Chemosphere xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Carbon and hydrogen isotope fractionation of benzene and toluene during hydrophobic sorption in multistep batch experiments G. Imfeld a,⇑, F.-D. Kopinke b, A. Fischer c,d, H.-H. Richnow c a

Laboratory of Hydrology and Geochemistry of Strasbourg (LHyGeS), University of Strasbourg/ENGEES, UMR 7517 CNRS, France Department of Environmental Engineering, Helmholtz Centre for Environmental Research – UFZ, Leipzig D-04318, Germany c Department of Isotope Biogeochemistry, Helmholtz Centre for Environmental Research – UFZ, Leipzig D-04318, Germany d Isodetect – Company for Isotope Monitoring, Leipzig D-04103, Germany b

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 Carbon and hydrogen isotope

fractionation during multistep sorption was measured for benzene and toluene.  Successive hydrophobic partitioning steps result in low isotope fractionation.  We observed carbon and hydrogen isotope fractionation for the benzene–octanol pair.  Functional groups of SOM may specifically interact with BTEX compounds.

a r t i c l e

i n f o

Article history: Received 10 July 2013 Received in revised form 24 January 2014 Accepted 28 January 2014 Available online xxxx Handling Editor: Klaus Kümmerer Keywords: Sorption Natural attenuation Aquifer Degradation Isotopic fractionation

a b s t r a c t The application of compound-specific stable isotope analysis (CSIA) for evaluating degradation of organic pollutants in the field implies that other processes affecting pollutant concentration are minor with respect to isotope fractionation. Sorption is associated with minor isotope fractionation and pollutants may undergo successive sorption-desorption steps during their migration in aquifers. However, little is known about isotope fractionation of BTEX compounds after consecutive sorption steps. Here, we show that partitioning of benzene and toluene between water and organic sorbents (i.e. 1-octanol, dichloromethane, cyclohexane, hexanoic acid and Amberlite XAD-2) generally exhibits very small carbon and hydrogen isotope effects in multistep batch experiments. However, carbon and hydrogen isotope fractionation was observed for the benzene–octanol pair after several sorption steps (Dd13C = 1.6 ± 0.3‰ and Dd2H = 88 ± 3‰), yielding isotope fractionation factors of aC = 1.0030 ± 0.0005 and aH = 1.195 ± ! 0.026. Our results indicate that the cumulative effect of successive hydrophobic partitioning steps in an aquifer generally results in insignificant isotope fractionation for benzene and toluene. However, significant carbon and hydrogen isotope fractionation cannot be excluded for specific sorbate–sorbent pairs, such as sorbates with p-electrons and sorbents with OH-groups. Consequently, functional groups of sedimentary organic matter (SOM) may specifically interact with BTEX compounds migrating in an aquifer, thereby resulting in potentially relevant isotope fractionation. Ó 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Address: Laboratory of Hydrology and Geochemistry of Strasbourg (LHyGeS), University of Strasbourg/ENGEES, CNRS, 1, rue Blessig, 67 084 Strasbourg Cedex, France. Tel.: +33 3 6885 0407; fax: +33 3 8824 8284. E-mail address: [email protected] (G. Imfeld). http://dx.doi.org/10.1016/j.chemosphere.2014.01.063 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Imfeld, G., et al. Carbon and hydrogen isotope fractionation of benzene and toluene during hydrophobic sorption in multistep batch experiments. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.01.063

2

G. Imfeld et al. / Chemosphere xxx (2014) xxx–xxx

1. Introduction Monitored natural attenuation (MNA) is an approach gaining increased attention to manage aquifers contaminated with organic pollutants. Successful implementation of MNA requires a monitoring strategy relying on several lines of evidence to assess pollutant depletion (Bombach et al., 2010). Compoundspecific stable isotope analysis (CSIA) is proposed to be a useful tool to assess pollutant biodegradation in aquifers (Elsner, 2010; Thullner et al., 2012). CSIA relies on the change in the stable isotope composition of a pollutant between its source and the plume as an indication of chemical or biochemical conversion. Since degradation can lead to stable isotope fractionation, an enrichment of heavier isotopes in the residual fraction of the degraded pollutant can be observed. Hence, stable isotope fractionation for carbon, hydrogen or other elements indicates in situ pollutant degradation, relying on the presumption that only degradation can significantly alter the isotope composition of a pollutant in aquifers. Physical attenuation processes, such as dilution, dispersion, volatilization and sorption, are generally considered not to result in significant isotope fractionation for groundwater pollutants (Dempster et al., 1997; Harrington et al., 1999; Schüth et al., 2003; Wang and Huang, 2003). However, little is known about isotope fractionation during sorption of hydrophobic pollutants in aquifers. Partitioning of hydrophobic pollutants between groundwater and sedimentary organic matter (SOM) in aquifers can control the pollutant distribution and may lead to isotope fractionation. Hydrophobic sorption does not significantly alter the pollutant isotope composition in single step batch experiments, which suggests that sorption-based isotope effects are negligible or small in aquifers (Poulson et al., 1997; Harrington et al., 1999; Slater et al., 2000; Schüth et al., 2003; Höhener and Yu, 2012). However, hydrophobic pollutants undergo successive sorption-desorption steps between the mobile water phase and the stationary SOM phase, which may result in small but significant isotope fractionation (Kopinke et al., 2005; Thullner et al., 2012). Predictive calculations assumed that under certain conditions isotope fractionation is similar to that observed during biodegradation at an expanding front of a plume or in the case of a single contamination event (Kopinke et al., 2005; Van Breukelen and Prommer, 2008; Höhener and Atteia, 2010). Clearly, the cumulative effect of a large number of successive partitioning steps may lead to measurable isotope effects, as previously observed in chromatography-based studies (see for a review Filer, 1999). Reversed-phase liquid chromatography (RP-LC) exhibits some analogies to the migration of a pollutant in an aquifer. Significant carbon and hydrogen isotope fractionation has been observed in RP-LC experiments (Caimi and Brenna, 1993, 1997; Poulson et al., 1997; Turowski et al., 2003; Kopinke et al., 2005; Valleix et al., 2006). It has been suggested that sorption can also cause significant shifts in pollutant isotope ratios in aquifers (Kopinke et al., 2005). However, the ‘chromatographic efficiency’ of an aquifer appeared to be smaller than that of RP-LC columns, thus leading to small or negligible isotope fractionation in aquifers (Schüth et al., 2003). Besides RP-LC-based studies, carbon isotope fractionation was also observed for benzene and toluene in multistep batch sorption experiments with dissolved humic acids (Kopinke et al., 2005). Although, the use of humic substances allows studying the environmental fate of organic compounds, variability in their structure and functional groups may limit the interpretation of the partitioning phenomenon resulting in isotope effects. Therefore, structurally well-defined sorbents or partitioning media, such as 1-octanol, can be more useful for studying sorption of organic pollutants with respect to isotope fractionation.

