Diffusive Fractionation of BTEX and Chlorinated Ethenes in Aqueous

Article pubs.acs.org/est

Diffusive Fractionation of BTEX and Chlorinated Ethenes in Aqueous Solution: Quantification of Spatial Isotope Gradients Biao Jin,† Massimo Rolle,*,†,‡ Ting Li,† and Stefan B. Haderlein† †

Center for Applied Geosciences, University of Tübingen, Hölderlinstrasse 12, D-72074 Tübingen, Germany Department of Civil and Environmental Engineering, Stanford University, 473 Via Ortega, 94305 Stanford, California United States

‡

S Supporting Information *

ABSTRACT: Laboratory experiments were performed to investigate and quantify the extent of diffusive isotope fractionation of organic contaminants in aqueous solution. We selected petroleum hydrocarbons (toluene and ethylbenzene, in 1:2 mixtures of labeled (perdeuterated) and nonlabeled isotopologues) and chlorinated solvents (trichloroethene, TCE, and cis-dichloroethene, cis-DCE, at their natural isotopic abundance) as model compounds. The experimental approach using gel diffusion tubes allowed us to resolve concentration and isotopic gradients induced by isotopologuespecific diffusion and to determine aqueous diffusion coefficients in agreement with the values calculated using published empirical correlations. The experimental results were quantitatively evaluated with numerical simulations to determine the aqueous diffusion coefficients, D, and the exponent of the inverse power-law relation between D and the molecular mass of the isotopologues. The results show remarkable diffusive isotope fractionation for all the investigated organic compounds; however, the extent of fractionation was found to be smaller for the chlorinated ethenes and remarkably deviating from an inverse square root relationship between the isotopologues diffusion coefficients and their molecular mass. The outcomes of this study are relevant for the interpretation of isotopic signatures of organic contaminants in environmental systems and for the quantitative application of compound specific isotope analysis (CSIA) that needs to take into account the fractionation effects of both physical and transformation processes.

■

the experiments of Richter et al.24 for aqueous diffusion of lithium, magnesium, and chloride and of Eggenkamp and Coleman23 for diffusion of chloride and bromide. The outcomes of these studies showed kinetic isotope fractionation during diffusion of the selected cations and anions in aqueous systems. However, the extent of aqueous diffusive isotope fractionation was different for the distinct charged species in aqueous solution and significantly lower than the effects observed in molten silicates liquids and from those predicted assuming a classical inverse square root relationship between the diffusion coefficient and the molecular mass. Both studies explained the experimental findings pointing to the important role of solute−solvent interactions and of the size and structure of the hydration shell surrounding the charged species in aqueous systems. Further insights that helped in interpreting the experimental results were provided by the work of Bourg and Sposito19,21 and Bourg et al.25 By means of molecular dynamics simulations these studies helped to interpret the experimental observations and allowed identifying the residence time of water molecules in the first solvation shell of the ions as key property quantifying the strength of solute−solvent

INTRODUCTION Compound specific isotope analysis (CSIA) is a powerful tool to assess transformations of organic pollutants in environmental systems.1−3 This technique is finding an increasing number of applications as summarized in recent reviews.2,4,5 The basic principle of the CSIA approach for organic contaminants is that light isotopes located at a certain reactive position react faster than the heavy ones. The isotopic composition in the remaining organic compound fraction is monitored by CSIA, which allows obtaining direct information on (bio)degradation and other transformation processes. An increasing number of investigations have demonstrated the important role of physical processes on isotope fractionation of organic contaminants. Fractionating effects have been reported for mass transfer limitations prior to contaminant transformation,6−8 sorption,9,10 volatilization,11,12 aqueous diffusion13 and transverse dispersion.14−16 Focusing on aqueous diffusion, despite the pivotal importance of this process for solute transport in many environmental systems, there is a lack of experimental data on diffusive isotope fractionation of organic pollutants. The research in this direction has been mainly focused on the investigation of diffusion-induced isotope fractionation of dissolved trace gases (e.g., noble gases, methane, and CO2)17−20 and ionic species (e.g., Br−, Cl−, SO42−, Li+, Mg2+, and Ca2+)21−27 in various diffusion-dominated laboratory and field systems. For ionic species important evidence was provided by © 2014 American Chemical Society

Received: Revised: Accepted: Published: 6141

October 21, 2013 April 28, 2014 May 8, 2014 May 8, 2014 dx.doi.org/10.1021/es4046956 | Environ. Sci. Technol. 2014, 48, 6141−6150

Environmental Science & Technology

Article

Table 1. Overview of the Experiments and Best-Fit Parameters for Toluene and Ethylbenzene Diffusion: Computed Average Values and Standard Deviations (±1σ) average compound toluene

ethylbenzene

exp.

time [days]

A B C D E F G H

10.0 10.8 12.5 12.8 10.0 10.5 12.0 12.3

DL [× 10−9 m2 s−1] DH [× 10−9 m2 s−1] 0.859 0.800 0.779 0.780 0.764 0.761 0.733 0.770

± ± ± ± ± ± ± ±

0.061 0.032 0.053 0.048 0.025 0.024 0.027 0.028

0.829 0.769 0.750 0.752 0.734 0.733 0.702 0.738

± ± ± ± ± ± ± ±

0.060 0.031 0.051 0.046 0.025 0.024 0.026 0.027

β [-] 0.430 0.485 0.455 0.450 0.450 0.421 0.483 0.466

± ± ± ± ± ± ± ±

0.074 0.031 0.054 0.027 0.076 0.080 0.088 0.086

DL [× 10−9 m2 s−1] DH [× 10−9 m2 s−1]

β [-]

0.805 ± 0.037

0.775 ± 0.037

0.455 ± 0.023

0.757 ± 0.017

0.727 ± 0.017

0.455 ± 0.027

experimental observations and to extrapolate the findings to a larger scale scenario relevant for practical application of CSIA in subsurface systems.

