C�eline Roux a, b, *, Corinne Le Gal La Salle a, c, Caroline Simonucci b, Nathalie Van Meir b,L. Keith Fifield d, ASTER Teama, 1Olivier Diez b, Sylvain Bassot b, Roland Simler e,Dmitri Bugai f, Valery Kashparov g, Jo€el Lancelot a, c
a Aix-Marseille Universit�e, CNRS-IRD-Coll�ege de France, UM 34 CEREGE, Technopole de l'Environnement Arbois-M�editerran�ee, BP80, 13545 Aix-en-Provence,Franceb Institute for Radioprotection and Nuclear Safety, PRP-DGE/SRTG, BP 17, F-92262 Fontenay-aux-Roses, Francec Universit�e de Nîmes, Laboratoire de G�eochimie Isotopique (GIS), 150 rue George Besse, 30035 Nîmes, Franced Department of Nuclear Physics, Research School of Physics and Engineering, The Australian National University, ACT 0200, Australiae Laboratoire d'Hydrologie d'Avignon, UMR EMMAH 11144 INRA, Universit�e d'Avignon, 84000 Avignon, Francef Institute of Geological Sciences, 55-b, Gonchara Str., Kiev 01054, Ukraineg Ukrainian Institute of Agricultural Radiology, UIAR NUBiP of Ukraine, Mashinobudivnykiv str. 7, Chabany, Kyiv-Svjatoshin, Ukraine
a r t i c l e i n f o
Article history:Received 16 September 2013Received in revised form4 July 2014Accepted 7 July 2014Available online
After the explosion of the Chernobyl Nuclear Power Plant in April 1986, contaminated material wasburied in shallow trenches within the exclusion zone. A 90Sr plume was evidenced downgradient of oneof these trenches, trench T22. Due to its conservative properties, 36Cl is investigated here as a potentialtracer to determine the maximal extent of the contamination plume from the trench in groundwater.36Cl/Cl ratios measured in groundwater, trench soil water and leaf leachates are 1e5 orders of magnitudehigher than the theoretical natural 36Cl/Cl ratio. This contamination occurred after the Chernobyl ex-plosion and currently persists. Trench T22 acts as an obvious modern point source of 36Cl, however othersources have to be involved to explain such contamination. 36Cl contamination of groundwater can beexplained by dilution of trench soil water by uncontaminated water (rainwater or deep groundwater).With a plume extending further than that of 90Sr, radionuclide which is impacted by retention and decayprocesses, 36Cl can be considered as a suitable tracer of contamination from the trench in groundwaterprovided that modern release processes of 36Cl from trench soil are better characterized.
After the explosion at the Chernobyl Nuclear Power Plant inApril 1986, debris, contaminated organic matter and contaminatedsoils were buried in about 800 shallow trenches within the exclu-sion zone. This clean-up operation aimed at reducing radiationexposure and atmospheric remobilization of radionuclides. Thesetrenches were dug in a permeable sandy formation, favorable to
CNRS-IRD-Coll�ege de France,is-M�editerran�ee, BP80, 13545
radionuclide migration through soils and in groundwater. Tomonitor these migrations, one of these trenches, trench T22, waschosen as a pilot site (Chernobyl Pilot Site) since 1999, in the frameof a collaborative program between the French Institute forRadiological Protection and Nuclear Safety (IRSN), the UkrainianInstitute of Agricultural Radiology (UIAR-NUBiP) and the Instituteof Geological Sciences (IGS). Studies were intensified in 2008through the funding of the TRASSE research group (on RadionuclideTransfers to the Soil, the ground and the Ecosystems), based on acollaboration between IRSN and the French National Center forScientific Research (CNRS).
A 90Sr plume was shown in groundwater downgradient of thetrench, with volumetric activities reaching 1000 Bq L�1
(2 � 10�9 mmol L�1) 15 m from the trench (Dewiere et al., 2004).Because of the reactivity of strontium, 90Sr migration is delayed bysorption processes (Szenknect, 2003; Dewiere et al., 2004; Bugaiet al., 2012a). Less reactive radionuclides, not impacted by suchprocesses, may have migrated over longer distance.
Fig. 1. Location and map of the Chernobyl Pilot Site (CPS). Piezometer locations are symbolized by crosses. For more details about piezometers on the CD profile, see Fig. 2.
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e3220
Due to its conservative properties, chloride (Cl�) is not impactedby sorption or other watererock interaction processes in theaquifer and could be a suitable tracer of to characterize themaximalextent of radionuclide contamination from the trench. More spe-cifically, its radioisotope chlorine-36 (36Cl) is a product of nuclearactivity and has most likely been released during the Chernobylaccident: Chant et al. (1996) showed high 36Cl/Cl in lichens linkedto the accident in Ukraine, Belarus and Russia. Moreover, 36Cl is ofinterest at such a time scale as its radioactive decay is negligible(half-life: 3.01 � 105 ± 0.04 � 105 years; Endt and Van der Leun,1973). However, the investigation of the maximal extent of thecontaminant plume using 36Cl as a tracer might be complicated. Forinstance, 25 or so years after the explosion, 36Cl pulse linked to theChernobyl explosion might have migrated outside the studied areaor be so diluted that this pulse is not observable anymore.
This study aims to determine if 36Cl is a suitable tracer of thecontamination from trench T22 under non reactive conditionsthrough the characterization of 36Cl content in Chernobyl Pilot Sitegroundwater.
2. Study site settings
The Chernobyl Pilot Site (CPS) is located 2.5 km South-Westfrom the Chernobyl Nuclear Power Plant (ChNPP) (Fig. 1). Thesefacilities are installed on the first alluvial terrace and on the rightbank of the Prypiat River, located at the top of sedimentary for-mations covering the North-western slope of the Ukrainian shield(Matoshko et al., 2004). The shallow 30-m-thick aquifer iscomposed of two facies, a 4-m-thick homogeneous aeolian layerunderlain by a heterogeneous alluvial layer, both favorable to
Fig. 2. Piezometers located on the CD profile of the CPS.
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e32 21
radionuclides migration. The aeolian layer, with a predominance offine and medium sand fraction, is characterized by high perme-abilities of 3e5 m d�1 and a total porosity of 34e36% (Matoshkoet al., 2004). The silty and clayey alluvial layer exhibits perme-abilities between 0.01 and 1 m d�1 and a total porosity of 35e37%(Matoshko et al., 2004). The aeolian layer is mainly composed ofquartz, whereas the alluvial layer is composed of quartz with someKeNa feldspars and accessory heavy minerals (Matoshko et al.,2004).
The CPS is located in the Red Forest area of the Chernobylexclusion zone. Regional groundwater flows roughly to north-enorth-east in the direction of the Prypiat river (Fig. 1; Matoshkoet al., 2004; Bugai et al., 2012a). In the Red Forest area, ground-water 90Sr volumetric activity ranged from 100 to 20,500 Bq L�1
(2 � 10�10 and 4 � 10�8 mmol L�1, respectively) between 1993 and1995 (Dzhepo and Skal'skii, 2002).
Trench T22 is 70-m-long over 8e10-m-large and 2e2.5-m-deep.Its direction is north-west/south-east, perpendicular to the maingroundwater flow direction (Fig. 1) (Bugai et al., 2005). Thegroundwater table range from 1 to 4m-depth and the limit aeolian/alluvial formation is around 110 m.a.s.l (Bugai et al., 2012a).
Roughly 100 piezometers were installed in the vicinity of thetrench (Fig. 1). Most of them were organized along two lines,called hereafter AB profile and CD profile (Fig. 1). Most of pie-zometers are made of PVC tubes with a diameter of 2.5 cm, a20 cm long stainless steel mesh screens and isolated by a 20 cmthick bentonite plug, in order to avoid cross-contamination be-tween piezometers (Dewiere et al., 2004). Clusters of 4 piezome-ters are organized to sample groundwater at the same spot at fourdifferent depths: the screen of the first piezometer is located at4 m depth in the aeolian layer, the screen of the second at theinterface between the aeolian and alluvial layers (around 5 mdepth), and the screens of the last two piezometers are in thealluvial layer, at 6 and 8 m depth respectively. At the end of the ABprofile, three piezometers (1-98-1, 1-98-2 and 1-98-3) wereinstalled to sample groundwater at greater depth, respectively at10, 20 and 30 m-depth. The screens of these last piezometers werenot isolated by a bentonite plug. Some piezometers were equippedwith TD-diver sensors to record groundwater table variations (VanEssen Instruments, The Netherlands).