In the present study, carbon and hydrogen isotope effects during partitioning of benzene and toluene between water and sorbents with different chemical properties were evaluated in multistep batch experiments. Benzene and toluene were selected as model sorbate compounds due to their environmental significance and simple chemical structure. The sorbents were chosen to represent typical structural components of natural organic matter (NOM), such as aliphatic and aromatic moieties, carboxylic groups, hydroxyl groups, and various types of interaction between sorbents and sorbates including (i) non-specific van der Waals interactions (i.e. cyclohexane, amberlite XAD2) and (ii) specific hydrogen-bridging interactions (i.e. octanol, hexanoic acid). Dichloromethane was added as representative of the important class of dipolar aprotic solvents, which are able to polar interactions except hydrogen bonding. Multistep batch experiments were performed in order to generate a maximum isotope shift (Dd13C or Dd2H) upon consecutive partitioning steps, while depletion of benzene and toluene during partitioning still enables measuring the isotope composition in the final solution. 2. Materials and methods 2.1. Chemicals All chemicals were obtained in reagent quality from Merck (Germany). The physico-chemical properties and chemical structures of tested sorbates and sorbents are provided in Table 1. Amberlite XAD-2 is a nonionic styrene-divinylbenzene copolymer in the form of 20 mesh beads. 2.2. GC/FID and GC/C/IRMS measurements For the analysis of concentrations as well as carbon and hydrogen isotope signatures of benzene and toluene, 10 mL aqueous samples were extracted with 500 lL of n-pentane at 80 rpm during 4 h. The concentrations of benzene and toluene were measured on a gas chromatograph equipped with a flame ionization detector (GC-FID) (CP-3800 GC, Varian Inc., USA). Solvent extraction using n-pentane is considered not to affect the isotope signature of BTEX (Dempster et al., 1997). The carbon and hydrogen isotope signatures of analytes were measured using two independent gas chromatography/isotope ratio mass spectrometry (GC/IRMS) systems as described elsewhere (Fischer et al., 2009). The GC/IRMS systems were calibrated using gases (CO2 and H2) with known isotope composition (Coplen et al., 2006; Coplen, 2011). The carbon and hydrogen isotope ratios were expressed in the delta notation (d13C and d2H) in per mil units (‰) relative to the international standards according to the following equation (Coplen, 2011):

d13 C or d2 H ¼

X dissolved 1  X dissolved Coe

ð1Þ

where Rsample and Rstandard are the 13C/12C ratios or 2H/1H ratios of the sample and an international standard, respectively. All samples were measured in at least three replicates. The total analytical uncertainty that incorporates both accuracy and reproducibility for the mean d13C-values was always better than ±0.5‰ and for the d2H-values better than ±10‰. 2.3. Batch experiments Deionized water (1250 mL) was spiked separately with stock solutions of benzene and toluene yielding a final concentration of

Please cite this article in press as: Imfeld, G., et al. Carbon and hydrogen isotope fractionation of benzene and toluene during hydrophobic sorption in multistep batch experiments. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.01.063

3

G. Imfeld et al. / Chemosphere xxx (2014) xxx–xxx Table 1 Physico-chemical properties and chemical structure of sorbates and sorbents. Sorbates

Sorbents

Benzene

Toluene

1-octanol

Dichloromethane

Cyclohexane

Hexanoic acid

Amberlite XAD2

Molecular formula Molecular structure

C6H6

C7H8

C8H18O

CH2Cl2

C6H12

C6H12O2

Nonionic styrene-divinylbenzene copolymer (20 mesh beads)

Molar mass (g mol1) Density (20 °C) (g cm3) Solubility in water (25 °C) (g L1) Log Kow (–)