interactions and determining the distinct extent of diffusive isotope fractionation of monatomic solutes in liquid water. Such advances are yet not well documented for diffusion of uncharged polyatomic species, such as organic contaminants, in aqueous systems although an improved understanding of diffusive isotopic fractionation of such compounds is of critical importance for a correct interpretation of isotopic data in groundwater systems. In fact, the importance of diffusion as a major transport mechanism in saturated porous media is increasingly acknowledged in studies on plume migration in groundwater. For instance, several investigations have pointed out the role of back diffusion from silt and clay aquitards for the persistence of organic contaminant plumes.13,28,29 Also, recent high-resolution studies in contaminated aquifers have shown an effect of diffusion on the observed isotopic signatures at the plume fringes.30,31 Moreover, experimental and modeling investigations of transverse hydrodynamic dispersion in advection-dominated flow-through systems have demonstrated an explicit dependence of mechanical dispersion (i.e., the term of the hydrodynamic dispersion coefficient function of the water flow velocity) on the aqueous diffusion coefficients of the transported species. These findings have significant implications for conservative and mixing-controlled reactive transport in porous media,32,33 for coupled displacement of charged species leading to multicomponent ionic dispersion,34 for transport of different organic contaminant isotopologues, and for the interpretation of isotopic signatures in groundwater.14,16 The latter task requires high-resolution data which are currently not available for most organic chemicals of concern. This hampers a quantitative understanding and interpretation of the effects of aqueous diffusion on isotope fractionation of organic pollutants in aqueous systems. In this work we performed controlled laboratory experiments to quantify the effect of diffusive isotope fractionation on the diffusive displacement of different organic compounds, and we provide first data for contaminants of great environmental concern. We selected, as model compounds, neutral organic contaminants frequently found in groundwater, including petroleum hydrocarbons (toluene; ethylbenzene) and chlorinated solvents (cis-dichloroethene; trichloroethene). The experiments with toluene and ethylbenzene were performed using mixtures of labeled (perdeuterated) and nonlabeled isotopologues, whereas the experiments with cis-DCE and TCE were carried out using these compounds at their natural isotopic abundance and determining the spatial and temporal evolution of the chlorine isotope ratio. The results allowed us to clearly capture the effects of diffusion-induced spatial isotope gradients and their temporal evolution. A one-dimensional modeling approach was used to quantitatively interpret the

■

MATERIALS AND METHODS Diffusion Experiments. The experiments were performed in cylindrical glass tubes filled with agarose gel. We prepared the gel using deionized water and a minimal concentration of phyto agar (1% w/w) to obtain a medium in which the extent of diffusion is very similar to the one that would be observed in purely aqueous systems.35 In such a medium, only a minimal correction taking into account the effective volume fraction of the gel and its coefficient of obstruction (accounting for the detour of the diffusing solute around the particles of the gel substance) need to be applied to compare the observed diffusion coefficients with values obtained for purely aqueous systems (Supporting Information, Section 3). The experimental setup is similar to the one used to investigate fractionation of chlorine and bromine stable isotopes during diffusive transport of chloride and bromide ions.23 The glass tubes (25 cm long; 1.1 cm inner diameter) are glass sealed on one end and can be closed using crimp caps on the open opposite side. We prepared solutions with agar containing the organic contaminants as well as blank solutions (i.e., without contaminant). The glass tubes were partially filled up to 15 cm with a volume of 14.2 mL of solution containing the blank liquid agar representing the pristine fraction of the test tube into which diffusion occurs. After solidification of this blank medium, agar solutions (9.5 mL) containing the individual dissolved organic contaminants were added to fill up the tubes. These solutions contained known concentrations of different isotopologues (1:2 mixture perdeuterated:nondeuterated for toluene and ethylbenzene; natural abundance of isotopologues of cis-DCE and TCE). Once solidified, these solutions acted as contamination sources, progressively releasing the organic contaminant to the adjacent blank medium. During the setup preparation, the hot agar medium was injected using a 25-mL gastight syringe (Hamilton, Bonaduz, Switzerland) with a 20 cm stainless steel needle (UNIMED; Lausanne, Switzerland). After filling the agar medium containing the contaminant, the glass tubes were first covered with aluminum foil and crimp-sealed using caps with Teflon-coated silicone septa. To prevent gas exchange and contaminant mass losses we applied an additional sealing using wax. For a given experiment, several tubes were prepared starting from the same contaminant solution. The tubes were placed horizontally in a thermally insulated box and kept at a constant temperature of 20 °C. As shown in the schematic description of the experimental protocol provided in the Supporting Information, after several days necessary for diffusive isotopic gradients to 6142

dx.doi.org/10.1021/es4046956 | Environ. Sci. Technol. 2014, 48, 6141−6150


Article

Table 2. Overview of the Experiments and Best-Fit Parameters for cis-DCE and TCE: Computed Average Values and Standard Deviations (±1σ) average compound

exp.

time [days]

cis-DCE

I J K L M N

7.0 9.0 12.8 8.0 9.0 10.3

TCE

D [× 10−9 m2 s−1] 1.098 1.105 1.120 0.898 0.899 0.896

± ± ± ± ± ±

0.109 0.128 0.141 0.252 0.523 0.218

β [-]

D [× 10−9 m2 s−1]

β [-]