A weather station was also set up to collect precipitations data,wind direction, wind speed, hours of sunshine, soil and air tem-peratures and air humidity. These data allowed to characterize therelationship between precipitations and groundwater level: arainwater infiltration rate of 200e300 mm y�1 was derived, cor-responding to 30e50% of the annual precipitations (Bugai et al.,
2012a,b). Porous cups were installed in the trench to collect soilsolution. A field laboratory was set up for sample filtration, on sitemeasurements (such as alkalinity, sulfur concentrations …) andmaterial storage.
3. Sampling and analyses
3.1. Sampling strategy
The preliminary investigations were carried out in October 2008and October 2009 and a last sampling campaign was performed inMay 2011 at the CPS in order to obtain a complete dataset of the 36Clcontent in groundwater along the CD-profile (Fig. 2).
In the early 2000s, tracer tests using 36Cl were performeddowngradient of trench T22, in the vicinity of the AB-profile (Fig.1),to determine groundwater velocity (1999e2001, Bugai et al.,2012b). 36Cl was chosen because of its conservative properties,because a small input mass was needed, analyses were easy tomake and it was also easily available (D. Bugai pers. comm.; seeBugai et al., 2012b for more details). Hence, in order to avoid any36Cl contamination from these tracer tests, most groundwatersamples were collected along the CD profile, further from the in-jection test (Figs. 1 and 2). Indeed the injected 36Cl cannot reach theCD profile because of the northenorth-east groundwater flowdirection.
Deeper groundwater samples were collected in several pointsalthough, none of them were located on the CD profile. Threeparticularly interesting deep piezometers (1-98-1, 1-98-2, 1-98-3piezometers) were located on the AB profile (Fig. 1). The absence ofcontamination from the tracer tests in groundwater sampled inthese three piezometers is assumed considering the main flowdirection, the apparent ages of groundwater of the two deepestpiezometers (older than 33 years, Le Gal La Salle et al., 2012) andthe 36Cl content in groundwater sampled in another 10-m-depthpiezometer located farther east from the injection point (1-01piezometer).
In addition, in May 2011, to better identify 36Cl origins, samplesfrom leaves (birch located on the trench) were collected and atrench soil water sample from a porous ceramic cup located 1.25 mdeep into the trench along the CD profile was also analyzed. Tocomplete the sampling, outside samples were collected: ground-water located upgradient the CPS, in the vicinity of waste mounds(IGS33 piezometer) and water of the Borschi River, draining adifferent neighboring small watershed just south of the CPS (Freedet al., 2004) were sampled for comparison (Fig. 1).
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e3222
3.2. Sampling procedures
During preliminary investigations carried out in October 2008and October 2009, groundwater samples were collected in a total of10 piezometers. In May 2011, 21 groundwater samples werecollected at the CPS along the CD-profile (Fig. 2).
Groundwater was sampled using a peristaltic pump. The sam-pling was carried out after pumping between 3 and 5 L of water topurge the piezometers and to clean the pumping device. Then,1.5e2 L of water were collected and filtered. 200 mL of the filteredsample water were divided into 125 mL polyethylene bottles, onefor anion analysis (including chloride) and the other for cationanalysis (including 90Sr volumetric activity analyses). Bottles forcation analysis were acidified at pH lower than 2.5 with concen-trated HNO3. For 36Cl/Cl analyses, between 1 and 8 L were pumpedand conditioned in polyethylene bottles. During the 2008campaign, samples for 36Cl/Cl analyses were preconcentrated onsite by evaporation to reduce sample volume by a factor 6. In May2011, given the high 36Cl concentrations measured in October 2008and October 2009, the sample volume was reduced to 250 mL (seebelow for the applied protocol, Section 3.3.3).
The trench soil water was sampled from porous cups using amanual vacuum pump. On the CD profile, trench soil water can becollected at several depths. Only the most contaminated with 90Sr,located at 1.25-m-depth, was analyzed in 36Cl. Water was filteredand conditioned in two 125 mL polyethylene bottles for anion andcation analysis. As before, the cation bottle was also acidified at pHbelow 2.5. 6 mL of the bottle reserved for anion analysis were usedfor 36Cl/Cl analysis.
Birch leaf samples were taken on the edge of the CD profileabove trench T22. A dose rate of 15 mSv h�1 was measured oncontact with a radiometer. Two different extraction protocols weretested. In the first protocol, approximately 350 raw leaves (about28 g) were crushed and leached with distillated water. In the sec-ond protocol, about 90 g of leaves were burnt in an oven at 400 �Cuntil reduction to ashes. Ashes were then leached with distilledwater. Then, both solutions were acidified to pH 4. In order toremove any radioactive cation from the solution, more particularly90Sr, the leachate was filtered through DOWEX resin (DOW inc.).The same resin was used for both leachates: crushed leaf leachatewas filtered first, then the resin was cleaned with HCl and finally,the ashed leaf leachate was filtered. Leachates were then condi-tioned in 250 mL bottles. Hereafter, the sample prepared followingthe first protocol is named TCV 1 and sample prepared followingthe second protocol is named TCV 2 (for TChernobyl Vegetation 1and 2).
The groundwater sampled outside the CPS (IGS33 piezometer)was collected using the manual peristaltic pump and after a purgeof the piezometer. The sample was then filtered and conditioned intwo 125 mL polyethylene bottles for anion and cation analysis. Thecation bottle was acidified at pH lower than 2.5 too. For 36Cl/Clanalysis, 250 mL of sample were conditioned separately.
Borschi river water was sampled, filtered and conditioned in125 mL polyethylene bottles for anion and cation analysis. Again,the cation bottle was acidified at pH lower than 2.5. For 36Cl/Clanalysis, a volume of 6 mL was taken from the 125 mL bottlereserved for anion analysis.
3.3. Analyses
3.3.1. Strontium-90 analyses90Sr volumetric activities were measured at the Analysis and
Experimental facilities Laboratory (IRSN) by liquid scintillation:10 mL of samples were mixed with 10 mL of Ultimagold AB andwere then analyzed on Liquid Scintillation Analyzer TRI-CARB 3170
TR/SL (Perkin Elmer Instruments, Courtaboeuf, France). Countingtime was a maximum of 240 min but it was reduced for sampleswith high 90Sr content. Detection limit was 0.03 or so Bq L�1. Forsamples collected in October 2008 and 2009, no quenchingcorrection was applied because quenching effect was supposed tonot imply a big variation (TSie between 250 and 350). For samplescollected in May 2011, a quenching correction was applied ac-cording to measurements of 90Sr standards with the same activitybut different quenching values.
In order to compare 90Sr contamination with 36Cl contamina-tion, 90Sr concentration ([90Sr]) are corrected by the radioactivedecay and calculated inmmol L�1 from these activities, according tothe following equation:
½90Sr� ¼ A90Sr � 1000� 365:25� 24� 3600l90Sr � NA
(1)
where A90Sr is the90Sr volumetric activity (Bq L�1), l90Sr the constant
of 90Sr radioactive decay, equal to 0.024 y�1, and NA the Avogadronumber 6.022 � 1023 mol�1.
3.3.2. Chloride analysesChloride concentrations ([Cl�]) were measured by ion chroma-
tography also at the Analysis and Experimental facilities Laboratoryon a Compact IC 861 Metrohm Ion Chromatograph (MetrohmFrance Inc, Villebon-rue-Yvette, France.). Three measurementswere performed on each sample to assess repeatability. Resultinganalytical uncertainties range between 0.0002 and 0.04 mmol L�1
(Tables 1, 2 and 4).
3.3.3. Chlorine-36 analyses
3.3.3.1. Australian National University measurements. The October2008 and October 2009 samples were analyzed by AcceleratorMass Spectrometry (AMS) at the Department of Nuclear Physics atthe Australian National University (ANU). The protocol used forsample preparation was adapted from Conard et al. (1986).
Before sample preparation, all laboratory ware (tubes, beakers)was cleaned in a bath of HNO3 65% solution diluted with tri-distilled water. Between 1 and 6 L of each sample were evapo-rated to approximately 300 mL in 1 L glass beakers. Two blanks ofWeeks Island halite standard were prepared in parallel. Chlorinewas precipitated as AgCl by adding an excess of AgNO3 (relative tochloride concentration). The precipitate was left to settle overnightand most of the liquid phase was removed. The precipitate wasthen dissolved by adding a fewmilliliters of an ammonia solution at25%. A fewmilliliters of a saturated solution of BaNO3 aq were addedto precipitate sulfate as barium sulfate, in order to minimize the 36Scontent of the sample and the solution was manually agitated. Thesolution was then filtered to remove the barium sulfate precipitate.This step was repeated until no more precipitate was observable.The filtrate was poured into a centrifuge tube and AgCl reprecipi-tated by adding HNO3 65%. The tube was then centrifuged, thesupernatant poured off, and the precipitate rinsed with tri-distilledwater and centrifuged again. This operationwas done twice. Finally,the precipitate was dried at 50e60 �C, with the tube wrapped in analuminum foil to protect the precipitate from direct light.
