additional line of evidence points towaoccurrence of degradation in these sampling points (bethe source and 350 m downgradient).

C isotopic ratios of cDCE vary from −38.5‰ (B23-2,downgradient) to −21.3‰ (B34-3, 1050 m downgraIsotopic values enriched 13C compared to the assumed initial PCE isotopic signa−25.0‰ are found in the source zone up to 3downgradient in the upper part of the aquifer, as well am downgradient. This suggests that cDCE is being degrathese parts of the aquifer. Lower C isotopic values founddeeper part of the aquifer 350 m to 750 m from the sourindicate that if cDCE is being degraded, this process islimited or slower than cDCE production from TCE. Cont

ld

2006, when C isotopic values of cDCE and VC

reached up to −18 ‰ and −13 ‰, respectively, at the plume front (N1400 m), such enriched values are not observed in

lues o.2 ‰resenlso bresener, VcDC

d in ar et a littwever thaies o

2008ormeead tpletemo

adienting

rest onablelace, cessere th

ition 2010tationing e

± 1sourcrang

fiel7–2.7 C\\Car, nstitutelinesin th

ed foe areportee TC sincious botlled bwhe wato 30whicurthee duplumpare

TC±1

ved irang.4–4.

Kuder et al., 2013), but is significantly different from the slope obtained for abiotic TCE reduction (5.2 ± 0.3, Audí-Miró et al.,

f biotic

of 2.1 2011) 050 m he time e cDCE

0.004 e dual ctively studied ecently ociated et al., s closer ination ination. on that ductive of a lent Fe though not the garded s likely ed iron 50 m though nation) ), C\\Cl loser to ient to

within ociated ro et al. pports

e of the ithout ence of ver the pletion ic cDCE 08 and

e area, at the osition sumed erefore e takes tive VC ducing, n was et al., nclear. as in ctually at O2 is 2010). bic or

12

the same part of the plume in 2014, when C isotopic vacDCE and VC reached up to −24.0 ‰ and −23respectively. This suggests that cDCE, or VC when it is pexperiences less degradation. These lower values could aattributed to the influx of the less degraded cDCE p1050 m downgradient from the source in 2006. Moreovis generally almost not depleted in comparison to which contradicts all trends of C isotope ratios observebiodegradation studies to date (Bloom et al., 2000; Slate2001; Abe et al., 2009; Fletcher et al., 2011). Suchdifference in cDCE and VC isotopic composition could horeflect the occurrence of abiotic cDCE degradation rathebiotic degradation. Indeed, this is consistent with studabiotic chlorinated ethene degradation (Elsner et al.,Audí-Miró et al., 2013) where VC and acetylene were fin parallel and where isotopically sensitive branching lthe unexpected situation where VC was scarcely decompared to its precursor. The 13C enriched values for cDCE are found 1050 m downgrof the source as well as close to the source, indicastronger degradation activity in these areas than in thethe aquifer. While single element isotopic analysis eidenti-fication of areas where degradation is taking pdoes not permit differentiation between various progoverning degradation, except in the case of cDCE whedata point toward abiotic degradation.

Dual C\\Cl isotope slopes may provide addinsight into such processes (Abe et al., 2009; Elsner,Cretnik et al., 2013; Kuder et al., 2013) although limiwere recently pointed out (Badin et al., 2014; Renpennal., 2014). For PCE, the dual C\\Cl isotope slope of 3.0associated with data points located between the area and 750 downgradient falls within the reported for microbial reductive dechlorination instudies (0.7–3.5) and in laboratory experiments (0.(Badin et al., 2014) (Fig. 1 and Fig. 6). Since no dualslope for abiotic PCE reduction has been reported so fcomparison is possible. These observations thus conan additional line of evidence that supports the likthat PCE undergoes biotic reductive dechlorination first 750 m of the plume.

The dual C\\Cl isotope slope of 2.7 ± 1.0 observTCE for sampling points located between the sourcand 1050 m downgradient cannot be compared to reranges of dual isotope slopes associated with soldegradation in a way as straightforward as for PCEhere TCE is both produced and consumed. It was prevshown that the dual isotope slope associated with aproduced and degraded chlorinated ethene was controits degradation when this step was rate-limiting, i.e. an accumulation of this intermediate compoundobserved (Hunkeler et al., 2009). TCE accounts for up 40% at 350 and 750 m downgradient from the source, suggests that it accumulates before becoming fdegraded to cDCE. It can thus be assumed that thisotope slope associated with TCE in this part of the is controlled by TCE degradation and may be comwith slopes associated with biotic reductivedechlorination. When taking the uncertainty of into account, the TCE dual C\\Cl isotope slope obserRødekro is not significantly different from the observed for biotic reductive TCE dechlorination (3

Cretnik et al., 2013;

f , t, e t C E, ll l., le r n n ; d o d st

t a f s it s e

al ; s t .2 e e d ) l o e s e

r a d E e ly h y n s –h r al e d E .0 n e 8,

2013). These observations support the occurrence oreductive TCE dechlorination in this part of the plume.