78.11 0.88 1.79 2.13

92.14 0.87 0.53 2.69

130.23 0.82 0.58 –

84.93 1.33 13.00 1.25

84.15 0.78 Insoluble 3.44

116.16 0.93 10.82 2.05

Not determined 1.08 Not determined Not determined

300 ppm in the aqueous solution. The solutions were agitated at 80 rpm in 1500 mL glass separation funnels (10 h at 20 ± 1 °C). Five parallel experiments, each with a different sorbent, were performed under the same conditions with 1-octanol, dichloromethane, cyclohexane, hexanoic acid and Amberlite XAD-2 (Table 1). The deionized water and the sorbent organic phase (16 g L1) were brought in contact by agitation at 80 rpm for 5 min (20 ± 1 °C), avoiding the formation of emulsions. After separation of the two phases in the separation funnel, 10 mL aliquote of water was collected for concentration and GC/IRMS measurements, and the organic sorbent was completely replaced for a next sorption step. Decrease in the dissolved concentrations of benzene and toluene was attributed to partitioning, and partly to volatilized fractions. The control experiments without sorbent were carried out in parallel using the same multistep batch experiment procedure to evaluate loss of benzene and toluene by volatilization. In parallel to the concentrations, the carbon and hydrogen isotope composition of the analytes was measured after each sorption step by GC/IRMS analysis of the n-pentane extracts until the limit of detection was reached. 2.4. Definitions and calculations Partitioning of analytes between water and dissolved or suspended organic matter is usually described using the sorption coefficient (Koc) according to the following equation:

K oc ¼

Csorbed 1  X dissolved 1 1 ¼  ½L kg  Coc Cdissolved X dissolved

ð2Þ

where Csorbed (mg kg1) is the analyte concentration in organic matter, Cdissolved (mg L1) is the analyte concentration in the water phase, Xdissolved is the dissolved fraction of analyte (Xdissolved + Xsorbed = 1) and Coc is the concentration of organic carbon (OC) in the sample. Koc values can be defined for any distinct species including isotopologues of an organic compound. For simplification, it was assumed that for low-molecular-weight compounds only two isotopically distinct species must be considered: one 12C-isotopologue and one 13C1-isotopologue containing one 13C-atom, regardless of its position in the carbon skeleton. For these isotopologues, Koc val13C ues are given as K 12C oc and K oc , respectively. The isotope fractionation factor asorption is defined according to the following equation:

asorption ¼

K 12C oc K 13C oc

ð3Þ

and can be approximated using Eq. (4), as described elsewhere (Kopinke et al., 2005):

asorption  1 þ Dd13 Cm =½1000 mð1  X dissolved Þ

ð4Þ

where Dd13Cm = d13Cfinal state  d13Cinitial state is the shift in the carbon isotope composition that is generated after m sorption steps during the multistep batch experiment. The bound fraction is repeatedly produced and removed from the system during the multistep batch experiment. m is the number of identical sorption steps with (1  Xdissolved) as the degree of analyte binding in each step. Hydrogen isotope fractionation factors were approximated similarly. The measured decrease of benzene and toluene concentrations (i.e. loss fraction) in the water phase after a partitioning and a phase separation include both analyte depletion by partitioning and by evaporation losses. The evaporation losses of the analytes are assumed to occur mainly during the pouring out of the water phase from the separation funnel, and evaporation losses was assumed to be similar in the control experiments without solvent partitioning. Hence, the loss fraction retrieved from the control experiments was subtracted from the total analyte loss observed in the corresponding sorption experiment, yielding the final Xdissolved value for each experiment. The isotope fractionation factor asorption was calculated while accounting for the sorbent-bound fraction of the analyte, since evaporation loss did not lead to measurable isotope fractionation in the control experiments (see Tables 2 and 3). 3. Results Carbon isotope fractionation factors (aC) (Table 2) and hydrogen isotope fractionation factors (aH) (Table 3) were obtained in the parallel multistep batch sorption experiments during partitioning of benzene and toluene between the sorbents and the water phases. Plots of Dd2H versus Dd13C values measured during the multistep batch experiments show significant carbon and hydrogen isotope fractionations in the case of the benzene–octanol pair (Fig. 1). 3.1. Sorbent-bound fractions The dissolved fractions of the analytes in the water phase varied between Xdissolved = 0.25 and 0.9 (Table 2). Taking into account the evaporation losses in the control experiments, sorbent-bound fractions between Xsorbed = 0.09 and 0.56 were obtained, with the exception of the experiment with dichloromethane. In this experiment, bound fractions were apparently lower due to a different handling of the water phase. Dichloromethane forms a bottom phase, which could be removed during phase separation without flushing out the water phase. This procedure prevents most of the benzene and toluene loss by evaporation. It results a sorbed

Please cite this article in press as: Imfeld, G., et al. Carbon and hydrogen isotope fractionation of benzene and toluene during hydrophobic sorption in multistep batch experiments. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.01.063

4

G. Imfeld et al. / Chemosphere xxx (2014) xxx–xxx

Table 2 Carbon isotope fractionation of benzene and toluene in multistep sorption experiments (isotope shifts (Dd13C) in the aqueous fraction). Benzene Xdissolved Control 1-Octanol Dichloromethane Cyclohexane Hexanoic acid Amberlite XAD-2

Toluene

a

13

Dd C (‰) (after m steps)

0.80 ± 0.10 m = 17 0.71 ± 0.15 m=6 0.89 ± 0.08d m=9 0.67 ± 0.17 m=3 0.32 ± 0.10 m=5 0.24 ± 0.16 m=4

b

0.3 ± 0.9 m = 17 1.6 ± 0.3 m=5 0.2 ± 0.2 m=9 0.3 ± 0.2 m=9 1.4 ± 0.2 m=5 1.1 ± 0.2 m=4