90Sr volumetric activities were measured in each removed su-pernatant of the most contaminated sample (4-00 piezometer) inorder to verify that most of the 90Sr was removed from the AgClprecipitate.
As these first measurements revealed unexpectedly high 36Cl/Clratios, approaching 10�8 at at�1 for the most 36Cl-concentratedsample, counting times were reduced to prevent contamination ofthe AMS ion source (counting times between 10 and 30 s). Large
Table 1[Cl�] concentrations, 90Sr volumetric activity, 36Cl/Cl measured ratios and laboratory where 36Cl/Cl were measured for samples collected in October 2008 and October 2009 atthe CPS (from upgradient to downgradient).
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e32 23
uncertainties resulted from these very short counting times. Thetwo blanks of Weeks Island halite were analyzed after the high-level samples, and gave 36Cl/Cl ratios of 1.4 � 10�13 and1.8� 10�13 at at�1. These high values were two orders of magnitudehigher than typical blanks, and resulted from cross-contaminationof the high-level samples in the ion source. The confusion between36S and 36Cl signals is not probable and cannot explain such ratiosbecause of the good distinction between 36Cl and 36S signals duringthe counting at AMS of the Australian National University (Fifieldet al., 2013).
3.3.3.2. ASTER measurements. To confirm the measurements andreduce the uncertainties resulting from the short counting time,tests were carried out with samples from the October 2008 fieldcampaign diluted with halite solutions of known chloride con-centrations (measured by ion chromatography in the hydrogeologylaboratory of Avignon University, LHA) and very low 36Cl/Cl ratios(close to 10�15 at at�1). Three mixing solutions were made andanalyzed on the French Accelerator Mass Spectrometer nationalfacility, ASTER, located at CEREGE (Aix-en-Provence). The aim ofthese isotopic dilutions was to optimize the counting time,analyzing at the first the most diluted sample (dilution A) andfinally the least diluted (dilution C). Based on the ratios obtainedfromOctober 2008 test analyses, these dilutions were done to reachin the diluted sample 36Cl/Cl ratios close to 10�12, 10�13 and5 � 10�14 at at�1, for each dilution respectively. 36Cl/Cl ratios in theoriginal sample are calculated and reported following the applieddilution in Tables 2e4. The 36Cl/Cl ratio in the original sample wasderived from the following equation:
�36ClCl
�measured
¼36Clsample þ 36ClhaliteClsample þ Clhalite
(2)
where 36Clsample and 36Clhalite were the amount in atoms of 36Cl inthe sample and in the halite solution, respectively and Clsample andClhalite were the amount in atoms of Cl in the sample and in thehalite solution, respectively.
The sample 36Cl/Cl ratio was written as (Equation (3)):
�36ClCl
�sample
¼�36Cl
Cl
�measured
� Clsample þ ClhaliteClsample
��36Cl
Cl
�halite
� ClhaliteClsample
(3)
For these tests, 36Cl/Cl ratios were calculated from measured36Cl/35Cl ratio and considering the average natural 35Cl/Cl ratio of0.7575 (Rosman and Taylor, 1998), 36Cl/35Cl was measured there-after. Calculated uncertainties on 36Cl/Cl ratios took into accountthe measurement uncertainties as well as the natural variation ofthe 37Cl/35Cl ratio (Rosman and Taylor, 1998). The 36Cl/Cl ratio ofhalite solutions were analyzed and showed as expected very low36Cl/Cl ratios of 5 � 10�16 ± 1 � 10�16 and2.1 � 10�15 ± 0.7 � 10�15 at at�1. The isotopic dilution with thesehalite solutions was tested on groundwater samples showing thelowest and the highest 36Cl/Cl ratios, collected in October 2008:groundwater sampled in piezometers 1-98-2 and 4-00, respec-tively. These samples were analyzed both at ANU and on ASTER andled to coherent values (Table 1).
Based on these results, the samples from May 2011 werehandled in a similar way: three isotopic dilutions were carried outfor each sample collected during the May 2011 sampling campaignand analyzed at ASTER. Dilutions were performed with two newsolutions showing two different concentrations in halite standard(NIST SRM 999b), of which [Cl�] were analyzed at the Analysis andExperimental facilities Laboratory (IRSN) and prepared as blanksfor ASTER analyses. 36Cl/Cl measured in these standard solutionswere similar to the previous halite solution 36Cl/Cl ratios:5�10�16 ± 2�10�16 and 2.4�10�15 ± 0.9� 10�15. All 36Cl/Cl ratiosdirectly measured are showed in appendix. Only 36Cl/Cl ratioscalculated based on the considered dilution are discussed hereafter.
36Cl concentrations ([36Cl]) are calculated from the inferredgroundwater 36Cl/Cl ratio and the measured [Cl�], in mmol L�1,using the following equation:
½36Cl� ¼�36Cl
Cl
�calculated
�hCl�
i(4)
4. Results
Results of the analyzes are presented in Tables 1e4, according tothe type of sample, time and location of sampling.
Table 1 shows sampled piezometers, piezometer screen eleva-tion, Cl� concentrations, 90Sr concentrations, measured 36Cl/Cl ra-tios and calculated 36Cl concentrations for samples collected inOctober 2008 and October 2009.
May 2011 sample results are shown in Tables 2e4. These tablespresent 36Cl/Cl ratios measured in the diluted sample and correctedfrom the dilution, [Cl�] and [90Sr]. Results for samples collected
Table 2[Cl�] concentrations, 36Cl/Cl ratios measured on ASTER (CEREGE) for each isotopic dilution of samples collected in May 2011 outside the CPS.
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e3224
outside the CPS are shown in Table 2. Calculated 36Cl/Cl ratios of thetwo vegetal samples are given in Table 3. Table 4 presents results ofgroundwater samples collected at the CPS.
90Sr concentrations range from below the detection limit(around 10�12 mmol L�1) to 4 � 10�8 mmol L�1 (Fig. 4). The highestvalue is measured in the trench soil water sampled in May 2011. Ingroundwater, the highest values are measured in shallow pie-zometers located downgradient of trench T22 (Fig. 4). However,groundwater sampled in the most upgradient piezometer (6-99)also shows some contamination in [90Sr] up to 10�10 mmol L�1.[90Sr] are lower in deeper groundwater (alluvial layer) and ingroundwater just upgradient of the trench (1-06-1, 1-06-2 pie-zometers). In most of the deepest groundwater, [90Sr] are below thedetection limit (1-01, 1-98-2, 1-98-3 piezometers) except forgroundwater sample collected in the 1-98-1 piezometer whichshows a [90Sr] of 5 � 10�12 mmol L�1.
With respect to Cl, the highest [Cl�] is measured in the trenchsoil water sample collected in May 2011 on the CD profile, with avalue of 0.35 mmol L�1. In groundwater sampled on the CD profile,most of [Cl�] is one order of magnitude lower, ranging from 0.008to 0.035 mmol L�1. The highest [Cl�] are observed downgradient oftrench T22 or in the most upgradient groundwater sample (6-99piezometer) also contaminated in 90Sr. The remaining samplesshow [Cl�] ranging between 0.01 and 0.02 mmol L�1. Deepgroundwater, collected in AB profile and 1-01 piezometers, showshigher [Cl�], reaching a maximum of 0.058 ± 0.006 mmol L�1,measured in the deep 1-98-2 piezometer in May 2011. All observed[Cl�] in the groundwater remainwithin the range of meteoric water(0.006e0.07 mmol L�1, Bugai et al., 2012a). In contrast, watersampled in the Borschi river shows a higher [Cl�] of0.17 ± 0.02 mmol L�1 (Table 2).