± ± ± ± ± ±

1.108 ± 0.011

0.088 ± 0.015

0.898 ± 0.002

0.043 ± 0.008

0.104 0.075 0.086 0.049 0.034 0.046

establish, the tubes were sacrificed and sampled. The sampling procedure started with breaking the glass tube and removing the gel that was successively cut into 1 cm slices with a scalpel. Each slice was immediately put in a 10 mL glass vial for subsequent headspace GC-analysis and closed with screw caps with Teflon coated silicone septa. To correct for small errors due to uneven cutting, we weighted each vial on a precision scale and used these data as a gravimetric correction of the concentration measurements. The 25 vials were heated to melt the gel, and measurements were performed using GC-MS at an incubation temperature of 50 °C. An overview of the experiments is provided in Table 1 and Table 2 for the petroleum hydrocarbons and the chlorinated solvents, respectively. Chemicals. The organic compounds (acronym; manufacturer) used in the experiments include the following: ethylbenzene (ETB, ACROS, New Jersey, USA), perdeuterated ethylbenzene (D-ETB, Sigma-Aldrich, Steinheim, Germany), toluene (TOL, Merck, Darmstadt, Germany), perdeuterated toluene (D-TOL, Sigma-Aldrich, Seelze, Germany), cis-1,2dichloroethene (cis-DCE, Sigma-Aldrich, Steinheim, Germany), trichloroethene (TCE, Merck, Darmstadt, Germany), and phyto agar in powder (Duchefa, Haarlem, Netherland). Analytical Methods. Gas chromatography (GC) was used to determine the concentrations of all analytes as well as the chlorine isotope ratios for the chlorinated compounds. Compound specific chlorine isotope analysis using quadrupole MS has recently been enabled by advances in analytical techniques36−40 and evaluation methods41,42 and fostered new experimental and modeling applications.43−46 In this study we used a 7890A gas chromatograph connected to a 5975C quadrupole mass selective detector (MSD) (Agilent, Santa Clara, CA, USA) for both concentration and chlorine isotope analysis. Automated headspace sample injection was performed with a COMBIPAL multipurpose autosampler (Gerstel, Australia). The GC was equipped with a split/splitless injector and a capillary column (60 m × 250 μm, 1.4 μm film thickness; Restek, USA). Helium at 1 mL/min was used as carrier gas. The GC oven program was set as follows: 40 °C (2 min) →110 °C at a rate of 25 °C/min →200 °C at a rate of 15 °C/min (5 min). 250 μL of headspace for each sample was injected for analysis. The quadrupole mass spectrometer were operated in the selected ion mode (SIM) with a dwell time of 50 ms applied to each target ion. The analysis was conducted considering the ions of different analytes: TOL (91.1 m/z) and D-TOL (98.1 m/z) for toluene and the most abundant ions ETB (91.1 m/z) and D-ETB (98.1 m/z) for ethylbenzene. The target ions for the chlorinated ethenes were as follows: TCE (130, 132, 134, 136 m/z) and cis-DCE (96, 98, 100 m/z). The sample concentration was determined by calibration with external standards.

0.006 0.006 0.007 0.007 0.001 0.006

The isotopic ratios of perdeuterated and light toluene and ethylbenzene were determined by calculating the ratio of the corresponding molar concentrations, and the chlorine isotope ratios of chlorinated organic compounds were determined using the method proposed by Jin et al.38 t

∑ j = 1 i · Ij Tot( 37Cl) R Cl = = t Tot( 35Cl) ∑ j = 1 (h − i ) · I j

(1) 37

35

where RCl is the chlorine isotope ratio, Tot( Cl) and Tot( Cl) are the total number of heavy and light isotopes, Ij is the ion abundance of a certain isotopologue (j), and h and i are the total number of chlorine atoms and the number of heavy chlorine atoms in a certain molecular ion. Modeling of Diffusive Isotope Fractionation. Diffusion of dissolved organic contaminants in the glass tubes filled with agarose gel is described by Fick’s second law in one dimension ∂c ∂ 2c = Dav 2 ∂t ∂x

(2)

where c = c(x,t) is the contaminant concentration, which is a function of space (x) and time (t), and Dav is the diffusion coefficient of the organic compound. Tracking independently the concentration of each isotopologue, eq 2 can be expressed as ∂cj ∂t

= Dj

∂ 2cj ∂x 2

(3) th

where cj is the concentration of the j isotopologue, and Dj is the isotopologue-specific diffusion coefficient. The fundamental interpretation of diffusion coefficients in dense fluid systems such as aqueous solutions is complex and has generated considerable debate in the literature.47 In particular, concerning the mass dependence of diffusion coefficients in dilute solutions two different pictures are given by hydrodynamic and kinetic theories. The former is based on Navier−Stokes hydrodynamics and predicts diffusion coefficients independent of the molecular mass of the solute, whereas the latter is based on binary collision dynamics and predicts an inverse square root dependence. More recent theoretical developments48 as well as experimental observations (e.g., refs 23−25 and 49) suggest that these theories provide two limiting cases describing the mass dependence of aqueous diffusion coefficients. A simple and practical model that has been proposed to capture the effect of mass dependence on diffusive isotope fractionation is the inverse power law relationship: Dj ∝ mj−β. In this study we adopt this model which, for any two distinct isotopologues of an organic compound, reads as β D1 ⎛ m2 ⎞ =⎜ ⎟ D2 ⎝ m1 ⎠

6143

(4)



Article

Figure 1. Spatial gradients of isotopologue concentrations (A1-D1 for toluene and E1-H1 for ethylbenzene) and ratios (A2-D2 for toluene and E2-H2 for ethylbenzene) in the diffusion experiments with BTEX compounds. The symbols (squares: light (nondeuterated) isotopologue; triangles: heavy (perdeuterated) isotopologue) represent the measured concentrations (normalized with respect to the initial source concentration of the light isotopologue) and isotope ratios (circles); the solid lines are the best-fit model results. The time frame of the experiments was between 10 and 12.8 days. The letters used to label the panels correspond to the experiments reported in Table 1, and the gray areas delineate the location of the source zone.

where D1 and D2 are the diffusion coefficients of the two isotopologues with molecular masses m1 and m2. β is the inverse power law exponent, which equals 0.5 in the ideal case of kinetic gas theory. The magnitude of β and the value of this

coefficient used to model diffusive fractionation in aqueous systems caused some discussion in the geochemistry literature.22,50,51 In the field of contaminant hydrology there is a lack of experimental data, and a power law exponent of 0.5 has been 6144