Xdissolveda

Dd13C (‰) (after m steps)b

asorptionc

0.9999 ± 0.0004

0.81 ± 0.07

0.3 ± 0.5

1.0001 ± 0.0001

1.0030 ± 0.0005

0.58 ± 0.05

1.4 ± 0.1

1.0012 ± 0.0001

1.0002 ± 0.0001

0.87 ± 0.07d

0.3 ± 0.2

1.0002 ± 0.0001

1.0009 ± 0.0003

0.54 ± 0.20

1.2 ± 0.3

1.0005 ± 0.0001

1.0006 ± 0.0001

0.26 ± 0.16

1.0 ± 0.6

1.0004 ± 0.0001

1.0005 ± 0.0001

0.39 ± 0.05

1.0 ± 0.4

1.0006 ± 0.0001

asorption

c

a

The error given for Xdissolved values corresponds to one standard deviation of the single values through all the sorption steps. The error given for the Dd13C values was calculated via error propagation based on ±one standard deviation of the mean d13C values from P3 measurements for each sample. c The errors given for the asorption values were calculated via error propagation. The error incorporates errors associated with the Xdissolved and the Dd13C values based on ±one standard deviation of the mean d13C values from P3 measurements for each sample. d Larger Xdissolved values for dichloromethane compared to those retrieved for the control are due to different handling of the water phase and explained in the Section 3.1. b

Table 3 Hydrogen isotope fractionation of benzene and toluene in multistep sorption experiments (isotope shifts (Dd2H) in the aqueous fraction). Benzene

Toluene

Xdissolveda

Dd2H (‰) (after m steps)b

asorptionc

Xdissolveda

Dd2H (‰) (after m steps)b

asorptionc

Control

0.80 ± 0.1

1.010 ± 0.002

0.81 ± 0.07

0.71 ± 0.15

1.195 ± 0.026

0.58 ± 0.05

6±3 m=7 2±4 m=2 6±4 m=3 11 ± 5 m=4 10 ± 8 m=2 12 ± 7 m=4

1.004 ± 0.001

1-octanol

14 ± 4 m=7 88 ± 3 m=5 7±2 m = 10 6±4 m=7 19 ± 5 m=2 20 ± 5 m=4

d

Dichloromethane

0.89 ± 0.08

Cyclohexane

0.67 ± 0.17

Hexanoic acid

0.32 ± 0.10

Amberlite XAD-2

0.24 ± 0.16

d

1.006 ± 0.002

0.87 ± 0.07

1.006 ± 0.002

0.54 ± 0.20

1.019 ± 0.001

0.26 ± 0.16

1.009 ± 0.004

0.39 ± 0.05

1.004 ± 0.009 1.015 ± 0.006 1.011 ± 0.002 1.009 ± 0.003 1.007 ± 0.001

a

The error given for Xdissolved values corresponds to one standard deviation of the single values through all the sorption steps. The error given for the Dd2H values was calculated via error propagation based on ±one standard deviation of the mean d2H values from P3 measurements for each sample. c The errors given for the asorption values were calculated via error propagation. The error incorporates errors associated with the Xdissolved and the Dd2H values based on ±one standard deviation of the mean d2H values from P3 measurements for each sample. d Larger Xdissolved values for dichloromethane compared to those retrieved for the control are due to different handling of the water phase and explained in the Section 3.1. b

fraction Xsorbed,benzene = 1  Xdissolved = 1  0.89 = 0.11, which is in the same order than that obtained for the other sorbents. Therefore, carbon and hydrogen isotope fractionation factors asorption were calculated while accounting for the evaporation losses of the analytes, except in the case of dichloromethane. The data show significantly lower sorbed fractions than those expected from known partitioning coefficients (e.g. Kow). For instance, the Kow value calculated for the experiment with benzene and 1-octanol (Kow = 8) was 17 times lower than the corresponding literature value (Kow = 135) (e.g. Sangster, 1989). This might be due to incomplete partitioning equilibration during the 5 min contact time between aqueous and sorbent phase. However, one can assume that this does not affect the conclusions drawn from the isotope analyses because the lower sorbed fractions are taken into account for the calculation of the isotope fractionation factors according to Eq. (4).

uncertainty of the mean d13C-values) (Table 2). A noticeable exception is the benzene–octanol pair, for which a significant carbon isotope fractionation (Dd13C = 1.6 ± 0.3‰) was observed after 6 sorption steps (Fig. 2a). No significant carbon isotope fractionation was observed for benzene and toluene in the control experiments (Table 2). Beyond the Dd13C values, the carbon isotope fractionation factors (aC) or the corresponding enrichment factors eC = (1  aC)  1000 are more conclusive. We retrieved two enrichment factors eC < 1‰, those for the partitioning of benzene and toluene with n-octanol: eCbenzene,octanol = 3.0 ± 0.5‰ and eCtoluene,octanol = 1.2 ± 0.1‰ respectively. No significant 13C-enrichment in the water fraction could be observed in the other sorption experiments when considering the total analytical uncertainty Dd13C < 1‰. 3.3. Hydrogen isotope fractionation factors

3.2. Carbon isotope fractionation factors The carbon isotope composition of benzene and toluene generally exhibited very small 13C-enrichment in the water fraction (Dd13C < 1‰ after m sorption steps, i.e. 2 times the total analytical

The hydrogen isotope compositions of benzene and toluene generally yielded very small 2H-enrichement in the water fraction (Dd2H < 20‰ after m sorption steps, i.e. 2 times the total analytical uncertainty of the mean d2H-values) (Table 3). Remarkably, a