36Cl/Cl ratios and [36Cl] range from 2 � 10�12 to 1 �10�8 at at�1
and 6 � 10�14 to 4 � 10�9 mmol L�1, respectively. 36Cl/Cl ratios arereported in �10�10 at at�1 along CD profile and with depth (ABprofile), in Fig. 3. The highest 36Cl contents are observed in the soilwater of the trench (Table 4) as well as in leaf leachates. Ground-water sampled at the CPS show 36Cl/Cl ratios at least one order ofmagnitude lower, ranging from 2.4 � 10�12 ± 1.7 � 10�12 to4 � 10�9 ± 2 � 10�9 at at�1 (from 6 � 10�14 ± 4 � 10�14 to1 � 10�10 ± 0.4 � 10�10 mmol L�1 for [36Cl]). In groundwater, thehighest values aremeasured in shallow groundwater downgradientof the trench (aeolian layer and aeolian/alluvial interface) (Fig. 3). Inother groundwater samples, collected along the CD profile, 36Cl/Clratios are close to 10�10 at at�1 (8 � 10�12 mmol L�1 for [36Cl]),respectively. Deeper, 36Cl/Cl ratios decrease: at 20-m-depth (1-98-2piezometer), 36Cl/Cl ratio in groundwater is close to 10�11 at at�1
Table 336Cl/Cl ratios measured on ASTER (CEREGE) for each isotopic dilution of vegetalsamples collected in May 2011.
with [36Cl] of 1�10�12 mmol L�1 (October 2008 andMay 2011) andat 30-m-depth (1-98-3 piezometer), 36Cl/Cl ratio is of2.4� 10�12 ± 1.7 � 10�12 at at�1 and [36Cl] is of 6� 10�14 mmol L�1.Groundwater sampled in the IGS33 piezometer (outside the CPS)and in the Borschi river show 36Cl/Cl ratios close to 10�11 at at�1.Borschi river shows [36Cl] close to 1 � 10�12 mmol L�1 whereas[36Cl] in groundwater sampled in IGS33 piezometer is two orders ofmagnitude lower (3 � 10�13 mmol L�1).
In summary, three observations can be made. First, both [90Sr],[Cl�] and 36Cl/Cl ratios increase downgradient of the trench. Sec-ondly, the most upgradient groundwater in the aeolian layer alsoshows high [90Sr] and [36Cl]. Finally, [90Sr] and [36Cl] decrease withincreasing depth, whereas [Cl�] increase slightly with respect to[Cl�] observed in the less contaminated groundwater.
5. Discussion
5.1. Origins of 36Cl in Chernobyl pilot site groundwater
Most of the water samples show particularly high 36Cl/Cl ratios,with several orders of magnitude between the lowest and thehighest ratios. Several sources may contribute to this high contentin 36Cl.
At the CPS, groundwater recharge is mainly due to infiltration ofmeteoric water (Bugai et al., 2012b; Le Gal La Salle et al., 2012). 36Cl/Cl in rainwater was not measured nevertheless, it can be estimatedaccording to cosmogenic 36Cl deposition models. Commonly, innatural meteoric water, the occurrence of 36Cl results from atmo-spheric production by cosmic ray spallation of 40Ar, 40Ar (n,p4n)36Cl, or by 36Ar neutron capture, 36Ar (n,p) 36Cl (Bentley et al., 1986;Phillips, 2000). One of the most recent deposition model wasproposed by Phillips (2000). According to this model, the cosmo-genic production can be estimated at 45 at m2 s�1 at Chernobyllatitude, which are in agreement with two corrected values ob-tained at this latitude (Phillips, 2000). The average [Cl�] in meteoricwater is 0.0209 mmol L�1 (Bugai et al., 2012a). The precipitationamount is considered 625 mm y�1, (c.f. Section 2 and Bugai et al.,2012b), the theoretical natural 36Cl/Cl ratio in rainwater should beclose to:
36ClCl
¼36Cl deposited over a year
Precipiations� 1�
Cl��meteoric water � NA
¼ 2� 10�13 at at�1
(5)
NA is Avrogado's number.This value is 1e4 orders of magnitude lower than the 36Cl/Cl
ratios measured in the CPS groundwater and cannot thereforeexplain the observed ratios. Even considering a higher depositionrate (such as 60 at m2 s�1) or a lower precipitation amount (such as500 mm y�1, which is still possible at the CPS), the theoretical 36Cl/Cl ratio in rainwater stays close to 2 � 10�13 at at�1.
There are other natural productions of 36Cl which couldcontribute to the 36Cl content in CPS groundwater. For instance, in
Table 4[Cl�] concentrations. [90Sr] concentrations. 36Cl/Cl ratios measured on ASTER (CEREGE) for each isotopic dilution of samples collected in May 2011 at the CPS.
Sampledpiezometer(profile)
Mean screenaltitude (m)
[Cl�](mmol L�1)
[90Sr] (mmol L�1) Calculated 36Cl/Cl [36Cl]
Dilution A (at at�1) Dilution B (at at�1) Dilution C (at at�1) (Dilution) (mmol L�1)
Fig. 3. 36Cl content in groundwater along the cross-section at the CPS in October 2008,October 2009 and May 2011. Values are given near the sampled piezometerin �10�10 at at�1. Note that the x-scale is oversized in comparison with y-scale.Groundwater from the 3-02-2 piezometer was sampled in October 2008 and October2009, this is why two values are reported.
Fig. 4. 90Sr content in groundwater along the cross-section at the CPS in October 2008,October 2009 and May 2011. Values are given near the sampled piezometerin �10�10 mmol L�1. Note that the x-scale is oversized in comparison with y-scale.Groundwater from the 3-02-2 piezometer was sampled in October 2008 and October2009, this is why two values are reported.
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e3226
the upper section of the lithosphere, the main reaction is neutronactivation of 35Cl, with a neutron flux production coming fromnatural uranium and thorium decays in rocks (Bentley et al., 1986).In groundwater at the vicinity of uranium ores, where naturalneutron flux are high, the highest 36Cl/Cl ratios reported from theliterature data are close to 10�11 at at�1 (Cornett et al., 1996;Songsheng et al., 1994). As the highest 36Cl/Cl ratios in CPSgroundwater reaches 10�9 at at�1, this natural contribution can beassumed to be quite low. Therefore, 36Cl at the CPS cannot originatefrom natural processes only.
5.1.1. Chernobyl explosion and trenches set upAnthropogenic contamination has to be considered to explain
such high 36Cl/Cl ratios (up to 10�9 at at�1). At the CPS, the mostobvious potential source is the release of radionuclides during theexplosion of Unit 4 in April 1986. In nuclear facilities, 36Cl is pro-duced by neutron activation of 35Cl (n,g) 36Cl or 36Ar (n,p) 36Cl, and
high contents were measured in reactors, concrete shielding,graphite rings, etc. (Bentley et al., 1986; Bessho et al., 2007;Bondar'kov et al., 2009; Hou et al., 2007; Milton et al., 1994;Phillips, 2000). Hence, the explosion of Chernobyl Unit 4 is mostlikely to have released 36Cl, as shown by the high 36Cl/Cl ratios from10�10 to 10�12 at at�1 measured in lichens in Ukraine, Belarus andthe Russian Federation after 1986 (Chant et al., 1996). It suggests adiffuse contamination in 36Cl over a large area.
Apparent ages of shallow groundwater measured at the CPS onthe AB profile in October 2008 range from a few years to 20e25years at 8 m depth, which implies an infiltration after 1986 (Le GalLa Salle et al., 2012). Assuming that groundwater at similar depthon the CD profile infiltrated during the same period, observed 36Clcontamination of groundwater occurred after the reactor explosionin 1986. The timing is also coherent with the trench set up in1987e1988.
Fig. 5. 36Cl/Cl ratios, [Cl�] and potential mixing processes at the CPS. Groundwater samples of CPS are represented by empty symbols. May 2011 leaf leachates are represented bytwo dark cross on the y axis as [Cl�] is not known, May 2011 trench soil water sample by a full square and May 2011 samples collected outside the CPS (river and groundwater fromIGS33 piezometer) by a dark cross and a full triangle, respectively. The theoretical natural rainwater range is represented by three points linked by a dark line, each point rep-resenting minimal, mean and maximum measured [Cl�] in rainwater (on 30 samples collected during the 2005e2006 period, Bugai et al., 2012a). The rainwater 36Cl/Cl ratio wascalculated earlier (Section 5.1) and is supposed constant for different [Cl�]. Potential mixing processes are shown by the dark lines (A) and the three dotted lines (B, C and D).
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e32 27
5.1.2. 36Cl contamination of recent groundwaterGroundwater which shows the highest 36Cl contamination was
apparently recharged recently, over the last 10 years (Le Gal La Salleet al., 2012). This implies that a source of 36Cl still persists.
In groundwater from the aeolian layer, 36Cl/Cl ratios increase bymore than a factor 6 from upgradient to downgradient of the trench(Fig. 3). Consequently, it can be assumed that trench T22 acts as apoint source and keeps releasing 36Cl. The processes involved inthis release could be linked to alteration or leaching of buriedmaterial with which 36Cl is associated. Indeed, 36Cl may be asso-ciated with debris of explosion as 36Cl was reported in corematerialof Chernobyl Nuclear Power Plant unit 2 (Bondar'kov et al., 2009) orwith buried organic matter (originating from degradation of treetrunks, litter, …) as high 36Cl/Cl ratios were shown in lichens afterChernobyl accident (Chant et al., 1996).
The retention of anthropic 36Cl in soils and biosphere is sup-ported by the 36Cl/Cl ratios measured in the trench soil watersample and leaf leachates collected on the top of the trench whichare at least one order of magnitude higher than 36Cl/Cl ratios ingroundwater.