Article

heavy (red triangle) toluene and ethylbenzene isotopologues show clear spatial gradients along the direction of the diffusive flux, from the source zone toward the initially pristine gel medium. Since the light isotopologues tend to diffuse slightly faster than the heavy ones, the isotopic ratio between the heavy and the light isotopologues changed with the distance from the interphase along the tube. Diffusive isotope fractionation results in a significant decrease of the isotope ratio (up to 0.08) for both toluene (Figure 1, A2-D2) and ethylbenzene (Figure 1, E2-H2). We fitted the one-dimensional diffusion model (eq 2) to the experimental data to estimate the diffusion coefficients of the light and heavy isotopologues (DL and DH), and we obtained the inverse power law exponent (β) expressing the dependence of D on the molecular mass. The outcomes of the evaluation are summarized in Table 1. The determined diffusion coefficients for the light and heavy isotopologues show a consistent trend with always slightly higher values obtained from the fitting of the concentrations of the light isotopologues. The statistical significance of the differences between the values of DL and DH is also substantiated by a nonparametric test (Welch’s t test). This two sample t test showed that the determined DL and DH coefficients are data from populations with different means. The result was confirmed at 5% significant level for the entire ethylbenzene data set and for the toluene data set excluding experiment A, for which a higher scatter of the nonlabeled compound and a less good agreement between the measured and simulated isotope ratios can be noticed in comparison to the other experiments (Figure 1). The average values from the sets of experiments are 0.805 × 10−9 m2/s and 0.775 × 10−9 m2/s for toluene (experiments A-D) and 0.757 × 10−9 m2/s and 0.727 × 10−9 m2/s for ethylbenzene (experiments E-H). The higher diffusivities of the nonlabeled compared to the labeled isotopologues substantiate, for both toluene (DL/DH = 1.039) and ethylbenzene (DL/DH = 1.041), the isotopic diffusive fractionation observed in the investigated experimental setup. Furthermore, the magnitude of the diffusion coefficient values is consistent with the values computed using the empirical correlations proposed by Wilke and Chang52 and Worch.53 The experimental results corrected by the effective volume fraction of the gel and the coefficient of obstruction according to Lauffer35 as well as the values computed using the abovementioned empirical correlations are reported in Table S1 of the Supporting Information. Only a minimal tortuosity adjustment is required (τ2 = 1.037), which has no influence on the quantification of the extent of diffusive isotope fractionation (eq S8, Supporting Information). An average β value of 0.455 was obtained for both toluene and ethylbenzene, close to the upper square root limit. This value is considerably larger than the ones reported for ionic species24 and also larger than the data reported for most noble gases19,49 with the exception of the measurements of helium in water in the experiments of Jähne et al.20 Diffusive Isotope Fractionation of Chlorinated Compounds at Natural Abundance. The experiments with chlorinated compounds were conducted in the same experimental setup using cis-DCE and TCE at their natural isotopic abundance and measuring the evolution of the chlorine isotope ratio. The measured concentrations and isotope ratios are reported in Figure 2. The results are qualitatively similar to the outcomes for the aromatic hydrocarbons. The concentrations of both cis-DCE and TCE show a diffusive profile from the source zone toward

used in recent modeling studies on transport of organic contaminants in groundwater in the context of CSIA.13,16,31 In this study we determine the value of the diffusion coefficients as well as the parameter β from high-resolution diffusion experiments. In the experiments with petroleum hydrocarbons each light and heavy (perdeuterated) isotopologue could be measured individually. Therefore, to evaluate the experimental results, we solved eq 3 for each toluene and ethylbenzene isotopologue in a one-dimensional system with the same geometry, initial and boundary conditions as the experimental setup. We used a numerical approach with an implicit finite-difference scheme to solve the forward problem and an automated procedure to fit the model to the isotopologues concentrations, using Dj as fitting parameter. In this way only the concentrations of the two isotopologues are used to determine the isotopologue-specific diffusion coefficients, and the isotope ratios predicted by the model can be directly compared with the experimental observations. The fitting procedure was implemented using the capability of the Matlab function lsqnonlin to solve nonlinear least-squares problems. Details on the optimization procedure and on error propagation are provided in the Supporting Information. In the experiments with chlorinated compounds at natural isotopic abundances, we could not measure independently the concentrations of each individual isotopologue, since there is no physical separation prior to detection in the mass spectrometer. Therefore, eq 2 was used to describe the spatial and temporal evolution of cis-DCE and TCE and to fit the values of Dav from the measured concentration data. The average diffusion coefficient of a given compound can be expressed by the geometric mean of the individual diffusion coefficients of each isotopologue j, weighted by their relative abundances (Fj): Dav =

1 F1 D1

+

F2 D2

+

F3 D3

+ ... +

Fj Dj

(5)

Assuming that, for the isotopologues of a given compound, different isotopic substitutions in different molecular positions result in the same extent of diffusive isotope fractionation, the power law relationship among different isotopologue pairs can be described using the same exponent β. Thus, combining eq 4 and eq 5, the isotopologue-specific diffusion coefficients (Dj) can be written as ⎛ ⎛ ⎞β ⎞ ⎛ m ⎞β ⎛ mj − 1 ⎞ β m ⎟⎟ + Fj⎟ Dj = Dav ·⎜⎜F1·⎜⎜ 1 ⎟⎟ + F2·⎜⎜ 2 ⎟⎟ + ···+ Fj − 1·⎜⎜ ⎟ m ⎝ mj ⎠ ⎝ mj ⎠ ⎝ ⎝ j⎠ ⎠

(6)

Using the Dav obtained from the measured concentrations and knowing the relative abundances of the chlorine isotopologues of cis-DCE and TCE as well as the measured isotope ratios, the value of the exponent β is determined for each experimental run.

■

RESULTS AND DISCUSSION Diffusive Isotope Fractionation of Isotopically Labeled Petroleum Hydrocarbons. Diffusion experiments using 1:2 mixtures of labeled (perdeuterated) and nonlabeled toluene and ethylbenzene were conducted, and the measured concentration and isotopic gradients are shown in Figure 1. All experiments showed consistent results and were highly reproducible. The concentration of both light (blue square) and 6145



Article

Figure 2. Spatial gradients of concentrations (I1-K1 for cis-DCE and L1-N1 for TCE) and isotope ratios (I2-K2 for cis-DCE and L2-N2 for TCE) in the diffusion experiments with chlorinated ethenes. The symbols represent the measured concentrations (squares) and chlorine isotope ratios (circles) sampled at different days; the solid lines are the best-fit model results (see Table 2 for further details). The time frame of the experiments was between 7 and 12.8 days. The letters used to label the panels correspond to the experiments reported in Table 2, and the gray areas delineate the location of the source zone.

in each experimental run (Figure 2). The change in chlorine isotope ratio is larger for cis-DCE than for TCE. The extent of change varies in a range from 0.0020 (experiment K, panel K2)

the initially pristine gel medium. Significant spatial isotope gradients are obtained due to the different displacement of cis-DCE and TCE isotopologues and can be clearly observed 6146