The formerly reported dual C\\Cl isotope slope (published εCl/εC = 0.48 ± 0.05, Hunkeler et al.,associated with cDCE from data points located 1downgradient of the source to the plume front was at tcompared with slopes associated with biotic reductivdechlorination (11.4 to 13.7, published εCl/εC = 0.088 ±to 0.073 ± 0.006). It was thus concluded based on thisotope approach that cDCE was either redudechlorinated by other strains than those previously or that abiotic degradation occurred. Audi-Miro et al. rreported a dual C\\Cl isotopic slope of 3.1 ± 0.2 asswith abiotic reductive cDCE dechlorination (Audí-Miró2013). The slope value observed in Rødekro in 2006 wato the one observed for abiotic reductive cDCE dechlorthan to those associated with biotic reductive dechlorA slope of 1.5 ± 0.2 associated with cDCE degradatioccurred by simultaneous biotic and abiotic redechlorination in the subsurface in the presencepermeable reactive barrier containing 3% (v/v) zero-vawas more recently reported (Audí-Miró et al., 2015). AlAudi-Miro et al. concluded that abiotic reduction was predominant degradation process, it cannot be disrethat this slope is influenced by the latter. cDCE was thuto be abiotically reduced by pyrite (and/or other reducminerals) in the sediment in 2006 between 10downgradient of the source and the plume front alacetylene (degradation product of abiotic dihaloelimiwas not detected. In 2014, except for one point (B23-3isotope data align with generally less enriched values cthe source and more enriched values further downgradgenerate a slope of 3.0 ± 0.7 (R2 = 0.82). This slope isthe 95% confidence interval of the slope of 3.1 ± 0.2 asswith abiotic cDCE degradation determined by Audi-Mi(Audí-Miró et al., 2015) (Fig. 6 and Table S 3, SI). This suthe predominance of abiotic cDCE reduction in the corplume in 2014 as also indicated by the C isotopic data. Wthe evidence brought forward by isotopic data, the presVC could be attributed to cDCE hydrogenolysis. Howelittle VC 13C depletion compared to cDCE 13C desuggests that VC may rather be produced during abiotdegradation, as previously observed by Elsner et al., 20Audi-Miro et al., 2013.

Apart from immediately downgradient of the sourcwhere the VC C isotopic composition is −22.9 ‰, anddeepest sampling point in B58, the VC C isotopic compdoes not become much more enriched than the asinitial isotopic signature of PCE (−25.0 ‰). It is thunlikely that further hydrogenolysis of VC to ethenplace despite more optimal redox conditions for reducdechlorination in 2014 than in 2006 (i.e. sulfate-reChapelle, 1996). Finally, while anaerobic oxidatiosuggested as a possible degradation pathway (Smits2011), the likeliness of such a process to occur remains uGossett et al. previously suggested that what wmicrocosms thought to be anaerobic oxidation might abe aerobic oxidation with O2 concentrations so low thquickly consumed and not measurable (Gossett, However, based on the current data, anaeromicroaerophilic oxidation cannot be ruled

out. Moreover, the conditions are not aerobic in this part of the aquifer, which means aerobic oxidation is not possible.

B34-13C oimite-tox via otop

luablerticchlor the eaccts oof TCntally values hment the C nature most alue is other

is less ecular

of PCE and rature arried

could f PCE eption was

PCE nding ation,

data. Results can be found in Table S 3, SI. For PCE, the extent ofdegradation determined with the C enrichment factor varies

t withst ande PCEated

eoveructive7 andallestocess,rada-

Fig. 7.Dhc population size and activity in the subsurface. Given values correspond to the ratio rRNA/DNA. High N1·106 gene copies·L−1; Intermediate ∈ [3·105–1·106]gene copies·L−1; Low ∈ [bdl - 3·105] gene copies·L−1. na: not analysed. dl: detection limit. Bdl: below detection limit. nd: not detected.

Fig. 8. Relative proportion of bacteria which have been identified to belong to a16S rRNA gene based phylogenetic group that has been shown to containbacteria involved in iron-reduction, sulfate-reduction, partial dechlorination,pyrite oxidation, biotic chlorinated ethene oxidation or complete dechlorina-tion processes. Only these phylogenetic groups (Table S 5, SI) are included inthe analysis.

13

The highest δ13Csum value at the site (−21.3 ‰ in Table S 3, SI) indicates a maximum enrichment in 3.7‰. This also supports the supposition that ltransformation of chlorinated ethenes to noncompounds, such as ethene, is occurring, unless it isdegradation process associated with low isfractionation.

Dual C\\Cl isotope slopes convey even more vainformation. Hunkeler et al., 2009 predicted that the vspacing between dual isotope slopes associated with nated ethenes during reductive dechlorination reflectedenrichment factors when the slopes run parallel toother. The reason is that the kinetic Cl isotope effect athe cleaved chloride so that the chlorine isotope ratio is expected to match that of PCE (as experimeconfirmed by Cretnik et al., 2014), whereas C isotope differ by the C enrichment factor εC. Here, a C enricfactor of −10.3 ‰may be determined when subtractingisotope value of the most depleted PCE C\\Cl sig(corresponding to the initial PCE signature) from thedepleted TCE C\\Cl signature (where the Cl isotope vthe closest to that of the initial PCE signature). On thehand, interpreting the spacing between TCE and cDCEstraightforward as it is also influenced by the intramolchlorine isotope distribution in TCE.

In order to determine the extent of degradation and cDCE in some sampling points, minimummaximum enrichment factor values reported in the litewere chosen since no microcosm experiment was cout based on which site specific enrichment factorsbe determined. The estimation of the extent odegradation based on C isotopic data is an excwhere the C enrichment factor determined aboveused. As the data support biotic reductivedechlorination and abiotic cDCE reduction, correspoenrichment factors were chosen. The extent of degradD, was determined both with C and Cl isotope

3, f d ic a ic

e al i-C h n E

from 3 to 59%with an average of 28%. This is in agreementhe average extent of degradation determined with largesmallest Cl enrichment factors associated with reductivdechlorination of 10 and 34%, respectively. The estimextent of degradation based on C and Cl data are morconsistent for each sampling point. Finally, for abiotic redcDCE dechlorination, average extents of degradation of19% could be determined based on the largest and smenrichment factors associated with this degradation prrespectively. This is in agreement with the observed degtion stall at cDCE.