Cornett et al. (1997) estimated a residence time of at least 25years for Cl in the terrestrial biosphere of the Eastern Ontario(Canada), increasing with the increasing biota. Studies of 36Cltransfer in the soil-plant system, carried out in the Chernobylexclusion zone, have shown retention of 36Cl by live biota and ex-changes are characterized by quick transfers from the soil solutionto plants and from dry vegetation to soil solution (Kashparov et al.,2005, 2007). Involved retention processes are atmosphere recy-cling, adsorption on soil, vegetation and microbial uptakes andemission under organochloride forms (Milton et al., 2003;Bastviken et al., 2006, 2007). Lee et al. (2001) showed the affinityof chlorine with low molecular weight humic substances and
suggested that chloride might not be the predominant species ofchlorine in soil, particularly because of microbial activity. Therelease of chlorine from this storage to groundwater is more or lessimportant, depending on concentration, soil moisture, biotic ac-tivity (Kashparov et al., 2005; Ashworth and Shaw, 2006; Bastvikenet al., 2006, 2007). For instance, this biogeochemical cycle of Cl ismentioned several times to explain some delay of 36Cl bomb-pulsemigration in groundwater, particularly relatively to 3H (Cornettet al., 1997; Milton et al., 2003).
It could explain the 36Cl contamination persisting in recentgroundwater: 36Cl could be trapped in the trench in organochlorineform and be gradually released.
High 36Cl contamination is also observed upgradient of thetrench (6-99 piezometer). Two hypotheses can be considered: 36Clmigration from upgradient trenches (the closest upgradient trenchis less than 100 m away according to Antropov et al., 2001) and/orleaching of potential residual contamination in soils, still heredespite of the cleanup procedures. The duration of these releasesmight have been extended due to potential recycling of chlorine insoil and/or biosphere too (Cornett et al., 1997). Another potentialsource of 36Cl can be supposed to explain the current 36Clcontamination of groundwater: the neutron flux originating fromburied fuel particles, which could interact with 35Cl and produce36Cl. According to equations published by Songsheng et al. (1994),the required neutron flux to obtain such 36Cl/Cl ratio can be esti-mated following (Equation (6)):
Fn ¼36ClCl
� l36
s35� NClN35
(6)
where 36Cl/Cl is the measured ratio, l36 is the decay constant of 36Cl(2.3 � 10�6 y�1; Songsheng et al., 1994), s35 the thermal neutron
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e3228
capture cross section of 35Cl (44 � 10�24 cm2; Songsheng et al.,1994) and N35/NCl the proportion of 35Cl in total Cl (0.7577;Rosman and Taylor, 1998)
At the CPS, a neutron flux of 4� 108 N/cm2 would be required toproduce alone the observed 36Cl/Cl ratio of 10�8 at at�1 (order ofmagnitude in trench soil water).
However, it is difficult to assess the possibility of such neutronflux in CPS soil because of the lack in the knowledge of transuranicelements in the trench and soils. The last inventory of the trenchwas performed in 2000 and the estimation on 238Pu, 239,240Pu and244Cm were 4.3, 2.4, 0.44 GBq, respectively (Levchuk et al., 2012).Uranium content is not known, particularly taking into account itspotential migration linked to dissolution of fuel particles(Kashparov et al., 2009, 2012; Van Meir et al., 2009).
Rq FP fission 36Cl > fission in soil?
5.1.3. Deep groundwater contaminationThe deeper piezometers (1-98-1, 1-01, 1-98-2 and 1-98-3 pie-
zometers) also show some contamination by 36Cl.Potential contamination by the tracer test carried out between
1999 and 2001 with 36Cl has already been discarded (Section 3.1).The apparent ages in the two deepest piezometers, older than
33 years old (Le Gal La Salle et al., 2012), suggest that 36Clcontamination would result either from the 36Cl ‘bomb-pulse’generated by nuclear tests between 1952 and 1971 (Bentley et al.,1982; Elmore et al., 1982; Finkel et al., 1980) or from the Cher-nobyl Nuclear Power Plant activity before the explosion (whichstarted in 1977, Shestopalov, 2002). Indeed, 36Cl contaminationwasreported in the environment around nuclear facilities (Milton et al.,1994; Seki et al., 2007). However, estimation cannot be madebecause the thermonuclear tests impact in the area and 36Cl re-leases by RBMK reactors are not known.
Moreover, cautions should be taken when interpreting 1-98-1,1-98-2 and 1-98-3 piezometers data because, unlike in the otherpiezometers, screens are not isolated by a plug of bentonite.Moreover, some mixing processes were envisaged in the deepgroundwater (Le Gal La Salle et al., 2012).
5.2. Transport processes
The study of 36Cl content compared to other elements shouldallow characterizing 36Cl contamination processes in groundwater.
5.2.1. Chlorine-36 and mixing processesAdvection is assumed to be the main transport process in CPS
groundwater however, the influence of mixing processes ingroundwater has to be investigated. 36Cl/Cl ratios are comparedwith the inverse of [Cl�], measured in groundwater, trench soilwater, river water and leaf leachate samples, collected in October2008, October 2009 and May 2011 (Fig. 5). Samples are representedaccording to their nature and the depth of sampling. Mixing linesare drawn based on the following equation:
where Y is the considered element, x varies between 0 and 1, mix isthe solution resulting from the mixing process and End-member1and End-member2 are the initial solutions considered for themixing.
36Cl/Cl ratios resulting from mixing processes are calculateddividing [36Cl]mix by [Cl�]mix.
As noticed before (c.f. Section 4), the most contaminated sam-ples in 36Cl are trench soil water and leaf leachates with 36Cl/Clratios close to 10�8 at at�1. [Cl�] in trench soil water is also higherthan that of the rainwater range (0.006e0.07 mmol L�1). Ground-water sample [Cl�] concentrations remain within this range with36Cl/Cl ratio 1 to 4 orders of magnitude higher.
The overall trend described by the dataset is a decrease of 36Cl/Clratios and [Cl�] from the trench soil water sample to shallow al-luvial groundwater from 10�8 at at�1 to 2 � 10�12 at at�1 and0.35 mmol L�1 to 0.010 mmol L�1 in 36Cl/Cl and [Cl�], respectively.Samples collected in deeper alluvial layer shows lower 36Cl/Cl ratio(down to 10�12 at at�1) for [Cl�] in the same range of concentra-tions. To explain these trends, mixing processes are considered. Thefirst is the dilution of trench soil water by rainwater (lines A inFig. 5). The rainwater end-member is considered with a constant36Cl/Cl ratio but [Cl�] varies naturally. Thus, two mixing lines aredrawn, each ending at the lowest [Cl�] and the highest [Cl�]measured in rainwater according to Bugai et al. (2012a). All CPSgroundwater samples are between these two lines. Consequently,groundwater contamination may result of the mixing of trench soilwater with rainwater and the resulting [Cl�] varies depending onthe [Cl�] in rainwater.
The understanding of the dataset scattering can be refinedconsidering additional processes. Most of the samples collected inthe aeolian layer and some of the samples collected in the shallowalluvial layer are scattered between two end-members, 4-02-1 and1-06-1 piezometers, respectively, describing a decrease in 36Cl/Clratios by one order of magnitude from 4-02-1 piezometer to 1-06-1piezometer with [Cl�] changing by a factor 4. Piezometer 4-02-1 islocated just downgradient of the trench and its contamination ismost likely due to migrations from the trench. In contrast, 1-06-1 islocated just upgradient of the trench and is contaminated too,because of migrations from another source located upgradient oftrench 22, such as upgradient trenches or a quick local preferentialflow in the trench linked to trench body heterogeneity. Ground-water sampled in the 10-m-deep piezometers (1-01 and 1-98-1,Table 4) and far upgradient of the trench (6-99 piezometer) falls onthis trend and are among the most contaminated (Tables 1 and 4,Fig. 3).
In groundwater sampled in the shallow alluvial layer, [Cl�]scatter within the rainwater range, while 36Cl/Cl ratios remainsrelatively constant between 2 � 10�10 and 8 � 10�10 at at�1. 36Clcontaminations in groundwater sampled in the two deepest pie-zometers (1-98-2, 1-98-3 piezometers) are really different fromother CPS groundwater samples. Groundwater sampled at 20-m-depth (1-98-2 piezometer, Table 4) shows 36Cl/Cl ratios at least oneorder of magnitude lower than in the other samples and the highest[Cl�] (still within the rainwater range). Groundwater collected inthe deepest 1-98-3 piezometer is the least contaminated of all witha 36Cl/Cl ratio of 2.4 � 10�12 at at�1 and its [Cl�] is close to theaverage rainwater [Cl�]. Water sampled outside the site (ground-water from the IGS33 piezometer andwater from the Borschi River)show 36Cl/Cl ratios in the same order of magnitude than ground-water sampled on site at 20-m depth in the 1-98-2 piezometer.Note, [Cl�] in the Borschi river is higher than the rainwater range.