Article

to 0.0035 (experiment I, panel I2) for cis-DCE and from 0.0007 (experiment N, panel N2) to 0.0011 (experiment L, panel L2) for TCE. The values of D and β were determined with the fitting procedure outlined above, based on eqs 2, 5, and 6. The values of diffusion coefficients and β obtained in the different experiments are reported in Table 2, whereas the detailed isotopologue-specific properties including the computed relative abundances of cis-DCE and TCE chlorine isotopologues and their aqueous diffusivity are summarized in Table S2 (Supporting Information). The average diffusion coefficients are 1.108 × 10−9 m2/s for cis-DCE and 0.898 × 10−9 m2/s for TCE. As in the case of toluene and ethylbenze, also the determined diffusion coefficients for the considered chlorinated compounds are in fairly good agreement with the values estimated using published empirical correlations (Table S1, Supporting Information). The average values of β for the chlorinated ethenes are 0.088 for cis-DCE and 0.043 for TCE. These values are larger but in the same order of magnitude of the ones reported for monovalent cations and anions,25 thus pointing to similar but less intense electrostatic interactions of the solute with the surrounding water molecules. Rationalizing, in analogy with mechanisms proposed for ionic species,54 that the solvation molecules follow the solute motion, the observed attenuated mass dependence of diffusive fractionation can be attributed to properties of the hydration shell such as its volume, structure, and residence time of water molecules. Simulations using the experimentally determined D and β values were performed to predict the spatial and temporal evolution of cis-DCE and TCE isotope ratios for simulation times that go significantly beyond the time frame of our experiments. The results are reported in the Supporting Information and besides the change in isotope ratios in the initially pristine medium also show a small enrichment in the source zone (Figure S4). Figure S4 reports the extent of diffusive isotope fractionation as chlorine isotope ratios and referenced δ37Cl scale. Referencing chlorine isotope data is important in particular to compare measurements using different instruments.39 However, linear deviations found for different analytical instruments and their relative corrections only affect the absolute values of chlorine isotope ratios but not their relative differences and the evaluation of the parameter β. In comparison to the results for toluene and ethylbenzene, the values of β found for cis-DCE and TCE are considerably smaller. A possible reason for such remarkable dissimilarity is that the deuterium substitutions in the case of labeled hydrocarbon molecules do not only imply a significant difference in the mass but also in the size of the toluene and ethylbenzene molecules. Recently, it was shown by high-resolution neutron powder diffractometry and by theoretical arguments55 that, starting from a temperature of 170 K, the molecular volume of fully deuterated benzene (C6D6) was greater than the one of the light molecule (C6H6). This could explain the large difference in self-diffusion coefficients between C6H6 and C6D6 reported in the early work of Mills56 as well as the extensive diffusive fractionation between labeled and nonlabeled petroleum hydrocarbons (toluene: C7H8/C7D8; ethylbenzene: C8H10/C8D10) observed in our diffusion experiments and quantified by an average β of 0.455 (Table 1). Such explanation and the finding of large β values are also in agreement with observations on aqueous diffusion of noble gases. In fact it was noted that the apparent mass-dependence of D for the major noble gas isotopes, where both solute mass and radius increased,

followed an inverse square root relationship and was larger than the true mass dependence of D for a set of isotopes of a single noble gas where mass increased but radius did not.19 The marked differences of β values between aromatic hydrocarbons and chlorinated ethenes can be further rationalized by the differences in polarity of these compound classes leading to different solute−solvent interactions and thus different effective volumes of the hydrated species. In fact, as mentioned above, the chlorine substitutions in the chlorinated molecules results in high polarity of the C−Cl bond and in uneven charge distribution within the molecules. This significantly affects the interactions of the chlorinated ethenes molecules with the water molecules in the hydration shell. A similar effect was reported for the hydration of peptides, for which the presence of polar groups and resulting partial charges has the effect of constricting solute molecules in the vicinity of such groups.57 Such effects strengthen and stabilize the hydration shell and result in more attenuated diffusive fractionation effects for the chlorinated ethenes. These arguments, although still rather qualitative and requiring further evidence, might also explain the relative difference between the β values of cis-DCE and TCE, with the latter being smaller due to stronger solute− solvent interaction. Environmental Significance. The development of significant diffusive isotope gradients during transport of organic contaminants has important environmental implications. For instance, many subsurface environments (e.g., clay and silt aquitards, unweathered and unfractured rocks, lake and marine sediments etc.) are diffusion-dominated systems for which a quantitative understanding of diffusive isotope fractionation is of primary importance. Specifically, the outcomes of our experiments with labeled compounds are relevant for the study of in situ processes in BTEX-contaminated aquifers since mixtures of deuterated and nondeuterated compounds are applied in both natural gradient and push−pull tracer tests to identify and quantify physical and transformation processes (e.g., refs 58−60). The results for chlorinated ethenes at their natural isotopic abundance have direct implications for the study of the fate and transport of these widespread organic contaminants in subsurface environments. To project the experimental findings of this study to larger scales we consider a modeling scenario of conservative transport in an unbounded aquitard. Simulations of diffusive transport were performed in a one-dimensional domain representing a vertical profile in a low-permeability medium. In the simulations, the contaminant source containing cis-DCE and TCE was initially located at the bottom 0.5 m of the profiles (Figure 3). The parameter D and β were selected identical to the ones experimentally determined in this study with the only exception that, in the simulations, the effective aqueous diffusion coefficients were corrected by a tortuosity of 2.5, taking into account that in this scenario diffusion occurs only through the void space of the porous medium. The resulting spatial isotopic gradients computed after a 10-year diffusion period are shown in Figure 3. Expressing the Cl isotope ratio in delta notation, significant Δδ37Cl gradients were obtained for both cis-DCE and TCE with more depleted values in the initially pristine region at the top, toward which the contaminants diffuse, and slightly enriched values at the bottom due to the faster migration of light isotopologues, which tend to leave the source zone at a slightly faster diffusive rate. We also compared these outcomes with the prediction of simulations based on the inverse square root relationship between the diffusion coefficients of the different isotopologues and their molecular 6147



Article

isotopic fractionation of organic compounds in aqueous solution. This will allow developing compound-specific molecular descriptors for different organic chemicals that could be directly correlated with the extent of observed diffusive isotope fractionation. Such advances will contribute to the development of a mechanistic theory of this environmentally relevant but still poorly understood process.