3.4. Input from molecular biology

A and alyses ion to nes in

qPCR m not gene es are while 050 m ities of pling

3, and iously nation 105 to nough ectiorrencctivit 2.48sourchighe welat DhDhc i-1 an), Dhder oet aDhc isotop–2 anble S

thA) are thotopnce o thnation A and trated

genus ion to

of the mpled ps are dizing, plete

acteria erobic , such ting as list of yrotag with ed in

Fig. 9. Proportion of bacteria that have been identified to belong to a 16S rRNA gene based phylogenetic group that has been shown to contain bacteria involved in sulfate-reduction, iron-reduction and pyrite oxidation normalized by the maximum total quantity of bacteria involved in such redox processes among all samples.

bioremediation (Fig. 10) could also be determined. Generally, low relative abundance (ranging from 0 to 4%) is found in all

ong all n- and ethene nation) latively based to be

wells during relative ndance redox

idation acteria und in B34-6, ient of isotope inates

some ductive ffler et

Fig. 10. Proportion of bacteria that have been identified to belong to a 16S rRNAgene based phylogenetic group that has been shown to contain bacteriainvolved in biotic chlorinated ethene oxidation, and partial or completedechlorination processes normalized by the maximum total quantity ofbacteria involved in chlorinated ethene bioremediation processes among allsamples.

14

454 pyrotag sequencing, Dhc DNA and mRNA, bvcvcrA functional genes, and gene transcript (mRNA) anwere performed to acquire complementary informatunder-stand the processes affecting chlorinated ethethe aquifer after the source thermal remediation.

The total Dhc population analysis performed byshows a relatively large range across sampling wells frodetected (below detection limit, bdl) to 1.73 · 106

copies⋅L−1 (Fig. 7 and Table S 4, SI). Higher abundancmeasured from the source area to 350 m downgradientlower abundances to non-detects are measured from 1downwards except in B34-4 where intermediate quant1.03 · 105 gene copies⋅L−1 are detected. Only a few samlocations (i.e. F4-3, B16-1, B17-1, B23-1, B23-2, B23-B34-4) indicate Dhc quantities in the ranges prevreported for other sites where biotic reductive dechlorihas naturally occurred (Damgaard et al., 2013b) (from106 gene copies ⋅-L−1,Table S 4, SI) and are thus high eto support some dechlorination potential. Yet, the detof Dhc DNA is not sufficient to confirm the actual occuof dechlorination as DNA does not indicate bacterial aOn the other hand, Dhc targeted rRNA is detected up to105 copies⋅L−1 in sampling points located between the zone to 1050 m downgradient (Table S4, SI). Relatively proportions of 16S rRNA than corresponding DNA inB16-1 and B23-2 compared with other wells indicate thin B16-1 and B23-2 are more metabolically active than some other locations. Although rRNA/DNA ratios in B16B23-2 are not much higher than 1 (max. ratio: 1.56activity was previously associated with ratios of this ormagnitude in a previous laboratory study (Wagner 2013). The presence and apparent high activity of B16-1 and B23-2 coincide with highly enriched C iratios of −17.6‰ for PCE and −19.3‰ for TCE in B23−22.9‰ for cDCE in B16-1 (Fig. 5, Fig. 7, Table S3 and TaSI).

Genes coding for the enzymes involved infinal transformation of VC to ethene (vcrA and bvcneither present (DNA) nor expressed (mRNA) abovdetection limit. This is consistent with the minor C isenrichment observed for VC, which supports an abseVC degradation by reductive dechlorination atsite. Nevertheless, further reductive VC dechloricannot be ruled out as VC rdhA genes other than vcrbvcA may exist. It was indeed recently demonsthat some members of the Dehalogenimonas (Dhg)respire VC and thus participate in VC dechlorinatethene (Yang and Löffler, 2015).

Pyrotag sequencing enabled the identification relative abundance of various bacteria in the salocations (Fig. 8 and Table S5, SI). As only specific groutargeted (i.e. iron-, sulfur-reducing, pyrite oxichlorinated ethene oxidizing, chlorinated ethene comand partial reductive dechlorinators), other groups of bwhich may be present in some samples, such as asamples, do not appear. In the absence of O2 and NO3−

bacteria may use iron as an electron acceptor, thus acan iron-reducer although it is not included among theiron-reducers established in this study. Based on psequencing data, relative proportions to the samplehighest bacteria quantity of various bacteria involv

redox processes (Fig. 9) and in chlorinated ethene

n e y. · e r ls c n d c f l., n e d4,

e e e ic f e

samples for bacteria likely involved in iron reduction amof the considered groups (i.e. bacteria involved in irosulfate-reduction, pyrite oxidation, chlorinated oxida-tion, and partial and complete reductive dechlori(Fig. 8 and Table S5, SI), which coincides with the rereduced conditions observed in most of the aquiferon redox species concentrations. Bacteria reported present in sulfate-reducing conditions (4 out of 10where the relative proportion is N10%) as well as pyrite oxidation (6 out of 10 wells where the proportion is N26%) are detected in high relative abu(Table S5 and Fig. 8, SI), which corroborates the data and indicates a high likelihood of pyrite oxand sulfate-reducing conditions. More specifically, breported to be involved in pyrite oxidation are forelative abundance higher than 20% in B34-2, B34-3,B58-2, B58-6 and B61-3, i.e. from 1050 m downgradthe source. This coincides with cDCE dual C\\Cl data that suggests abiotic cDCE reduction predomin these locations. Bacterial genera among whichmembers are known to catalyze complete redechlorination of chlorinated ethene (i.e. Dhc, Löal., 2013, and Dhg, Yang and Löffler, 2015) are

detected in the majority of sampling locations (Fig. 8). In particular, the relative abundance of bacteria potentially able

igh i0 mpletthes

do noed b

4 anencinfferenDhg igh icterieducted i. 8 anble tut thy an-3 ans thapart oved iof thwhicinate

than an, F4-

th 105ot bdatio wer

thidizedwhereve

par

2006

inatein th thesplum

thfferenitionecula

walikeln th

influx of low background DOC or to the release of DOC during thermal remediation (2014). Redox conditions are

han in inant

source TCE Biotic ut to a tically hortly ns are

more ted by m the uctive s here disap-00 m es the does is part CcDE b duced

luated better cluded ed by hough of VC led to d the area. ruptly icating itation s (and d iron SO4−

o the NO3−

to the II) and in this 050 m e data sotope th the gested 4. This hich is he low t the ssible ich is ounds