Several mixing lines are considered to fit with the datasettrends. Two lines represent the mixing of the most contaminatedsample in 36Cl (4-02-1) with two different rainwater end-members,the least [Cl�] concentrated and an adjusted rainwater end-member (lines B and C in Fig. 5). The mixing line between theadjusted rainwater end-member and the most contaminatedgroundwater in 36Cl sampled in the 4-02-1 piezometer fits partic-ularly well with most of the shallow groundwater samples, mostlycollected downgradient of trench T22. The contamination of sam-ples falling on these mixing lines could be the result of the mixing
Fig. 6. 90Sr and 36Cl concentrations and potential mixing processes. Groundwater samples of CPS are represented by empty symbols. May 2011 trench soil water sample is rep-resented by a full square and May 2011 samples collected outside the CPS (river and groundwater from IGS33 piezometer) by a dark cross and a full triangle, respectively. Sampleswith [90Sr] below the detection limit are represented on the left outline of the diagram: this representation shows [36Cl] in the sample even if the [90Sr] is below the detection limit.May 2011 vegetation samples are not shown because their [Cl�] and [90Sr] are not known. Potential mixing processes are shown by the dark lines (A) and the three dotted lines (Band C).
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e32 29
between shallow groundwater, contaminated because of migra-tions from the trench (4-02-1 piezometer is located just down-gradient of the trench), with modern rainwater.
The last represented mixing process is the mixing of shallowgroundwater located upgradient of the trench (1-06-1 piezometer)with the rainwater end-member showing the highest [Cl�] (line Din Fig. 5). This mixing line fits quitewell with the trend described bysamples collected in the deeper alluvial layer. Here, the rainwaterend-member could also represent uncontaminated deep ground-water, resulting from uncontaminated rainwater infiltration priorto the Chernobyl accident in 1986, because of the 36Cl contamina-tion decreases with depth and as groundwater is mainly rechargedby rainwater infiltration and because of the quite high [Cl�].However, in the shallow alluvial layer, 36Cl groundwater contami-nation is too constant and in the deepest part, a number of samplesare too low to make this mixing process a strong hypothesis.
To summarize, CPS groundwater contamination might be theresult of the mixing of a more-or-less contaminated shallowgroundwater end-member with an uncontaminated water end-member. These mixing processes should be dispersive mixing ifadvective transport is assumed to be the main transport ofelements.
Samples collected outside the CPS do not fall on these mixingtrends, suggesting other contamination processes can occur at thescale of the area.
5.2.2. Chlorine-36 and Strontium-90 behavior comparison[36Cl] are compared to [90Sr] in Fig. 6 to investigate the differ-
ences in groundwater contamination by these two radionuclides,both assumed to be released from the trench.
The trench soil water sample is the most contaminated sampleboth in 36Cl and in 90Sr. Groundwater samples show an overalldecreasing trend in [36Cl] and [90Sr] from the most contaminated
groundwater (4-02-1 piezometer) to concentrations of1�10�11 mmol L�1 in [36Cl] and below the detection limit for [90Sr].Two different trends can be distinguished in the two differentlayers. In shallow groundwater (aeolian layer and aeolian/alluvialinterface), [36Cl] and [90Sr] fall in between the trench soil water andshallow alluvial samples concentrations and are relatively scat-tered. In the alluvial layer, only [90Sr] decreases while [36Cl] re-mains almost constant, at least in the shallowest samples (thedeepest samples are less contaminated in 36Cl).
The previously considered mixing processes are represented inorder to investigate the observed [36Cl] and [90Sr] trends (c.f. Sec-tion 5.2.1). The following end-members are considered:
- Uncontaminated water, which may correspond to rainwater: anaverage [36Cl] is considered, calculated from the theoretical 36Cl/Cl deposit (c.f. Section 5.1) and the average [Cl�] measured inrainwater (Bugai et al., 2012a). It could also correspond todeeper uncontaminated groundwater, recharged by rainwaterinfiltration prior to groundwater contamination. 90Sr is assumednot to exist in natural environment but as the diagram is pre-sented in log scale, [90Sr] is set at 1 at L�1, corresponding to1.7 � 10�21 mmol L�1. This end-member is not shown in Fig. 6because it is out of the chosen axis ranges.
- The trench soil water sample.- The most contaminated shallow groundwater, represented bythe groundwater sampled in the 4-02-1 piezometer.
To draw the mixing line, [90Sr] are calculated following (Equa-tion (7)) and the three potential mixing processes are reported inFig. 6.
The first observation is that the most contaminated ground-water, shallow groundwater located downgradient of trench T22(4-02-1 and 12-02-1 piezometers, Fig. 2), shows [90Sr] close to
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e3230
[90Sr] in the trench soil water whereas their [36Cl] are almost twoorders of magnitude lower: the high concentrations in these pie-zometers do not seem to be the result of direct dilution of thetrench soil water sample by uncontaminated rainwater (line A inFig. 6). This tends to suggest that the release of 36Cl and 90Sr fromthe source term could be different: 36Cl could be retarded in trenchpreferentially rather than 90Sr, which is almost entirely diluted ininfiltrated rainwater. Publications about 36Cl behavior in soilssupport the hypothesis of 36Cl retention in soils (Cornett et al., 1997;Milton et al., 2003, …). First, 36Cl thermonuclear pulse was shownto be retarded in soils compared to 3H thermonuclear pulse(Cornett et al., 1997; Milton et al., 2003). Then, the affinity of 36Clwith organic matter and vegetation can affect the migration of 36Cl(see Section 5.1.2).
To take into account that difference in contamination, anothercontaminated end-member is considered: groundwater locatedjust downgradient of the trench and assumed to be impacted bysuch processes. A mixing line between this most contaminatedgroundwater end-member and the uncontaminated end-memberis considered (line B; Fig. 6). Actually, most of the shallowgroundwater (in the aeolian and in the shallow alluvial layers),Borschi River water and groundwater sampled in IGS33 piezometerfall between this mixing line and the mixing line between trenchsoil water and uncontaminated water end-members. This obser-vation suggests a variation in 36Cl and/or 90Sr contaminations ingroundwater. Several authors showed that in soil, 36Cl release de-pends on the [Cl�] in infiltrated rainwater and the residence time ofwater in soil (Kashparov et al., 2005; Bastviken et al., 2006).
In contrast, the groundwater samples from the alluvial layer donot fall on these mixing lines: they show an excess in 36Cl or adeficit in 90Sr. Two processes can explain the low [90Sr] whereas[36Cl] remains almost constant in groundwater of the alluvial layer:
- mixing of highly contaminated end-member with water onlycontaminated in 36Cl (3-02-2 piezometer), shown in Fig. 6 by theline C. Upgradient contaminationwouldn't be the direct result ofthis downgradient process but would result from a similarupgradient process.
- loss of 90Sr, by retention process and/or decay.
Retention processes of Sr were already assessed to explain thedelay of 90Sr relatively to the groundwater flow (Dewiere et al.,2004). Moreover, cation exchanges were shown in the aeolianlayer (Szenknect, 2003), the cation exchange capacity increases inthe alluvial layer, and as 90Sr is radioactive and decays (after 25years, nearly half of the initial 90Sr has decayed, the 90Sr half-lifebeing 28.8 y), the second option is the most likely.
Regarding the deepest alluvial piezometers (1-01,1-98-1,1-98-2and 1-98-3 piezometers), [90Sr] is below the detection limit formost samples, consequently processes are not observable. Onlygroundwater sampled in 1-98-1 piezometer, located on the ABprofile, shows [90Sr] above the detection limit, with a high [36Cl],close to 10�10 mmol L�1 and is not on the considered mixing pro-cesses. This particular case could be explained by the trench het-erogeneity, as it was reported for instance for 137Cs repartitionwithin the trench body (Bugai et al., 2005) and could result inhigher 36Cl contamination.
In summary, the study of [36Cl] and [90Sr] comforts the hy-pothesis of a groundwater contamination resulting from thedilution of a highly contaminated end-member with uncon-taminated rainwater, particularly in shallow groundwater. How-ever, 36Cl could be retained in the trench soil and it could resultin a lower contamination in 36Cl of the groundwater of theaeolian layer compared to 90Sr contamination. Next, in ground-water sampled in the alluvial layer, 90Sr is most likely lost by
retention processes, as an increase of CEC from 1 meq/100 g inthe aeolian layer to 5e10 meq/100 g in the alluvial layer(Matoshko et al., 2004), and radioactive decay relative to 36Cl. As[36Cl] remains nearly constant in the shallow alluvial layer while[90Sr] decrease, 36Cl conservative properties can still be assumedin groundwater.