■

ASSOCIATED CONTENT

* Supporting Information S

Schematic descriptions of the experimental protocol, the parameter-fitting approach, the comparison of the experimental results with the values of aqueous diffusion coefficients based on empirical correlations, and a summary of the isotopologuespecific properties of the compounds at natural isotopic abundance. This material is available free of charge via the Internet at http://pubs.acs.org.

■

Figure 3. Diffusion-induced chlorine isotope fractionation in a lowpermeability subsurface layer. The continuous lines represent the values for cis-DCE and TCE, after a simulation time of 10 years, with the parameters D and β determined in this study; the dashed lines represent the results of simulations with the same D but β = 0.5 assuming an inverse square root relationship and the dash-dot line is the case with no diffusive isotope fractionation. The gray area delineates the location of the initial source zone.

AUTHOR INFORMATION

Corresponding Author

*Phone: 6507211118. E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank three anonymous reviewers for their detailed comments that significantly helped improve the manuscript. B.J. acknowledges his Ph.D. scholarship from the IPSWaT program of the German Federal Ministry of Education and Research (BMBF). M.R. acknowledges the support of the Marie Curie International Outgoing Fellowship (DILREACT project) within the seventh European Community Framework Programme. M.R. and S.B.H. also acknowledge the support of the Deutsche Forschungsgemeinschaft (Grant RO4169/2-1 and ENST 37/740-1).

masses. With the latter approach the extent of chlorine isotope depletion in the initially pristine region at the top of the domain as well as the isotopic enrichment in the source zone would be remarkably overestimated compared to the results of the model based on the β values determined in our diffusion experiments. The outcomes of the present work help understanding and quantifying the significance of diffusive isotope fractionation of organic contaminants in various environmental settings. Although this process might be as important as other phenomena (e.g., isotope fractionation during contaminant transformation), it has received little attention by the scientific community. Our high-resolution data for BTEX and chlorinated ethenes show first experimental evidence of the relevance and complexity of diffusive isotope fractionation for organic contaminants of primary concern. Due to the scarcity of currently available data on diffusive isotope fractionation of organic compounds in water and to the crucial importance of diffusion for contaminant mass transport in many environmental systems such as geologic formations, further research is required to extend the investigation to other organic contaminant classes and to spatial and temporal scales larger than the ones considered in the experimental investigation carried out in this study. In particular, since our experimental results clearly show a distinct extent of diffusive fractionation for different compounds, a highresolution database including most of the organic pollutants frequently found in groundwater systems will be instrumental for an integrated interpretation of isotopic signatures and for the quantitative application of CSIA taking into account the fractionation effects of both physical and transformation processes. Moreover, we think that further experimental investigations and modeling studies using tools and approaches of a broad range of scientific disciplines (e.g., from theoretical and computational chemistry to applied contaminant hydrogeology) will help to shed light on the role of factors such as the molecular structure of the solute, temperature, polarity, and the interaction between solute and solvents molecules, on diffusive

■

REFERENCES

(1) Elsner, M.; Zwank, L.; Hunkeler, D.; Schwarzenbach, R. P. A new concept linking observable stable isotope fractionation to transformation pathways of organic pollutants. Environ. Sci. Technol. 2005, 39 (18), 6896−6916. (2) Elsner, M. Stable isotope fractionation to investigate natural transformation mechanisms of organic contaminants: principles, prospects and limitations. J. Environ. Monit. 2010, 12 (11), 2005− 2031. (3) Schmidt, T. C.; Zwank, L.; Elsner, M.; Berg, M.; Meckenstock, R. U.; Haderlein, S. B. Compound-specific stable isotope analysis of organic contaminants in natural environments: a critical review of the state of the art, prospects, and future challenges. Anal. Bioanal. Chem. 2004, 378 (2), 283−300. (4) Thullner, M.; Centler, F.; Richnow, H. H.; Fischer, A. Quantification of organic pollutant degradation in contaminated aquifers using compound specific stable isotope analysis - Review of recent developments. Org. Geochem. 2012, 42 (12), 1440−1460. (5) Hatzinger, P. B.; Bohlke, J.; Sturchio, N. C. Application of stable isotope ratio analysis for biodegradation monitoring in groundwater. Curr. Opin. Biotechnol. 2013, 24 (3), 542−9. (6) Thullner, M.; Kampara, M.; Richnow, H. H.; Harms, H.; Wick, L. Y. Impact of bioavailability restrictions on microbially induced stable isotope fractionation. 1. Theoretical calculation. Environ. Sci. Technol. 2008, 42 (17), 6544−6551. (7) Tobler, N. B.; Hofstetter, T. B.; Schwarzenbach, R. P. Carbon and Hydrogen Isotope Fractionation during Anaerobic Toluene Oxidation by Geobacter metallireducens with Different Fe(III) Phases as Terminal Electron Acceptors. Environ. Sci. Technol. 2008, 42 (21), 7786−7792.