Please cite this article in press as: Imfeld, G., et al. Carbon and hydrogen isotope fractionation of benzene and toluene during hydrophobic sorption in multistep batch experiments. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.01.063

G. Imfeld et al. / Chemosphere xxx (2014) xxx–xxx

Fig. 1. Plots of Dd2H versus Dd13C for benzene (a) and toluene (b) during their partitioning between the sorbents (i.e., 1-octanol, dichloromethane, cyclohexane, hexanoic acid and Amberlite XAD-2) and the water phase in multistep sorption experiments. The sorption steps m for benzene in the experiment with 1-octanol are indicated in bold close to the symbol. Error bars correspond to the total analytical uncertainty (±0.5‰ for d13C, and ±10‰ for d2H values), which incorporates both accuracy and reproducibility of n P 3 measurements.

5

molecules with heavy isotopes are less strongly sorbing than molecules consisting in only light isotopes. These results are in agreement with previous studies where in single step batch experiments with carbonaceous materials (i.e., graphite, activated carbon, lignite coke and lignite) the measured isotope fractionation were in the range of the accuracy and reproducibility limits of the analytical instruments (0.5‰ for d13C and 10‰ for d2H) (Slater et al., 2000; Schüth et al., 2003). Recently, carbon and hydrogen isotope enrichment factors for equilibrium sorption where estimated using linear free energy relationships (LFERs), and sorption experiments with a soil and activated carbon as sorbents were carried out for perdeuterated cyclohexane and perdeuterated toluene (Höhener and Yu, 2012). This study suggests that equilibrium sorption does create only very small isotope shifts for 13C in groundwater pollutants in aquifers, whereas deuterium shifts are expected to be higher. In another study, significant carbon isotope fractionation was determined for benzene and toluene (aCbenzene = 1.00044 ± 0.0015 and aCtoluene = 1.00060 ± 0.00010) when the cumulative effect of successive partitioning steps with a dissolved humic acid was accounted for (Kopinke et al., 2005). To compare the carbon isotope fractionation factors determined in the present study with those obtained in the latter study, it should be taken into account that an aqueous humic acid is significantly different from octanol as partitioning media. Hydrogen isotope fractionation was not considered in the latter study. The exception of the benzene–octanol and the toluene–octanol pairs with respect to carbon isotope fractionation likely reflects specific interactions, which usually do not occur in hydrophobic sorption. Hydrophobic sorption is driven by a gain in the free enthalpy of the aqueous phase and to a minor extent by non-specific van der Waals interactions in the organic phase (Goss and Schwarzenbach, 2003). In addition, electron donor–acceptor interactions between an aromatic ring (behaving as the electron donor) and an alcohol hydroxyl group may result in predominant electron donor–acceptor interactions (Djordjevic et al., 2009).

significant hydrogen isotope fractionation was observed for the benzene–octanol pair with Dd2H = 88 ± 3‰ after 5 sorption steps. No significant hydrogen isotope fractionation occurred for benzene and toluene in the control experiments (Table 3). This indicates that evaporation of benzene and toluene from the water phase does not result in significant isotope fractionation in multistep batch experiments. The hydrogen isotope fractionation factor (aH) for benzene in octanol calculated from the final Dd2H value using Eq. (4) is aH = 1.195 ± 0.026. Alternatively, one can consider the plot of Dd2H versus m (sorption steps) including all the analyzed benzene fractions. According to Eq. (4) this plot should be linear rather than curved (Fig. 2b). If one disregards the final data point and takes a linear regression line through the remaining four points it results a lower value of aH = 1.100 ± 0.030, which still reflects a significant hydrogen isotope fractionation. No significant 2H-enrichment in the water fraction could be observed in the other sorption experiments when considering the total analytical uncertainty (Dd2H < 20‰). 4. Discussion Our results show that partitioning of benzene and toluene between water and structurally different organic sorbents (i.e., 1-octanol, dichloromethane, cyclohexane, hexanoic acid and Amberlite XAD-2, see Table 1) yields insignificant carbon and hydrogen isotope effects in multistep partitioning experiments for most of the sorbents. However, both carbon and hydrogen isotope fractionation of benzene are significant during partitioning between water and 1-octanol. Thus, this study underlines that

Fig. 2. Plots of Dd13C (a) and Dd2H (b) versus m (sorption steps) for benzene, during its partitioning between 1-octanol and the water phase in multistep sorption experiments. Error bars correspond to the total analytical uncertainty (±0.5‰ for d13C, and ±10‰ for d2H values), which incorporates both accuracy and reproducibility of n P 3 measurements.