6. Summary and conclusion
The content in 36Cl is investigated in the Chernobyl Pilot Site(CPS) groundwater. This investigation aimed at determining if 36Clmight be a suitable tracer of the contamination from the trench inconditions less reactive as possible.
36Cl/Cl ratios were measured in groundwater samples, a trenchsoil water sample, leaf leachate solutions and the Borschi river,collected in October 2008, October 2009 and May 2011.
In groundwater, the obtained 36Cl/Cl ratios are 1e4 orders ofmagnitude higher than the natural theoretical ratio. Such ratiosshow an anthropogenic origin of the 36Cl in groundwater. Ac-cording to groundwater apparent ages (Le Gal La Salle et al.,2012), this high contamination of the groundwater is prevalentin groundwater recharged after the Chernobyl explosion. TrenchT22 acts as an obvious current point source of 36Cl: soil watersampled in the trench body shows 36Cl/Cl ratio five orders ofmagnitude higher than the natural theoretical ratio and oneorder of magnitude higher than the 36Cl/Cl ratio in the mostcontaminated groundwater samples, located downgradient ofthe trench. Additional sources of 36Cl are supposed to contributeto the total 36Cl contamination of groundwater, such as upgra-dient contamination from upgradient trenches and/or of residual36Cl by biogeochemical processes in soils, or diffuse contami-nation over the whole area, either due to the Chernobyl explo-sion, or due to the nuclear activity at the Power Plant before theaccident, or even to the thermonuclear test. The 36Cl modernrelease has to be investigated, most likely linked to Cl cyclingwith biosphere though a study extended to 36Cl/Cl ratio mea-surements in biosphere, soils and rainwater. The source ofupgradient contamination still needs to be studied as well as itsinfluence on 36Cl contamination in CPS groundwater. Thecontribution of other 36Cl sources, particularly diffuse contami-nation at the time of the reactor explosion, has to be investi-gated too.
The study of the mixing process influence in groundwatershowed that the 36Cl contamination in CPS groundwater can beexplained by the dilution of trench soil water or contaminatedgroundwater located downgradient of the trench, by uncontami-nated water (such as modern rainwater). In the shallow alluviallayer, groundwater 36Cl contamination remains almost constant incomparison with 90Sr content, which decreases because of reten-tion and/or decay processes.
In conclusion, 36Cl could be a good tracer of the contaminationfrom the trench in groundwater in conditions less reactive aspossible because of the persistence of the contamination, mostlylinked to migration from the trench and a more extended plume ingroundwater than 90Sr. However, 36Cl migration from the trenchinto groundwater is not as simple as it seems and to investigate themaximal extent of the 36Cl plume and as the 36Cl contaminantplume clearly exists out of the considered profile, additional pie-zometers would help to define the contaminant plume from trenchT22.
Acknowledgments
The authors would like to thank French Institute for RadiationProtection and Nuclear Safety (IRSN) and the French National
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e32 31
Center for Scientific Research (CNRS) for funding this project.Chloride analyses were performed at the Analysis and experimentalfacilities Laboratory (IRSN) and the laboratory of hydrogeology atAvignon University. 36Cl measurements were performed at theDepartment of Nuclear Physics (Australian National University e
Australia) and the ASTER AMS national facility (CEREGE, Aix enProvence), which is supported by the INSU/CNRS, the FrenchMinistry of Research and Higher Education, IRD, and CEA. Wewould like to thank all these teams for these analyses. Finally, theauthors would also like to thank the GNR-TRASSE project, IGS andUIAR Ukrainian institutes and Nîmes University for their activecollaborations. A special thanks to Yuri Kubko for the map of thearea.
Appendix 1
Direct 36Cl/Cl ratios measurements and details aboutcalculation.
To obtain 36Cl/Cl in the initial sample (Rsample), the followingequation is applied:
Solution Rhal ±Uncer
Mother solution 2.4 � 10�15 ±3 � 10Daughter solution e
Sample Dilution A e prepared with daughter solution
With Rmix, 36Cl/Cl in the sample diluted with the halite solution(at at�1); Clmix, the total amount of chlorine in the samplediluted with the halite solution (mol); Clech, the part of chlorinefor the sample solution (mol); Rhalite, 36Cl/Cl in the halite solu-tion (at at�1); Clhalite, the part of chlorine for the halite solution(mol).
Measurement of halite solutions
Two halite solutions were made to optimize volumes of dilutedsample solution and the used volumes of samples. The moreconcentrated is called “mother solution” and the second, madewith the first, is called “daughter solution”.
These solutions were analyzed several times with the analyzesof samples.
C. Roux et al. / Journal of Environmental Radioactivity 138 (2014) 19e3232
References
Antropov, V.M., Bugai, D.A., Dutton, M.C., Gerchikov, M.Y., Kennett, E.J., Ledenev, A.I.,Novikov, A.A., Rudko, V., Ziegenhagen, Y., 2001. Review and analysis of solidlong-lived and high level radioactive waste arising at the Chernobyl nuclearpower plant and the restricted zone. European Commission, 294p.
Ashworth, D.J., Shaw, G., 2006. A comparison of the soil migration and plant uptakeof radioactive chlorine and iodine from contaminated groundwater. J. Environ.Radioact. 89, 61e80. http://dx.doi.org/10.1016/j.jenvrad.2006.03.006.
Bastviken, D., Sanden, P., Svensson, T., Stahlberg, C., Magounakis, M., Oberg, G.,2006. Chloride retention and release in a boreal forest soil: effects of soil waterresidence time and nitrogen and chloride loads. Environ. Sci. Technol. 40,2977e2982.
Bastviken, D., Thomsen, F., Svensson, T., Karlsson, S., Sanden, P., Shaw, G.,Matucha, M., Oberg, G., 2007. Chloride retention in forest soil by microbialuptake and by natural chlorination of organic matter. Geochem. Cosmochim.Acta 71, 3182e3192. http://dx.doi.org/10.1016/j.gca.2007.04.028.
Bentley, H.W., Phillips, F.M., Davis, S.N., 1986. Chlorine-36 in the terrestrial envi-ronment. In: Fritz, P., Fontes, J.C. (Eds.), Handbook of Environmental IsotopeGeochemistry. Elsevier, pp. 427e480.
Bessho, K., Matsumura, H., Miura, T., Wang, Q., Masumoto, K., Hagura, H.,Nagashima, N., Seki, R., Takahashi, T., Sasa, K., Sueki, K., Matsuhiro, T., Tosaki, Y.,2007. Estimation of thermal neutron fluences in the concrete of proton accel-erator facilities from 36Cl production. Nucl. Instrum. Methods Phys. Res. B 259,702e707.
Bondar'kov, M.D., Bondar'kov, D.M., Maksimenko, A.M., Zheltonozhskii, V.A.,Zheltonozhskaya, M.V., Petrov, V.V., Savin, A.I., 2009. Activity study of graphitefrom the Chernobyl NPP reactor. Bull. Russ. Acad. Sci. Phys. 73, 261e265. http://dx.doi.org/10.3103/S1062873809020300.
Bugai, D., Kashparov, V., Dewi�ere, L., Khomutinin, Y., Levchuk, S., Yoschenko, V.,2005. Characterization of subsurface geometry and radioactivity distribution inthe trench containing Chernobyl clean-up wastes. Environ. Geol. 47, 869e881.
Bugai, D., Tkachenko, E., Van Meir, N., Simonucci, C., Martin-Garin, A., Roux, C., LeGal La Salle, C., Kubko, Y., 2012a. Geochemical influences of the waste trenchno.22T at CPS at the aquifer: long-term trends, governing process, and impli-cations for radionuclide migration. Appl. Geochem. 27, 1320e1338. http://dx.doi.org/10.1016/j.apgeochem.2011.09.028.
Bugai, D., Skalskyy, A., Dzhepo, S., Kubko, Y., Kashparov, V., Van Meir, N.,Stammose, D., Simonucci, C., Martin-Garin, A., 2012b. Radionuclide migra-tion at experimental polygon at red Forest waste site in Chernobyl zone.Part 2: hydrogeological characterization and groundwater transportmodeling. Appl. Geochem. 27, 1359e1374. http://dx.doi.org/10.1016/j.apgeochem.2011.09.028.
Chant, L.A., Andrews, H.R., Cornett, R.J., Koslowky, V., Milton, J.C.D., Van derberg, J.G., Verburg, T.G., Wolterbeek, H.T., 1996. 129I and 36Cl concentrations inlichens collected in 1990 from three regions around Chernobyl. Appl. Radiat.Isot. 47, 933e937.