15

to completely dechlorinate chlorinated ethene is hB23-2 (350 m downgradient) and B34-6 (105downgradient), indicating the possibility that comreductive dechlorina-tion may have occurred in locations. The counts of Dhc from pyrotag sequencing exactly coincide with the copy number of Dhc determintargeted quantitative PCR after DNA extraction (Table STable S 5, SI), which may be because 454 pyrotag sequis considered a semi-quanti-tative method and diamplification primers were used. Contrary to Dhc, detected in all samples, and reads are particularly hB23–2 (1.7·106 cells⋅L−1, Table S5, SI). Among the bagenera among which some members are known to rchlorinated ethenes only partially, several are detecparticular in B23-3, B34-2, B34-4, 34-6 and B61-1 (FigTable S5, SI), which indicates that bacteria potentially adechlorinate chlorinated ethenes are present throughoentire plume. Finally, highest quantities of partialltotally dechlorinating bacteria are found in B23-2, B23B34-2 (Fig. 10), which supports the higher likelinesbiotic reductive dechlorination is occurring in the first the plume. The highest proportions of bacteria involpyrite oxidation are found from the core to the front plume (B34-2, B34-3, B58-2 and B58-6, Fig. 9), supports the predominant occurrence of abiotic chlorethene reduction in this part of the plume.

Generally, the molecular biology data suggestsreduc-tive dechlorination could potentially occur isampling location, with a higher likelihood in B16-1(which are located immediately downgradient fromsource), B23-2 and B34-4 (which are located 350 andm downgradient from the source). Moreover, it canndefinitively deter-mined whether complete degrais possible because neither bvcA nor vcrA genesdetected.

Pyrotag sequencing data also highlight possibility that cDCE and/or VC may have been oxespecially downgradient (from B58 to B61) Polaromonas reads are the highest (Table S 5, SI). Howthis is more plausible in the upper rather than deeperof the aquifer, where oxic conditions are unlikely.

4. Summary of redox and chlorinated ethene degradation processes occurring in the aquifer from to 2014

Drawings that describe redox and chlorethene degradation processes that likely occurred subsurface in 2006 and 2014 and the consequences ofprocesses on chlorinated ethene distribution in the are given in Fig. 11. These models result fromcombination of lines of evidence brought by the dimethods applied in this study (i.e. redox condcontaminant concentration, iso-tope analysis and molbiology).

4.1. First 750 m

Based on the performed investigations, itconfirmed that PCE and TCE are very dechlorinated by biotic reductive dechlorination i

upper part of the plume, in the first 350 to 750 mdowngradient from the source, either due to the formepresence of hydrocarbons (2006) and natural

n

e e t y d g t is n al e n d o e d d t f n e h d

t y 3 e 0 e n e

e , e r, ts

d e e e e t s, r

s y e

2

2

2

2

generally more reduced in the source area in 2014 t2006 due to the DOC release. cDCE is the domchlorinated ethene immediately downgradient of the in 2014 as a result of biotic reductive PCE anddechlorination occurring in more reduced conditions. reductive dechlorination of cDCE also likely occurs blesser extent, and both cDCE and VC may be biooxidized in the upper part of the aquifer sdowngradient of the source where aerobic conditiopresent.

4.2. 750 to 900 m downgradient

As the plume moves forward, it dips and enters areduced zone starting from ~15 m depth that is creapyrite oxidation consuming the influx of O2 and NO3− frowater recharge (both in 2006 and 2014). Biotic reddechlorination of PCE and TCE takes place as conditionbecome manganese/iron reducing, which explains the pearance of these compounds between ~700 and ~9downgradient from the source, where the plume crosspyrite oxidation zone. Reductive dechlorination of cDCEnot appear to have occurred to a significant extent in thof the plume in 2006 nor in 2014, as supported by δ13

δ13-13Csource and the presence of insufficiently reconditions.

4.3. From 1000 m to 1500 m

The fate of cDCE in 2006 after 1000 m was better evabased on studies performed since 2011 that allow a dual C\\Cl isotope slope interpretation. It could be conthat cDCE was likely predominantly abiotically reducpyrite in 2006 between 1000 m and the plume front altbiotic degradation may have occurred as well (presenceand detection of Dhc). In 2014, the DOC release that more reduced conditions seems to have impactesubsurface to ~1500 m downgradient of the sourceIndeed, no Fe(II) and lower SO4− concentrations are abobserved between ~1000 and ~1500 m in 2014, indthat SO4− reduction followed by iron sulphide precipsuch as mackinawite and/or Fe(II) sorption on mineralpotential formation of green rust and other reducespecies) likely occurred. Such a drop in Fe(II) andconcentrations could also be partially attributed tabsence of pyrite oxidation that resulted from O2 andconsumption in the upper part of the aquifer close source after the DOC release, and therefore limited Fe(SO4− influx. cDCE degradation is probably occurring part of the plume, as indicated by δ13CcDE N δ13Csource 1downgradient from the source. Based on C\\Cl isotopand on the assumption that currently available dual islopes represent the range of slopes associated wicorrespond-ing degradation process, it was also sugthat cDCE is predominantly abiotically reduced in 201is also supported by the VC C isotopic composition wgenerally almost depleted compared to that of cDCE. TVC concentrations could additionally documenoccurrence of abiotic degradation to which the popresence of non-measurable surface-bound Fe(II) whknown to catalyze abiotic degradation of some comp

r

(Elsner et al., 2004; Han et al., 2012), could contribute. On the other hand, the

Fig. 11. Overview summarising redox conditions and processes as well as processes affecting chlorinated ethenes (CE) in the subsurface in 2006 and in 2014.