6148



Article

former brackish lagoon in The Netherlands. Appl. Geochem. 2011, 26 (3), 257−268. (28) Rasa, E.; Chapman, S. W.; Bekins, B. A.; Fogg, G. E.; Scow, K. M.; Mackay, D. M. Role of back diffusion and biodegradation reactions in sustaining an MTBE/TBA plume in alluvial media. J. Contam. Hydrol. 2011, 126 (3−4), 235−247. (29) Mackay, D. M.; Cherry, J. A. Groundwater Contamination Pump-and-Treat Remediation. 2. Environ. Sci. Technol. 1989, 23 (6), 630−636. (30) Hunkeler, D.; Chollet, N.; Pittet, X.; Aravena, R.; Cherry, J. A.; Parker, B. L. Effect of source variability and transport processes on carbon isotope ratios of TCE and PCE in two sandy aquifers. J. Contam. Hydrol. 2004, 74 (1−4), 265−282. (31) Hunkeler, D.; Aravena, R.; Shouakar-Stash, O.; Weisbrod, N.; Nasser, A.; Netzer, L.; Ronen, D. Carbon and Chlorine Isotope Ratios of Chlorinated Ethenes Migrating through a Thick Unsaturated Zone of a Sandy Aquifer. Environ. Sci. Technol. 2011, 45 (19), 8247−8253. (32) Rolle, M.; Hochstetler, D.; Chiogna, G.; Kitanidis, P. K.; Grathwohl, P. Experimental Investigation and Pore-Scale Modeling Interpretation of Compound-Specific Transverse Dispersion in Porous Media. Transp. Porous Media 2012, 93 (3), 347−362. (33) Hochstetler, D. L.; Rolle, M.; Chiogna, G.; Haberer, C. M.; Grathwohl, P.; Kitanidis, P. K. Effects of compound-specific transverse mixing on steady-state reactive plumes: Insights from pore-scale simulations and Darcy-scale experiments. Adv. Water Resour. 2013, 54, 1−10. (34) Rolle, M.; Muniruzzaman, M.; Haberer, C. M.; Grathwohl, P. Coulombic effects in advection-dominated transport of eletrolytes in porous media: Multicomponent ionic dispersion. Geochim. Cosmochim. Acta 2013, 120, 195−205. (35) Lauffer, M. A. Theory of diffusion in gels. Biophys. J. 1961, 1 (3), 205−213. (36) Sakaguchi-Soder, K.; Jager, J.; Grund, H.; Matthaus, F.; Schuth, C. Monitoring and evaluation of dechlorination processes using compound-specific chlorine isotope analysis. Rapid Commun. Mass Spectrom. 2007, 21 (18), 3077−3084. (37) Aeppli, C.; Holmstrand, H.; Andersson, P.; Gustafsson, O. Direct Compound-Specific Stable Chlorine Isotope Analysis of Organic Compounds with Quadrupole GC/MS Using Standard Isotope Bracketing. Anal. Chem. 2010, 82 (1), 420−426. (38) Jin, B. A.; Laskov, C.; Rolle, M.; Haderlein, S. B. Chlorine Isotope Analysis of Organic Contaminants Using GC-qMS: Method Optimization and Comparison of Different Evaluation Schemes. Environ. Sci. Technol. 2011, 45 (12), 5279−5286. (39) Bernstein, A.; Shouakar-Stash, O.; Ebert, K.; Laskov, C.; Hunkeler, D.; Jeannottat, S.; Sakaguchi-Soder, K.; Laaks, J.; Jochmann, M. A.; Cretnik, S.; Jager, J.; Haderlein, S. B.; Schmidt, T. C.; Aravena, R.; Elsner, M. Compound-Specific Chlorine Isotope Analysis: A Comparison of Gas Chromatography/Isotope Ratio Mass Spectrometry and Gas Chromatography/Quadrupole Mass Spectrometry Methods in an Interlaboratory Study. Anal. Chem. 2011, 83 (20), 7624−7634. (40) Cincinelli, A.; Pieri, F.; Zhang, Y.; Seed, M.; Jones, K. C. Compound Specific Isotope Analysis (CSIA) for chlorine and bromine: A review of techniques and applications to elucidate environmental sources and processes. Environ. Pollut. 2012, 169, 112−127. (41) Hofstetter, T. B.; Reddy, C. M.; Heraty, L. J.; Berg, M.; Sturchio, N. C. Carbon and chlorine isotope effects during abiotic reductive dechlorination of polychlorinated ethanes. Environ. Sci. Technol. 2007, 41 (13), 4662−4668. (42) Elsner, M.; Hunkeler, D. Evaluating chlorine isotope effects from isotope ratios and mass spectra of polychlorinated molecules. Anal. Chem. 2008, 80 (12), 4731−4740. (43) Heraty, L. J.; Fuller, M. E.; Huang, L.; Abrajano, T.; Sturchio, N. C. Isotopic fractionation of carbon and chlorine by microbial degradation of dichloromethane. Org. Geochem. 1999, 30 (8A), 793−799.

(8) Aeppli, C.; Berg, M.; Cirpka, O. A.; Holliger, C.; Schwarzenbach, R. P.; Hofstetter, T. B. Influence of Mass-Transfer Limitations on Carbon Isotope Fractionation during Microbial Dechlorination of Trichloroethene. Environ. Sci. Technol. 2009, 43 (23), 8813−8820. (9) Kopinke, F. D.; Georgi, A.; Voskamp, M.; Richnow, H. H. Carbon isotope fractionation of organic contaminants due to retardation on humic substances: Implications for natural attenuation studies in aquifers. Environ. Sci. Technol. 2005, 39 (16), 6052−6062. (10) Hohener, P.; Yu, X. J. Stable carbon and hydrogen isotope fractionation of dissolved organic groundwater pollutants by equilibrium sorption. J. Contam. Hydrol. 2012, 129, 54−61. (11) Bouchard, D.; Hohener, P.; Hunkeler, D. Carbon Isotope Fractionation During Volatilization of Petroleum Hydrocarbons and Diffusion Across a Porous Medium: A Column Experiment. Environ. Sci. Technol. 2008, 42 (21), 7801−7806. (12) Jeannottat, S.; Hunkeler, D. Chlorine and Carbon Isotopes Fractionation during Volatilization and Diffusive Transport of Trichloroethene in the Unsaturated Zone. Environ. Sci. Technol. 2012, 46 (6), 3169−3176. (13) LaBolle, E. M.; Fogg, G. E.; Eweis, J. B.; Gravner, J.; Leaist, D. G. Isotopic fractionation by diffusion in groundwater. Water Resour. Res. 2008, 44, (7). (14) Rolle, M.; Chiogna, G.; Bauer, R.; Griebler, C.; Grathwohl, P. Isotopic Fractionation by Transverse Dispersion: Flow-through Microcosms and Reactive Transport Modeling Study. Environ. Sci. Technol. 2010, 44 (16), 6167−6173. (15) Eckert, D.; Rolle, M.; Cirpka, O. A. Numerical simulation of isotope fractionation in steady-state bioreactive transport controlled by transverse mixing. J. Contam. Hydrol. 2012, 140, 95−106. (16) Van Breukelen, B. M.; Rolle, M. Transverse Hydrodynamic Dispersion Effects on Isotope Signals in Groundwater Chlorinated Solvents’ Plumes. Environ. Sci. Technol. 2012, 46 (14), 7700−7708. (17) Schloemer, S.; Krooss, B. M. Molecular transport of methane, ethane and nitrogen and the influence of diffusion on the chemical and isotopic composition of natural gas accumulations. Geofluids 2004, 4 (1), 81−108. (18) Risk, D.; Kellman, L. Isotopic fractionation in non-equilibrium diffusive environments. Geophys. Res. Lett. 2008, 35, (2). (19) Bourg, I. C.; Sposito, G. Isotopic fractionation of noble gases by diffusion in liquid water: Molecular dynamics simulations and hydrologic applications. Geochim. Cosmochim. Acta 2008, 72 (9), 2237−2247. (20) Jähne, B.; Heinz, G.; Dietrich, W. Measurement of the diffusion coefficients of sparingly soluble gases in water. J. Geophys. Res.: Oceans 1987, 92 (C10), 10767−10776. (21) Bourg, I. C.; Sposito, G. Molecular dynamics simulations of kinetic isotope fractionation during the diffusion of ionic species in liquid water. Geochim. Cosmochim. Acta 2007, 71 (23), 5583−5589. (22) Donahue, M. A.; Werne, J. P.; Meile, C.; Lyons, T. W. Modeling sulfur isotope fractionation and differential diffusion during sulfate reduction in sediments of the Cariaco Basin. Geochim. Cosmochim. Acta 2008, 72 (9), 2287−2297. (23) Eggenkamp, H. G. M.; Coleman, M. L. The effect of aqueous diffusion on the fractionation of chlorine and bromine stable isotopes. Geochim. Cosmochim. Acta 2009, 73 (12), 3539−3548. (24) Richter, F. M.; Mendybaev, R. A.; Christensen, J. N.; Hutcheon, I. D.; Williams, R. W.; Sturchio, N. C.; Beloso, A. D. Kinetic isotopic fractionation during diffusion of ionic species in water. Geochim. Cosmochim. Acta 2006, 70 (2), 277−289. (25) Bourg, I. C.; Richter, F. M.; Christensen, J. N.; Sposito, G. Isotopic mass dependence of metal cation diffusion coefficients in liquid water. Geochim. Cosmochim. Acta 2010, 74 (8), 2249−2256. (26) Wortmann, U. G.; Chernyavsky, B. M. The significance of isotope specific diffusion coefficients for reaction-transport models of sulfate reduction in marine sediments. Geochim. Cosmochim. Acta 2011, 75 (11), 3046−3056. (27) Beekman, H. E.; Eggenkamp, H. G. M.; Appelo, C. A. J. An integrated modelling approach to reconstruct complex solute transport mechanisms - Cl and delta Cl-37 in pore water of sediments from a 6149