Conard, N.J., Elmore, D., Kubik, P.W., Gove, H.E., Tubbs, L.E., Chrunyk, B.A.,Wahlen, M., 1986. The chemical preparation of AgCl for measuring 36Cl in polarice with accelerator mass spectrometry. Radiocarbon 28, 556e560.
Cornett, R.J., Andrews, H.R., Chant, L.A., Davies, W.G., Greiner, B.F., Imahori, Y.,Koslowsky, V.T., Kotzer, T., Milton, J.C.D., Milton, G.M., 1997. Is 36Cl fromWeapons' test fallout still cycling in the atmosphere? Nucl. Instrum. MethodsPhys. Res. B 123, 378e381.
Cornett, R.J., Cramer, J., Andrews, H.R., Chant, L.A., Davies, W., Greiner, B.F.,Imahori, Y., Koslowski, V., McKay, J., Milton, G.M., Milton, J.C.D., 1996. In situproduction of 36Cl in uranium ore: a hydrogeological assessment tool. WaterResour. Res. 32, 1511e1518, 0043-1397/96/95WR-03821$09.00.
Dewiere, L., Bugai, D., Grenier, C., Kashparov, V., Ahamdach, N., 2004. 90Sr migrationto the geo-sphere from a waste burial in the Chernobyl exclusion zone.J. Environ. Radioact. 74, 139e150. http://dx.doi.org/10.1016/j.jenvrad.2004.01.019.
Dzhepo, S., Skal'skii, A., 2002. Radioactive contamination of groundwater within theChernobyl exclusion zone. In: Shestopalov, V. (Ed.), Chernobyl e Disaster andGroundwater, pp. 25e70.
Elmore, D., Tubbs, L.E., Newmann, D., Ma, X.Z., Finkel, R., Nishiizumi, K., Beer, J.,Oeschger, H., Andree, M., 1982. 36Cl bomb pulse measured in shallow Ice corefrom Dye 3. Greenl. Nat. 300, 735e737.
Endt, P.M., Van der Leun, C., 1973. Energy levels of a ¼ 21e44 nuclei (V). Nucl. Phys.A214, 1e625.
Finkel, R.C., Nishiizumi, K., Elmore, D., Ferraro, R.D., Gove, H.E., 1980. 36Cl in polarice, rainwater and seawater. Geophys. Res. Lett. 7, 983e986.
Fifield, L.K., Tims, S.G., Stone, J.O., Argento, D.C., De Cesare, M., 2013. Ultra-sensitivemeasurements of 36Cl and 236U at the Australian National University. Nucl.Instrum. Methods Phys. Res. Sect. B: Beam Interact. Mater. Atoms 294, 126e131.http://dx.doi.org/10.1016/j.nimb.2012.04.028.
Freed, R., Smith, L., Bugai, D., 2004. The effective source area of 90Sr for a streamnear Chernobyl, Ukraine. J. Contamin. Hydrol. 71, 1e26. http://dx.doi.org/10.1016/j.jconhyd.2003.07.002.
Hou, X., Ostergaard, L. Frosig, Nielsen, S.P., 2007. Determination of 36Cl in nuclearwaste from reactor decommissioning. Anal. Chem. 79, 3126e3134.
Kashparov, V., Colle, C., Levchuk, S., Yoschenko, V., Zvarich, S., 2007. Radiochlorineconcentration ratios for agricultural plants in various soil conditions. J. Environ.Radioact. 95, 10e22. http://dx.doi.org/10.1016/j.jenvrad.2007.01.008.
Kashparov, V., Colle, C., Zvarich, S., Yoschenko, V., Levchuk, S., Lundin, S., 2005. Soil-to-plant halogens transfer studies 2. Root uptake of radiochlorine by plants.J. Environ. Radioact. 79, 233e253. http://dx.doi.org/10.1016/j.jenrad.2004.07.001.
Kashparov, V., Ahamdach, N., Levchuk, S., Yoschenko, V., Fesenko, S., Maloshtan, I.,2009. Dissolution of particles of irradiated nuclear fuel in the temporary stor-ages of radioactive waste in Chernobyl zone: sources of radionuclides migra-tion. In: Oughton, Deborah H., Kashparov, Valery (Eds.), Radioactive Particles inthe Environment, pp. 139e156.
Kashparov, V., Yoschenko, V., Levchuk, S., Bugai, D., Van Meir, N., Simonucci, C.,Martin-Garin, A., 2012. Radionuclide migration in the experimental polygon ofthe red forest waste site in the Chernobyl zone. Part 1: characterization of wastetrench, fuel particle transformation processes in soils, biogenic fluxes and ef-fects on biota. Appl. Geochem. 27, 1348e1358. http://dx.doi.org/10.1016/j.apgeochem.2011.11.004.
Le Gal La Salle, C., Aquilina, L., Fourre, E., Jean-Baptiste, P., Michelot, J.L., Roux, C.,Bugai, D., Labasque, T., Simonucci, C., Van Meir, N., Noret, A., Bassot, S.,Dapoigny, A., Baumier, D., Verdoux, P., Stammose, D., Lancelot, J., 2012.Groundwater residence time down gradient of the T22 trench at the CPS.Constraints on groundwater models. Appl. Geochem. 27, 1304e1319. http://dx.doi.org/10.1016/j.apgeochem.2011.12.006.
Lee, R.T., Shaw, G., Wadey, P., Wang, X., 2001. Specific association of 36Cl with lowmolecular weight humic substances in soils. Chemosphere 43, 1063e1070.
Levchuk, S., Kashparov, V., Maloshtan, I., Yoschenko, V., Van Meir, N., 2012. Migra-tion of transuranic elements in groundwater from the near-surface radioactivewaste site. Appl. Geochem. 27, 1339e1347.
Matoshko, A., Bugai, D., Dewiere, L., Skalskyy, A., 2004. Sedimentological study ofthe Chernobyl NPP site to schematise radionuclide migration conditions. En-viron. Geol. 46, 820e830. http://dx.doi.org/10.1007/s00254-004-1067-3.
Milton, G.M., Andrews, H.R., Causey, S.E., Chant, L.A., Cornett, R.J., Davies, W.G.,Greiner, B.F., Koslowsky, S.J., Imahori, Y., Kramer, S.J., McKay, J.W., Milton, J.C.D.,1994. Chlorine-36 dispersion in the chalk river area. Nucl. Instrum. MethodsPhys. Res. B 92, 376e379.
Milton, G.M., Milton, J.C.D., Sciff, S., Cook, P., Kotzer, T.G., Cecil, L.D., 2003. Evidencefor chlorine recycling e hydrosphere, biosphere, atmosphere e in a forestedwet zone on the Canadian shield. Appl. Geochem. 18, 1027e1042. http://dx.doi.org/10.1016/S0883-2927(02)00240-8.
Phillips, F.M., 2000. Chlorine-36. In: Cook, P., Herczeg, A.L. (Eds.), EnvironmentalTracers in Subsurface Hydrology, pp. 299e348.
Rosman, K.J.R., Taylor, P.D.P., 1998. Isotopic compositions of the elements 1997. PureAppl. Chem. 70, 217e235.
Seki, R., Matsuhiro, T., Nagashima, Y., Takahashi, T., Sasa, K., Sueki, K., Tosaki, Y.,Bessho, K., Matsumura, H., Miura, T., 2007. Isotopic ratios of 36Cl/Cl in Japanesesurface soil. Nucl. Instrum. Methods Phys. Res. B 259, 486e490. http://dx.doi.org/10.1016/j.nimb.2007.01.227.
Shestopalov, V., 2002. Chernobyl Disaster and Groundwater. Taylor & Francis,p. 289.
Songsheng, J., Shan, J., Hong, G., Shubin, D., 1994. Determination of 36Cl in thegroundwaters and ores around a uranium deposit. Nucl. Instrum. Methods Phys.Res. B 92, 385e388.
Szenknect, S., 2003. Transferts de Radio�el�ements en Zone Non Satur�ee. EtudeExp�erimentale et Mod�elisation Appliqu�ees Au Site Pilote de Tchernobyl. Univ.Joseph Fournier e Grenoble 1 (in French).
Van Meir, N., Bugaï, D., Kashparov, V., 2009. The experimental platform in Cher-nobyl: an international research polygon in the exclusion zone for soil andgroundwater contamination. In: Radioactive Particles Released from DifferentNuclear Sources. Nato Science for Peace and Security Series: C. Springer Verlag,pp. 197e208.
![[CL-0232] - Icsid](https://static.documents.pub/doc/80x56/63259306cedd78c2b50ca837/cl-0232-icsid.jpg)
