Carbon Isotopic Fractionation of CFCs during ... - ACS Publications

Article pubs.acs.org/est

Carbon Isotopic Fractionation of CFCs during Abiotic and Biotic Degradation Marie E. Archbold, Trevor Elliot,* and Robert M. Kalin† Environmental Engineering Research Centre (EERC), Queen’s University Belfast, School of Planning, Architecture & Civil Engineering, David Keir Building, Stranmillis Road, Belfast, Northern Ireland, BT9 5AG, United Kingdom S Supporting Information *

ABSTRACT: Carbon stable isotope (13C) fractionation in chlorofluorocarbon (CFC) compounds arising from abiotic (chemical) degradation using zero-valent iron (ZVI) and biotic (landfill gas attenuation) processes is investigated. Batch tests (at 25 °C) for CFC-113 and CFC-11 using ZVI show quantitative degradation of CFC-113 to HCFC123a and CFC-1113 following pseudo-first-order kinetics corresponding to a half-life (τ1/2) of 20.5 h, and a ZVI surface-area normalized rate constant (kSA) of −(9.8 ± 0.5) × 10−5 L m−2 h−1. CFC-11 degraded to trace HCFC-21 and HCFC-31 following pseudofirst-order kinetics corresponding to τ1/2 = 17.3 h and kSA = −(1.2 ± 0.5) × 10−4 L m−2 h−1. Significant kinetic isotope effects of ε(‰) = −5.0 ± 0.3 (CFC-113) and −17.8 ± 4.8 (CFC-11) were observed. Compound-specific carbon isotope analyses also have been used here to characterize source signatures of CFC gases (HCFC-22, CFC-12, HFC-134a, HCFC-142b, CFC-114, CFC-11, CFC-113) for urban (UAA), rural/remote (RAA), and landfill (LAA) ambient air samples, as well as in situ surface flux chamber (FLUX; NO FLUX) and landfill gas (LFG) samples at the Dargan Road site, Northern Ireland. The latter values reflect biotic degradation and isotopic fractionation in LFG production, and local atmospheric impact of landfill emissions through the cover. Isotopic fractionations of Δ13C ∼ −13‰ (HCFC-22), Δ13C ∼ −35‰ (CFC-12) and Δ13C ∼ −15‰ (CFC-11) were observed for LFG in comparison to characteristic solvent source signatures, with the magnitude of the isotopic effect for CFC-11 apparently similar to the kinetic isotope effect for (abiotic) ZVI degradation.

■

INTRODUCTION Chlorofluorocarbons (or freons) have been manufactured since the early 1930s and used worldwide as aerosol propellants, refrigerants, foam-blowing agents, and solvents before their phasing out as a result of the 1987 Montreal Protocol on Substances that Deplete the Ozone Layer. However, there are now a number of CFC-contaminated aquifers1 requiring remediation intervention strategies. Landfill gas (LFG) also invariably contains trace gases including volatile organic compounds (VOCs) such as hydrocarbons, aromatics, and halogenated hydrocarbons which can migrate in the subsurface and vent through the surrounding soil. Of the halogenated VOCs, CFCs typically account for up to 95% of the total chlorine measured in LFG2 predominantly from the volatilization of compounds contained in waste refrigerators, plastic foam, and aerosol propellants. Emissions of substantial levels of landfill CFCs may continue for up to 50 years after waste has been deposited in the landfill due to slow release from waste sources,3,4 potentially impacting both the surrounding atmosphere and groundwaters.5 Dependent on the magnitude of fractionation and/or the kinetics of reaction, compoundspecific stable isotope analyses generically can help determine the provenance of organic compounds and identify transformation pathways (biotic and abiotic) in the environment.6 © 2011 American Chemical Society

Moreover, if contaminant compounds are to be remediated in situ by engineered technologies such as flow-through permeable reactive barriers (PRBs) in groundwaters7 or cap systems in landfills,8 monitoring difficulties may need to be surmounted, cf. identifying contaminant bypass (via in situ fracture flow) or incomplete reaction due to insufficient reactor thickness, for which compound-specific isotopic fractionation may be deployed as a monitoring tool to determine the extent of reaction. Enhanced bioremediation and cleanup of contaminant CFCs in situ in groundwaters has been attempted using, e.g., hydrogen-release compound.9 Flow-through PRBs containing zero-valent iron (ZVI or Fe0) as a passive remediation strategy also have been deployed generally to abiotically treat groundwaters contaminated with dissolved chlorinated solvents.10,11 Scheutz et al.8 have shown that ZVI is capable of degrading CFC-11 under saturated batch (1 g of ZVI:37 mL of water; half-life ∼22 h) and column study, although the degradation products were not identified. Vidumsky and Received: Revised: Accepted: Published: 1764

September 26, 2011 December 19, 2011 December 21, 2011 December 21, 2011 dx.doi.org/10.1021/es203386a | Environ. Sci. Technol. 2012, 46, 1764−1773