Article

(44) Numata, M.; Nakamura, N.; Koshikawa, H.; Terashima, Y. Chlorine isotope fractionation during reductive dechlorination of chlorinated ethenes by anaerobic bacteria. Environ. Sci. Technol. 2002, 36 (20), 4389−4394. (45) Hunkeler, D.; Van Breukelen, B. M.; Elsner, M. Modeling Chlorine Isotope Trends during Sequential Transformation of Chlorinated Ethenes. Environ. Sci. Technol. 2009, 43 (17), 6750−6756. (46) Jin, B.; Haderlein, S. B.; Rolle, M. Integrated carbon and chlorine isotope modeling: applications to chlorinated aliphatic hydrocarbons dechlorination. Environ. Sci. Technol. 2013, 47 (3), 1443−51. (47) Tyrrell, H. J. V.; Harris, K. R. Diffusion in Liquids. A theoretical and experimental study; Butterworths: London, 1984. (48) Bhattacharyya, S.; Bagchi, B. Power law mass dependence of diffusion: A mode coupling theory analysis. Phys. Rev. E 2000, 61 (4), 3850−3856. (49) Tempest, K. E.; Emerson, S. Kinetic isotopic fractionation of argon and neon during air-water gas transfer. Mar. Chem. 2013, 153, 39−47. (50) Donahue, M. A.; Werne, J. P.; Meile, C.; Lyons, T. W. Response to comment by I.C. Bourg on ″Modeling sulfur isotope fractionation and differential diffusion during sulfate reduction in sediments of the Cariaco Basin″ by M.A. Donahue, J.P. Werne, C. Meile, and T.W. Lyons. Geochim. Cosmochim. Acta 2008, 72 (23), 5855−5856. (51) Bourg, I. C. Comment on ″Modeling sulfur isotope fractionation and differential diffusion during sulfate reduction in sediments of the Cariaco Basin″ by M.A. Donahue, J.P. Werne, C. Meile and T.W. Lyons. Geochim. Cosmochim. Acta 2008, 72 (23), 5852−5854. (52) Wilke, C. R.; Chang, P. C. AIChE J. 1955, 1, 264. (53) Worch, E. A new equation for the calculation of diffusion coefficients for dissolved substances. In 1993; Vol. 81, pp 289−297. (54) Møller, K. B.; Rey, R.; Masia, M.; Hynes, J. T. On the coupling between molecular diffusion and solvation shell exchange. J. Chem. Phys. 2005, 122, (11). (55) Dunitz, J. D.; Ibberson, R. M. Is deuterium always smaller than protium? Angew. Chem., Int. Ed. 2008, 47 (22), 3. (56) Mills, R. Diffusion relations in the binary system benzeneperdeuteriobenzene at 25 °C. J. Phys. Chem. 1976, 80 (8), 888−890. (57) Paterson, Y.; Nemethy, G.; Scheraga, H. A. Hydration of amino acids, peptides and model compounds. Ann. N. Y. Acad. Sci. 1981, 367, 132−150. (58) Thierrin, J.; Davis, G. B.; Barber, C. A Groundwater Tracer Test with Deuterated Compounds for Monitoring In-Situ Biodegradation and Retardation of Aromatic Hydrocarbons. Ground Water 1995, 33 (3), 469−475. (59) Reusser, D. E.; Istok, J. D.; Beller, H. R.; Field, J. A. In situ transformation of deuterated toluene and xylene to benzylsuccinic acid analogues in BTEX-contaminated aquifers. Environ. Sci. Technol. 2002, 36 (19), 4127−4134. (60) Gieg, L. M.; Alumbaugh, R. E.; Field, J.; Jones, J.; Istok, J. D.; Suflita, J. M. Assessing in situ rates of anaerobic hydrocarbon bioremediation. Microb. Biotechnol. 2009, 2 (2), 222−233.

6150


Diffusive Fractionation of BTEX and Chlorinated Ethenes in Aqueous

Recommend Documents