16

observed redox conditions are more favorable for bioticreductive cDCE and VC dechlorination than before remediation,explaining the occurrence of Dhc and Dhc activity around1050 m downgradient in 2014.

4.4. From 1500 m to the plume front

Further downgradient where the source remediation didnot impact the plume (N1500 m), cDCE does not seem to bedegraded, as indicated by the presence of little VC as well asδ13CcDE ≈ δ13Csource. Moreover, that the plume front extended~200 m further in 2014 compared to 2006 (concentrationcontours, Fig. 5) together with the presence of cDCE in slightlyhigher concentrations in 2014 in the downgradient part of theplume compared to 2006 suggests that cDCE is still expanding,though slowly (almost steady state). This coincides with theobservation of the small extent of cDCE degradation based onisotopedata ranging from7 to 19%. Such anobserved differencein the cDCE concentration at theplume front between 2006 and2014 might also be attributed to a slight lateral change in flowdirection. An explanation to the lack of further degradationcould be the absence of active degraders in 2014 as depicted inFig. 7. Finally, it was concluded based on isotope andmolecularbiology data that the small amounts of VC produced in thedowngradient part of the plume is scarcely degraded further.

5. Assessment of source thermal remediation effect onredox conditions and the fate of chlorinated ethenes

Due to the contaminant source removal, the overallchlorinated ethene concentrations in and downgradient of thesource area dramatically decreased between 2006 and 2014.Although no plume detachment was observed, the sourcethermal remediation seems to have triggered a change in redoxspecies concentrations, more particularly of iron. Indeed, anincrease in DOC in the source zone, likely due to the release oforganic matter during the treatment, could support microbialgrowth triggering in turn changes in redox conditions as aresult of electron acceptors consumption such as O2 andNO3−. Achain of redox reactions influenced by the additional presenceof pyrite in the aquifer eventually affected the degradation ofchlorinated ethenes,which led primarily to an increase of bioticPCE and TCE degradation to cDCE immediately downgradientof the source area and predominantly to an abiotic reduction ofcDCE in the plume centre. Intricate situations similar to that ofthe current study, where iron plays a role in redox reactionsand both bacterial and abiotic degradation of chlorinatedcontaminants occur, have been previously reported (Elsner etal., 2004; Shani et al., 2013; Broholm et al., 2014). Thisunderlines the need for studies that explore the dynamics ofgeochemical systems where iron is present.

This study thus demonstrates the strength of complemen-tary application of analytical and molecular biology tools togain insight to processes occurring in the subsurface where aplume of chlorinated ethenes flows in a complex geochemicalsystem.

Acknowledgements

The authors would like to acknowledge Niels Just and theRegion of Southern Denmark for providing a sampling

opportunity in the study site and additional funding as well assharing information relative to the site, and Jesper Gregersenfor his precious help in organising and performing fieldsampling. Orfan Shouakar-Stash and Mirna Stas are acknowl-edged formeasuring Cl isotopic ratios of cDCE. JulienMaillard isacknowledged for critical reading of the manuscript. TorbenDolin (DTU) is thanked for the plume illustrations (Figs. 3, 4, 5and 7). Alexandra Murray (DTU) is acknowledged forreviewing the English. Three anonymous reviewers are ac-knowledged for their very helpful comments. This researchwascompleted within the framework of the Marie Curie InitialTraining Network ADVOCATE - Advancing sustainable in situremediation for contaminated land and groundwater, fundedby the European Commission, Marie Curie Actions ProjectNo. 265063.

Appendix A. Supplementary data

Supplementary data to this article can be found online athttp://dx.doi.org/10.1016/j.jconhyd.2016.05.003.

References

Abe, Y., Aravena, R., Zopfi, J., Shouakar-Stash, O., Cox, E., Roberts, J.D., Hunkeler,D., 2009. Carbon and chlorine isotope fractionation during aerobicoxidation and reductive dechlorination of vinyl chloride and cis-1,2-dichloroethene. Environ. Sci. Technol. 43 (1), 101–107.

Aeppli, C., Holmstrand, H., Andersson, P., Gustafsson, O., 2010. Directcompound-specific stable chlorine isotope analysis of organic compoundswith quadrupole GC/MS using standard isotope bracketing. Anal. Chem. 82(1), 420–426.

Audí-Miró, C., Cretnik, S., Otero, N., Palau, J., Shouakar-Stash, O., Soler, A., Elsner,M., 2013. Cl and C isotope analysis to assess the effectiveness of chlorinatedethene degradation by zero-valent iron: evidence fro dual element andproduct isotope values. Appl. Geochem. 32, 175–183.

Audí-Miró, C., Cretnik, S., Torrentó, C., Rosell, M., Shouakar-Stash, O., Otero, N.,Palau, J., Elsner, M., Soler, A., 2015. C, Cl and H compound-specific isotopeanalysis to assess natural versus Fe(0) barrier-induced degradation ofchlorinated ethenes at a contaminated site. J. Hazard. Mater. 299, 747–754.

Aulenta, F., Majone, M., Verbo, P., Tandoi, V., 2002. Complete dechlorination oftetrachloroethene to ethene in presence of methanogenesis andacetogenesis by an anaerobic sediment microcosm. Biodegradation 13(6), 411–424.