Please cite this article in press as: Imfeld, G., et al. Carbon and hydrogen isotope fractionation of benzene and toluene during hydrophobic sorption in multistep batch experiments. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.01.063

6

G. Imfeld et al. / Chemosphere xxx (2014) xxx–xxx

In the case of the benzene–octanol pair, one can assume additional donor-acceptor interactions between the p-electrons of the aromatic ring of benzene and the OH group of the 1-octanol. These specific sorbate–sorbent interactions may yield isotope fractionation. It is reasonable to assume a slightly higher polarizability of the p-electrons in the light isotopologues, leading to a slightly higher electron donor capability (Wade, 1999). This slight difference in the polarizability of the p-electrons among isotopologues seems to be the reason for isotope fractionation observed for the benzene–octanol and the toluene–octanol pairs. Consequently, partitioning of isotopically lighter BTEX molecules to the nonpolar organic phase is expected to be larger compared to that of their heavier isotopologues (Turowski et al., 2003; Valleix et al., 2006). Two more findings are worth mentioning: (i) Toluene (an aromatic compound with an aliphatic group) exhibits a significantly lower isotope fractionation in octanol compared to that observed for benzene (a pure aromatic). The difference is most striking for hydrogen isotope fractionation (aH = 1.004 for toluene vs. aH = 1.195 for benzene). Since the hydrophobic effect of alkylated monoaromatic compounds, such as toluene, is larger than that of benzene, similar or lower carbon and hydrogen isotope fractionations are expected during partitioning to aliphatic or humic sorbents. Apparently, the methyl group and its interactions play a significant role for the net effect. Differences in the governing characteristics for aqueous solubility between benzene and toluene, such as excess molar refraction, polarizability, effective hydrogen-bound acidity and basicity, and the McGowan characteristic volumes, might also influence isotope fractionation (Abraham et al., 1994). (ii) Hexanoic acid is a solvent with electron acceptor (H-donor) properties similar to those of octanol. However, its interactions with benzene revealed a lower isotope fractionation (aC = 1.0006 and aH = 1.0019 vs. aC = 1.0030 and aH = 1.195 for octanol). This suggests that electron donor– acceptor interactions between an aromatic ring (behaving as the electron donor) and an hydroxyl group may not systematically result in predominant electron donor–acceptor interactions leading to significant isotope fractionation. Apparently, the presence of a carboxylate in the case of hexanoic acid and its interactions play a significant role, which may lower the net effect. When comparing the carbon isotope fractionation factors determined in the present study with those from Kopinke et al. (2005) determined for multistep sorption of benzene and toluene with dissolved humic acids (aC = 1.00044 ± 0.00015 for benzene and aC = 1.00060 ± 0.00010 for toluene), one can notice that carbon isotope fractionation is in the same order of magnitude than those with hexanoic acid, but much lower than those with octanol (both from the present study). This underscores the high specificity of the sorbent-sorbate interactions with respect to carbon and hydrogen isotope fractionation. Three main processes should also be considered in the interpretation of the results: (i) diffusion controlled rather than equilibrium conditions, (ii) co-solvent effects, and (iii) concurrent volatilization. These processes are discussed below. Diffusion of heavy isotopologues is expected to be slower than that of lighter isotopologues (LaBolle et al., 2008). This may result in enrichment of the heavier isotopologues in the water phase under diffusion controlled extraction conditions because the overall direction of diffusion processes is from the water into the organic phases. However, the mass transfer processes did not result in significant isotope fractionation for most of the investigated watersolvent pairs. Therefore, one can conclude that mass-exchanges

controlled by diffusion are not responsible for the observed isotope fractionation, assuming that the overall mass transfer is controlled by the diffusion of benzene and toluene through the stagnant boundary layer of the water phase. Diffusion through the stagnant boundary layer of the water phase was likely the prevailing process in our case because water has (i) a much lower solute concentration gradient inside the boundary layer, which is the driving force of diffusion, (ii) a lower diffusivity for organic solutes than the organic solvents, and (iii) a higher viscosity than the organic solvents. A diffusion step can only lead to isotope fractionation if it is involved in the rate-controlling step of the overall mass transfer. Since the water boundary layer is the same for all water-solvent systems, a putative isotope fractionation generated by diffusion through the water boundary layer would have been observed for all water-solvent pairs, and not only for specific pairs. Similarly, admitting diffusion in the stagnant boundary layer on the solvent side of the interface as possibly rate controlling step, one could speculate on specific solvent effects on diffusion coefficients. However, if the diffusion through solvent side of the interface would lead to isotope fractionation, it would be the case for all investigated water-solvent pairs, which does not correspond to our observation. Co-solvent effects (i.e., the effects of water-saturation of the organic phase, and vice-versa) may occur in partition systems with significant miscibility of the phases and cannot be totally excluded in the present study. However, the water solubility of the chosen solvents is relatively low, and its effect on the partition properties of the solvents with respect to benzene and toluene are expected to be low. For instance, a comparison between the results obtained in the experiments with 1-octanol (water solubility in 1-octanol: 4.7 g L1) and dichloromethane (water solubility in dichloromethane: 15 g L1) supports that possible co-solvent effects did not result in significant isotope fractionation in our case. It is known that volatilization of BTEX from water cause only insignificant isotope fractionation (Harrington et al., 1999; Wang and Huang, 2003), which is in agreement with the control experiments without sorbent carried out in our study using the multistep batch experiment procedure. The control experiments clearly show that losses of benzene and toluene by volatilization does not result in significant isotope fractionation. Moreover, both solvent-air and water-air volatilization of benzene and toluene inside the separation funnel are assumed to be negligible in our experiments, since the volume ratio of gas phase to liquid phase times Henry’s law coefficients for benzene and toluene was low. Analytes volatilization occurred mainly during the pouring out step from the aqueous solution. Therefore, the observed isotope fractionation was assigned exclusively to the partitioning step in this study. Overall, our results underscore that specific sorbent–sorbate interactions with respect to carbon and hydrogen isotope fractionation may occur in aquifer during the migration of BTEX compounds in groundwater. During plume migration in an aquifer, benzene, toluene and other organic compounds may undergo several successive partitioning steps between the mobile water phase and the stationary SOM phase, which eventually may lead to measurable isotope fractionation (Kopinke et al., 2005; Van Breukelen and Prommer, 2008; Höhener and Atteia, 2010). Our results indicate that specific interactions between benzene and toluene and some functional groups may lead to unexpectedly relevant isotope effects. The values of Dd2H/Dd13C slopes for benzene during partitioning between 1-octanol and the water phase in multistep sorption experiments (Dd2H/Dd13C slope = 66 ± 9, calculated based on linear regression analysis of Dd2H versus Dd13C values for benzene for the different sorption steps) are higher than those obtained for benzene biodegradation (Dd2H/Dd13C slope ranging from 0 ± 5 to 39 ± 5) (Fischer et al., 2008; Mancini et al., 2008) (Fig. 3) indicating that both processes seem to be distinguishable using carbon and