Badin, A., Buttet, G., Maillard, J., Holliger, C., Hunkeler, D., 2014. Multiple dualC–Cl isotope patterns associated with reductive dechlorination oftetrachloroethene. Environ. Sci. Technol. 48 (16), 9179–9186.

Bælum, J., Chambon, J.C., Scheutz, C., Binning, P.J., Laier, T., Bjerg, P.L., Jacobsen,C.S., 2013. A conceptual model linking functional gene expression andreductive dechlorination rates of chlorinated ethenes in clay richgroundwater sediment. Water Res. 47 (7), 2467–2478.

Ballapragada, B.S., Stensel, H.D., Puhakka, J.A., Ferguson, J.F., 1997. Effect ofhydrogen on reductive dechlorination of chlorinated ethenes. Environ. Sci.Technol. 31 (6), 1728–1734.

BIPM, IEC, IFCC, ISO, IUPAC, IUPAP, OIML, 1993. Guide to the Expression ofUncertainty in Measurement. International Organization for Standardiza-tion, Geneva, Switzerland.

Bloom, Y., Aravena, R., Hunkeler, D., Edwards, E., Frape, S.K., 2000. Carbonisotope fractionation during microbial dechlorination of trichloroethene,cis-1,2-dichloroethene, and vinyl chloride: implications for assessment ofnatural attenuation. Environ. Sci. Technol. 34 (13), 2768–2772.

Bradley, P.M., 2000. Microbial degradation of chloroethenes in groundwatersystems. Hydrogeol. J. 8 (1), 104–111.

Bradley, P.M., Chapelle, F.H., 1998. Microbial mineralization of VC and DCEunder different terminal electron accepting conditions. Anaerobe 4 (2),81–87.

Bradley, P.M., Chapelle, F.H., 2000. Aerobic microbial mineralization ofdichloroethene as sole carbon substrate. Environ. Sci. Technol. 34 (1),221–223.

Broholm, M.M., Hunkeler, D., Tuxen, N., Jeannottat, S., Scheutz, C., 2014. Stablecarbon isotope analysis to distinguish biotic and abiotic degradation of1,1,1-trichloroethane in groundwater sediments. Chemosphere 108 (0),265–273.

17

Butler, E.C., Hayes, K.F., 1999. Kinetics of the transformation of trichloroethyleneand tetrachloroethylene by iron sulfide. Environ. Sci. Technol. 33 (12),2021–2027.

Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello,E.K., et al., 2010. QIIME allows analysis of high-throughput communitysequencing data. Nat. Methods 7 (5), 335–336.

Chapelle, F.H., 1996. Identifying redox conditions that favor the naturalattenuation of chlorinated ethenes in contaminated ground-water systems.Symposium on Natural Attenuation of Chlorinated Organics inGroundwater.US EPA, Washington, DC, pp. 17–20

Cretnik, S., Thoreson, K.A., Bernstein, A., Ebert, K., Buchner, D., Laskov, C.,Haderlein, S., Shouakar-Stash, O., Kliegman, S., McNeill, K., Elsner, M., 2013.Reductive dechlorination of TCE by chemicalmodel Systems inComparisonto Dehalogenating bacteria: insights from dual element isotope analysis(13C/12C, 37Cl/35Cl). Environ. Sci. Technol. 47 (13), 6855–6863.

Cretnik, S., Bernstein, A., Shouakar-Stash, O., Löffler, F., Elsner, M., 2014.Chlorine isotope effects from isotope ratio mass spectrometry suggestintramolecular C-Cl bond competition in trichloroethene (TCE) reductivedehalogenation. Molecules 19 (5), 6450–6473.

Damgaard, I., Bjerg, P.L., Bælum, J., Scheutz, C., Hunkeler, D., Jacobsen, C.S.,Tuxen, N., Broholm, M.M., 2013a. Identification of chlorinated solventsdegradation zones in clay till by high resolution chemical, microbial andcompound specific isotope analysis. J. Contam. Hydrol. 146 (0), 37–50.

Damgaard, I., Bjerg, P.L., Jacobsen, C.S., Tsitonaki, A., Kerrn-Jespersen, H.,Broholm, M.M., 2013b. Performance of full-scale enhanced reductivedechlorination in clay till. Ground Water Monit. Rem. 33 (1), 48–61.

Duhamel, M., Wehr, S.D., Yu, L., Rizvi, H., Seepersad, D., Dworatzek, S., Cox, E.E.,Edwards, E.A., 2002. Comparison of anaerobic dechlorinating enrichmentcultures maintained on tetrachloroethene, trichloroethene, cis-dichloroethene and vinyl chloride. Water Res. 36 (17), 4193–4202.

Edgar, R.C., Haas, B.J., Clemente, J.C., Quince, C., Knight, R., 2011. UCHIMEimproves sensitivity and speed of chimera detection. Bioinformatics 27(16), 2194–2200.

Elsner, M., 2010. Stable isotope fractionation to investigate natural transfor-mation mechanisms of organic contaminants: principles, prospects andlimitations. J. Environ. Monit. 12 (11), 2005–2031.

Elsner, M., Schwarzenbach, R.P., Haderlein, S.B., 2004. Reactivity of Fe(II)-bearing minerals toward reductive transformation of organic contami-nants. Environ. Sci. Technol. 38 (3), 799–807.

Elsner, M., Zwank, L., Hunkeler, D., Schwarzenbach, R.P., 2005. A new conceptlinking observable stable isotope fractionation to transformation pathwaysof organic pollutants. Environ. Sci. Technol. 39 (18), 6896–6916.