Please cite this article in press as: Imfeld, G., et al. Carbon and hydrogen isotope fractionation of benzene and toluene during hydrophobic sorption in multistep batch experiments. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.01.063

G. Imfeld et al. / Chemosphere xxx (2014) xxx–xxx

7

MEST-CT-2004-8332) fellowship. We thank M. Gehre, U. Günther, S. Woszidlo and K. Ethner for their technical support in analytical and stable isotope measurements. References

Fig. 3. Plots of Dd2H versus Dd13C values for benzene measured during partitioning between 1-octanol and the water phase in multistep sorption experiments compared to the range of values determined for benzene biodegradation (Mancini et al., 2008 and Fischer et al., 2008). The solid lines represent linear regressions of d2H and d13C values. The Dd2H/Dd13C slope was calculated based on linear regression analysis of Dd2H versus Dd13C values for benzene for the different sorption steps. Prediction intervals (dashed lines) of the Dd2H/Dd13C slope for multistep benzene–octanol sorption are plotted correspondingly.

hydrogen CSIA. However, variation of sorption-induced isotope fractionation may occur depending on the structural properties of the pollutant and the sorbing material and, thus, the differentiation between degradation and sorption may be limited. 5. Summary and conclusions During migration in aquifers, BTEX compounds may undergo successive sorption-desorption steps between water and organic sorbents, potentially leading to measurable isotope fractionation. In this study, multistep batch experiments were conducted to assess carbon and hydrogen isotope effects during partitioning of benzene and toluene between water and structurally different sorbents. Our results showed that hydrophobic sorption of benzene and toluene exhibited minor isotope fractionation within accuracy and reproducibility of GC/IRMS measurements, even after consecutive sorption steps. However, a significant carbon and hydrogen isotope fractionation was observed in the case of the benzene–octanol pair (Dd13C = 1.6 ± 0.3‰ and Dd2H = 88 ± 3‰), yielding isotope fractionation factors of aC = 1.0030 ± 0.0005 and aH = 1.195 ± 0.026. This indicates that sorption-based isotope fractionation of dissolved benzene and toluene is weak if hydrophobic interactions dominate the sorption process, whereas specific interactions may lead to isotope fractionation. Consequently, functional groups of SOM may specifically interact with benzene and toluene migrating in aquifer, thereby resulting in isotope fractionation and confounding the evaluation of benzene and toluene degradation by the mean of CSIA. The heterogeneous nature of SOM generally hinders a detailed interpretation of sorption mechanisms involved in isotope fractionation. Therefore, the use of structurally well-defined sorbents, such as those in the present study, helps understanding isotope fractionation in sorption processes. More generally, considering the properties of SOM with respect to isotope fractionation may enhance the prediction of in situ isotope fractionation. We anticipate our results to be a starting point for considering sorption-based isotope fractionation in the case of specific sorbate–sorbent interactions occurring in aquifers. Acknowledgements G. Imfeld was supported by a European Union Marie Curie Early Stage Training Fellowship (AXIOM, contract No.