Elsner, M., Chartrand, M., Vanstone, N., Couloume, G.L., Lollar, B.S., 2008.Identifying abiotic chlorinated ethene degradation: characteristic isotopepatterns in reaction products with nanoscale zero-valent iron. Environ. Sci.Technol. 42 (16), 5963–5970.

Fletcher, K.E., Nijenhuis, I., Richnow, H.H., Loffler, F.E., 2011. Stable carbonisotope enrichment factors for cis-1,2-dichloroethene and vinyl chloridereductive dechlorination by dehalococcoides. Environ. Sci. Technol. 45 (7),2951–2957.

Flynn, S.J., Löffler, F.E., Tiedje, J.M., 2000. Microbial community changesassociated with a shift from reductive dechlorination of PCE toreductive dechlorination of cis-DCE and VC. Environ. Sci. Technol. 34 (6),1056–1061.

Friis, A.K., Albrechtsen, H.J., Heron, G., Bjerg, P.L., 2005. Redox processes andrelease of organic matter after thermal treatment of a TCE-contaminatedaquifer. Environ. Sci. Technol. 39 (15), 5787–5795.

Gossett, J.M., 2010. Sustained aerobic oxidation of vinyl chloride at low oxygenconcentrations. Environ. Sci. Technol. 44 (4), 1405–1411.

Haas, B.J., Gevers, D., Earl, A.M., Feldgarden, M., Ward, D.V., Giannoukos, G.,Ciulla, D., Tabbaa, D., Highlander, S.K., Sodergren, E., Methé, B., DeSantis,T.Z., Consortium, T.H.M., Petrosino, J.F., Knight, R., Birren, B.W., 2011.Chimeric 16S rRNA sequence formation and detection in Sanger and 454-pyrosequenced PCR amplicons. Genome Res. 21 (3), 494–504.

Han, Y.-S., Hyun, S.P., Jeong, H.Y., Hayes, K.F., 2012. Kinetic study of cis-dichloroethylene (cis-DCE) and vinyl chloride (VC) dechlorination usinggreen rusts formed under varying conditions. Water Res. 46 (19),6339–6350.

Hartmans, S., de Bont, J.A.M., Tramper, J., Luyben, K.C.A.M., 1985. Bacterialdegradation of vinyl chloride. Biotechnol. Lett. 7 (6), 383–388.

Hoelen, T.P., Reinhard, M., 2004. Complete biological dehalogenation ofchlorinated ethylenes in sulfate containing groundwater. Biodegradation15 (6), 395–403.

Holt, B.D., Sturchio, N.C., Abrajano, T.A., Heraty, L.J., 1997. Conversion ofchlorinated volatile organic compounds to carbon dioxide and methylchloride for isotopic analysis of carbon and chlorine. Anal. Chem. 69 (14),2727–2733.

Hunkeler, D., Van Breukelen, B.M., Elsner, M., 2009. Modeling chlorine isotopetrends during sequential transformation of chlorinated ethenes. Environ.Sci. Technol. 43 (17), 6750–6756.

Hunkeler, D., Elsner, M., Aelion, C.M., Hohener, P., Hunkeler, D., Aravena, R.,2010. Environmental Isotopes in Biodegradation and Bioremediation.

Hunkeler, D., Abe, Y., Broholm, M.M., Jeannottat, S., Westergaard, C., Jacobsen,C.S., Aravena, R., Bjerg, P.L., 2011. Assessing chlorinated ethene degradationin a large scale contaminant plume by dual carbon–chlorine isotopeanalysis and quantitative PCR. J. Contam. Hydrol. 119 (1–4), 69–79.

Hunkeler, D., Laier, T., Breider, F., Jacobsen, O.S., 2012. Demonstrating a naturalorigin of chloroform in groundwater using stable carbon isotopes. Environ.Sci. Technol. 46 (11), 6096–6101.

Jeong, H.Y., Anantharaman, K., Han, Y.-S., Hayes, K.F., 2011. Abiotic reductivedechlorination of cis-dichloroethylene by Fe species formedduring iron- orsulfate-reduction. Environ. Sci. Technol. 45 (12), 5186–5194.

Krajmalnik-Brown, R., Hölscher, T., Thomson, I.N., Saunders, F.M., Ritalahti, K.M.,Löffler, F.E., 2004. Genetic identification of a putative vinyl chloridereductase in dehalococcoides sp. strain BAV1. Appl. Environ. Microbiol. 70(10), 6347–6351.

Kuder, T., van Breukelen, B.M., Vanderford, M., Philp, P., 2013. 3D-CSIA: carbon,chlorine, and hydrogen isotope fractionation in transformation of TCE toethene by a dehalococcoides culture. Environ. Sci. Technol. 47 (17),9668–9677.

Lee, W., Batchelor, B., 2002a. Abiotic reductive dechlorination of chlorinatedethylenes by iron-bearing soil minerals. 1. Pyrite and magnetite. Environ.Sci. Technol. 36 (23), 5147–5154.

Lee, W., Batchelor, B., 2002b. Abiotic reductive dechlorination of chlorinatedethylenes by iron-bearing soilminerals. 2. Green rust. Environ. Sci. Technol.36 (24), 5348–5354.

Liang, X., Paul Philp, R., Butler, E.C., 2009. Kinetic and isotope analyses oftetrachloroethylene and trichloroethylene degradation by model Fe(II)-bearing minerals. Chemosphere 75 (1), 63–69.

Löffler, F.E., Ritalahti, K.M., Zinder, S.H., 2013. Dehalococcoides and reductivedechlorination of chlorinated solvents. In: Stroo, H.F., Leeson, A., Ward, C.H.(Eds.), SERDP ESTCP Environmental Remediation Technology Bioaugmen-tation for Groundwater Remediation. Springer, New York, NY, pp. 39–88.

Martins, P., Cleary, D.F.R., Pires, A.C.C., Rodrigues, A.M., Quintino, V., Calado, R.,Gomes, N.C.M., 2013. Molecular analysis of bacterial communities anddetection of potential pathogens in a recirculating aquaculture system forScophthalmus maximus and Solea senegalensis. PLoS ONE 8 (11), e80847.

McCormick,M.L., Adriaens, P., 2004. Carbon tetrachloride transformation on thesurface of nanoscale biogenic magnetite particles. Environ. Sci. Technol. 38(4), 1045–1053.

Meckenstock, R.U., Elsner,M., Griebler, C., Lueders, T., Stumpp, C., Dejonghe,W.,Bastiaens, L.L., Springael, D., Smolders, E., Boon, N., Agathos, S.N., Sorensen,S.R., Aamand, J., Albrechtsen, H.J., Bjerg, P.L., Schmidt, S., Huang, W.E., VanBreukelen, B.M., 2015. Biodegradation: updating the concepts of control formicrobial clean-up in contaminated aquifers. Environ. Sci. Technol.

Müller, J.A., Rosner, B.M., vonAbendroth, G., Meshulam-Simon, G.,McCarty, P.L.,Spormann, A.M., 2004. Molecular identification of the catabolic vinylchloride reductase from dehalococcoides sp. strain VS and its environmen-tal distribution. Appl. Environ. Microbiol. 70 (8), 4880–4888.

Newmark, R.L., Aines, R.D., 1995. Summary of the LLNL Gasoline SpillDemonstration - Dynamic Underground Stripping Project. U. S. D. o.Energy, Lawrence Livermore National Laboratory.

Novais, R.C., Thorstenson, Y.R., 2011. The evolution of Pyrosequencing® formicrobiology: from genes to genomes. J. Microbiol. Methods 86 (1), 1–7.

Paul, E.A.a.,C.,.F.E., 1996. Soil Microbiology and Biochemistry. San Diego, CA,USA.

Paul, L., Herrmann, S., BenderKoch, C., Philips, J., Smolders, E., 2013. Inhibition ofmicrobial trichloroethylene dechorination by Fe (III) reduction depends onFe mineralogy: a batch study using the bioaugmentation culture KB-1.Water Res. 47 (7), 2543–2554.

Pilloni, G., Granitsiotis, M.S., Engel, M., Lueders, T., 2012. Testing the limits of454 pyrotag sequencing: Reproducibility, quantitative assessment andcomparison to T-RFLP fingerprinting of aquifer microbes. PLoS ONE 7 (7),e40467.

Postma, D., Boesen, C., Kristiansen, H., Larsen, F., 1991. Nitrate reduction in anunconfined Sandy aquifer: water chemistry, reduction processes, andgeochemical modeling. Water Resour. Res. 27 (8), 1944–7973.

Reddy, C., Drenzek, N., Eglinton, T., Heraty, L., Sturchio, N., Shiner, V., 2002.Stable chlorine intramolecular kinetic isotope effects from the abioticdehydrochlorination of DDT. Environ. Sci. Pollut. Res. 9 (3), 183–186.

Renpenning, J., Keller, S., Cretnik, S., Shouakar-Stash, O., Elsner, M., Schubert, T.,Nijenhuis, I., 2014. Combined C and Cl isotope effects indicate differencesbetween corrinoids and enzyme (Sulfurospirillum multivorans PceA) inreductive dehalogenation of tetrachloroethene, but not trichloroethene.Environ. Sci. Technol. 48 (20), 11837–11845.

Scheutz, C., Durant, N.D., Dennis, P., Hansen, M.H., Jørgensen, T., Jakobsen, R.,Cox, E.E., Bjerg, P.L., 2008. Concurrent ethene generation and growth ofdehalococcoides containing vinyl chloride reductive dehalogenase genesduring an enhanced reductive dechlorination field demonstration. Environ.Sci. Technol. 42 (24), 9302–9309.

18

Shani, N., Rossi, P., Holliger, C., 2013. Correlations between environmentalvariables and bacterial community structures suggest Fe(III) and vinylchloride reduction as antagonistic terminal electron-accepting processes.Environ. Sci. Technol. 47 (13), 6836–6845.

Shouakar-Stash, O., Drimmie, R.J., Zhang, M., Frape, S.K., 2006. Compound-specific chlorine isotope ratios of TCE, PCE and DCE isomers by directinjection using CF-IRMS. Appl. Geochem. 21 (5), 766–781.

Slater, G.F., Lollar, B.S., Sleep, B.E., Edwards, E.A., 2001. Variability in carbonisotopic fractionation during biodegradation of chlorinated ethenes:implications for field applications. Environ. Sci. Technol. 35 (5), 901–907.

Sleep, B.E., Ma, Y., 1997. Thermal variation of organic fluid properties andimpact on thermal remediation feasibility. J. Soil Contam. 6 (3), 281–306.

Smits, T.H.M., Assal, A., Hunkeler, D., Holliger, C., 2011. Anaerobic degradation ofvinyl chloride in aquifer microcosms. J. Environ. Qual. 40, 915–922.

Tiehm, A., Schmidt, K.R., 2011. Sequential anaerobic/aerobic biodegradation ofchloroethenes—aspects of field application. Curr. Opin. Biotechnol. 22 (3),415–421.

Tobiszewski, M., Namieśnik, J., 2012. Abiotic degradation of chlorinated ethanesand ethenes in water. Environ. Sci. Pollut. Res. Int. 19 (6), 1994–2006.

van der Zaan, B., Hannes, F., Hoekstra, N., Rijnaarts, H., de Vos, W.M., Smidt, H.,Gerritse, J., 2010. Correlation of dehalococcoides 16S rRNA andchloroethene-reductive dehalogenase genes with geochemical conditionsin ch