Abraham, M.H., Andonianhaftvan, J., Whiting, G.S., 1994. The factors that influences the solubility of gases and vapors in water at 298 K, and a new method for its determination. J. Chem. Soc., Perkin Trans. 2 (8), 1777–1791. Bombach, P., Richnow, H.H., Kästner, M., Fischer, A., 2010. Current approaches for the assessment of in situ biodegradation. Appl. Microbiol. Biot. 86, 839– 852. Caimi, R.J., Brenna, J.T., 1993. High-precision liquid chromatography-combustion isotope ratio mass-spectrometry. Anal. Chem. 65, 3497–3500. Caimi, R.J., Brenna, J.T., 1997. Quantitative evaluation of carbon isotopic fractionation during reversed-phase high-performance liquid chromatography. J. Chromatogr. A 757, 307–310. Coplen, T.B., 2011. Guidelines and recommended terms for expression of stableisotope-ratio and gas-ratio measurement results. Rapid Commun. Mass Spectrom. 25, 2538–2560. Coplen, T.B., Brand, W.A., Gehre, M., Groning, M., Meijer, H.A.J., Toman, B., Verkouteren, R.M., 2006. New guidelines for delta C-13 measurements. Anal. Chem. 78, 2439–2441. Dempster, H.S., Sherwood Lollar, B., Feenstra, S., 1997. Tracing organic contaminants in groundwater: a new methodology using compound-specific isotopic analysis. Environ. Sci. Technol. 31, 3193–3197. Djordjevic´, B.D., Radovicˇ, I.R., Kijevcˇanin, M.L.J., Tasic´, A.Z., Serbanovic´, S.P., 2009. Molecular interaction studies of the volumetric behaviour of binary liquid mixtures containing alcohols. J. Serb. Chem. Soc. 74, 477–491. Elsner, M., 2010. Stable isotope fractionation to investigate natural transformation mechanisms of organic contaminants: principles, prospects and limitations. J. Environ. Monitor. 12, 2005–2031. Filer, C.N., 1999. Isotopic fractionation of organic compounds in chromatography. J. Labelled Compd. Rad. 42, 169–197. Fischer, A., Herklotz, I., Herrmann, S., Thullner, M., Weelink, S.A.B., Stams, A.J.M., Schlömann, M., Richnow, H.H., Vogt, C., 2008. Combined carbon and hydrogen isotope fractionation investigations for elucidating benzene biodegradation pathways. Environ. Sci. Technol. 42, 4356–4363. Fischer, A., Gehre, M., Breitfeld, J., Richnow, H.H., Vogt, C., 2009. Carbon and hydrogen isotope fractionation of benzene during biodegradation under sulfate-reducing conditions: a laboratory to field site approach. Rapid Commun. Mass Spectrom. 23, 2439–2447. Goss, K.U., Schwarzenbach, R.P., 2003. Rules of thumb for assessing equilibrium partitioning of organic compounds: successes and pitfalls. J. Chem. Educ. 80, 450–455. Harrington, R.R., Poulson, S.R., Drever, J.I., Colberg, P.J.S., Kelly, E.F., 1999. Carbon isotope systematics of monoaromatic hydrocarbons: vaporization and adsorption experiments. Org. Geochem. 30, 765–775. Höhener, P., Atteia, O., 2010. Multidimensional analytical models for isotope ratios in groundwater pollutant plumes of organic contaminants undergoing different biodegradation kinetics. Adv. Water Resour. 33, 740–751. Höhener, P., Yu, X., 2012. Stable carbon and hydrogen isotope fractionation of dissolved organic groundwater pollutants by equilibrium sorption. J. Contam. Hydrol. 129–130, 54–61. Kopinke, F.-D., Georgi, A., Voskamp, M., Richnow, H.H., 2005. Carbon isotope fractionation of organic contaminants due to retardation on humic substances: implications for natural attenuation studies in aquifers. Environ. Sci. Technol. 39, 6052–6062. LaBolle, E.M., Fogg, G.E., Eweis, J.B., Gravner, J., Leaist, D.G., 2008. Isotopic fractionation by diffusion in groundwater. Water Resour. Res. 44, W07405. Mancini, S.A., Cheryl, E.D., Elsner, M., Nandi, M.E., Ulrich, A.C., Edwards, E.A., Sherwood Lollar, B., 2008. Isotopic evidence suggests different initial reaction mechanisms for anaerobic benzene biodegradation. Environ. Sci. Technol. 42, 8290–8296. Poulson, S.R., Drever, J.I., Colberg, P.J.S., 1997. Estimation of K-oc values for deuterated benzene, toluene, and ethylbenzene, and application to ground water contamination studies. Chemosphere 35, 2215–2224. Sangster, J., 1989. Octanol–water partition coefficient of simple organic compounds. J. Phys. Chem. Ref. Data 18, 1111. Schüth, C., Taubald, H., Bolano, N., Maciejczyk, K., 2003. Carbon and hydrogen isotope effects during sorption of organic contaminants on carbonaceous materials. J. Contam. Hydrol. 64, 269–281. Slater, G.F., Ahad, J.M.E., Lollar, B.S., Allen-King, R., Sleep, B., 2000. Carbon isotope effects resulting from equilibrium sorption of dissolved VOCs. Anal. Chem. 72, 5669–5672. Thullner, M., Centler, F., Richnow, H.H., Fischer, A., 2012. Quantification of organic pollutant degradation in contaminated aquifers using compound specific stable isotope analysis – review of recent developments. Org. Geochem. 42, 1440– 1460. Turowski, M., Yamakawa, N., Meller, J., Kimata, K., Ikegami, T., Hosoya, K., Tanaka, N., Thornton, E.R., 2003. Deuterium isotope effects on hydrophobic interactions: the importance of dispersion interactions in the hydrophobic phase. J. Am. Chem. Soc. 125, 13836–13849.

Please cite this article in press as: Imfeld, G., et al. Carbon and hydrogen isotope fractionation of benzene and toluene during hydrophobic sorption in multistep batch experiments. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.01.063

8

G. Imfeld et al. / Chemosphere xxx (2014) xxx–xxx

Valleix, A., Carrat, S., Caussignac, C., Leonce, E., Tchapla, A., 2006. Secondary isotope effects in liquid chromatography behaviour of H-2 and H-3 labelled solutes and solvents. J. Chromatogr., A 1116, 109–126. Van Breukelen, B.M., Prommer, H., 2008. Beyond the Rayleigh equation: reactive transport modeling of isotope fractionation effects to improve quantification of biodegradation. Environ. Sci. Technol. 42, 2457–2463.

Wade, D., 1999. Deuterium isotope effects on noncovalent interactions between molecules. Chem-Biol. Interact. 117, 191–217. Wang, Y., Huang, Y.S., 2003. Hydrogen isotopic fractionation of petroleum hydrocarbons during vaporization: implications for assessing artificial and natural remediation of petroleum contamination. Appl. Geochem. 18, 1641– 1651.

Please cite this article in press as: Imfeld, G., et al. Carbon and hydrogen isotope fractionation of benzene and toluene during hydrophobic sorption in multistep batch experiments. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.01.063