Environmental Science & Technology

Article

Thomson12 have conducted column treatability studies to evaluate the potential effectiveness of ZVI technologies for treating CFC-11 and CFC-113, which identified reductive dechlorination of CFC-11 to HCFC-21 and HCFC-31 and CFC-113 to HCFC-123a and CFC-1113. Further summary detail and a schematic of potential pathways of degradation for CFC-113 and CFC-11 using ZVI are given in the Supporting Information (SI − Potential Degradation Products of CFC-113 and CFC-11 using ZVI) and Figure S1. Vidumsky and Thomson12 did not report on the kinetics of degradation, only the products and pathway identification, although a later report cites rapid kinetics (Connelly-GPM iron; half-lives ∼3.55 h for CFC-113 and ∼0.9 h for CFC-11; see ref 13); nevertheless, since they used throughflow columns rather than batch study, any (dynamic) rate constants reported likely could be without meaningful direct correlation to field situations. Moreover, the potential for using specifically compound stable (carbon) isotopic signatures to monitor the degradation process has not been investigated to date. The potential for using compound stable (carbon) isotopic signatures of CFCs are also investigated here following source characterizations within and around a landfill site in Northern Ireland. Sources of CFCs both in the atmosphere and LFG are discussed in further detail in SI (CFC Sources in Atmospheric and Landfill Gases). LFG invariably contains trace compounds including CFCs which, however, may biodegrade and naturally attenuate under the strongly reducing conditions potentially produced within a landfill. Under simulated anaerobic landfill conditions it has been shown that CFC-12 biodegrades predominantly to HCFC-22 (which itself can degrade), and CFC-11 sequentially to HCFC-21, HCFC-31, and possibly HCF-41,14,15 although the effects in situ are only recently being investigated (cf. ref 16). Potentially then, carbon stable isotopes again may be used here to monitor and discriminate, e.g., CFCs volatilizing directly from solvent sources versus those subject to (bio)degradative attenuation; since the molecular weights of the CFCs are all relatively large (∼100 g/mol), the relative mass differences between their carbon isotopic species are small such that for the same kinetic energy the ratio of their velocities during volatilization correspondingly should be close to unity thus restricting fractionation effects. The isotopic data presented here for urban air samples from Belfast originally provided by Archbold17 for HCFC-22, CFC12, HCFC-134a, CFC-11, and CFC-113 already have been extended to include concentration data by Redeker et al.18 for source air mass characterization per se. The Belfast urban atmospheric samples data reported here in the current paper form a subset of these Irish atmospheric values which, however, relate only to samples taken during the period of the landfill study and are used herein to characterize the local environmental impacts. Prior to these studies, an atmospheric δ13C value of −23.3 ± 1.6‰ (2σ, standard error) had reported for baseline CFC-113 sampled in both the Northern and Southern Hemispheres.19 Average values in Bristol Urban Area of δ13C = −33.5‰ ± 0.8‰ (n = 14) for CFC-12 and δ13C = −33.9‰ ± 1.0‰ (n = 14) for HCFC-22 have also been reported.20 However, apart from the work presented by these authors, no other atmospheric δ13C values for CFCs, HCFCs, and HFCs have been described (including CFC-114 as reported here), and particularly no data have yet been presented associated with isotopically characterizing in situ LFG sources and local atmospheric impacts as reported here.

Therefore, as abiotic and biologic studies have yet to show whether significant isotope fractionations indeed occur for the CFCs it is not yet possible to carry out mass balance and source apportionment calculations. The potential for carbon stable isotopic fractionation signatures arising in chlorofluorocarbon (CFC) compounds as a result of abiotic and biotic processes is investigated in the current study. Abiotic (chemical degradation) laboratory batch tests have been conducted both on CFC113 and CFC-11 to determine rates of reaction, degradation products, and isotopic fractionations following degradative action using ZVI. Although specific microbial species responsible for biotic degradation of CFCs have not been identified, degradation under sulfate-reducing and methanogenic conditions generally are recorded, thus carbon isotopic signatures of CFCs (especially CFC-12) are also juxtaposed here following source characterization within and around a landfill site in Northern Ireland to characterize biotic source signatures (following landfill gas attenuation and emissions under methanogenic conditions) and to highlight also local environmental (atmospheric) impacts.

■

MATERIALS AND METHODS ZVI Batch Studies. A batch test methodology was adapted from that of Gillham and O’Hannesin.21 Full details are given in the Supporting Information. Two separate stock solutions (Scott Specialty Gases, Sheffield, UK) of CFC-11 (33 ppm) and CFC-113 (90 ppm) were prepared. Five g of ZVI (Connelly, Chicago, IL, ETI CC-1004, 90%) was weighed and placed in 20-mL vials. Samples were stored in the dark in a rotary incubator (Sanyo, UK) held at 25 °C and at 100 rpm. Reaction progress was then monitored by sacrificing vials for analyses over the course of the batch test. Two controls also were monitored: vials filled with stock solution only, and vials containing 5 g of ZVI with only water added. Duplicate controls and triplicate sample analyses were made at each batch sampling point. Headspaced samples (see SI − Experimental Set-up) were shaken, then placed inverted in a 25 °C incubator and left for 3 h to allow the CFCs in solution to equilibrate prior to subsampling of the headspace for analysis. Both pH and Eh (redox, mV) evolution were monitored during the batch tests using a combined pH and Eh meter (TPS model 90FLMV microprocessor field analyzer). The pH measurements were made using a combination pH/reference electrode standardized with pH buffers of 4.00 and pH 6.88. Chloride and fluoride anions concentrations (as degradation products) were analyzed by ion chromatography (IC: Dionex DX500 with CD20 conductivity detector, GP50 gradient pump, AS3500 autosampler). For the batch experiments, a Rayleigh-type distillation equation (eq 1) is used to describe the partitioning and fractionation of isotopes between reservoirs as one (reactant) reservoir decreases in size and may be used to interpret changes in δ13C (isotopic fractionation) with respect to the natural log of the fraction ( f) for the residual compound in the tests:

R ≈ R of (α− 1)

(1)

where R, Ro = isotopic ratio (13C/12C) at times t and t = 0 (initial) f = X1/X0, the residual fraction X1 = the concentration or amount of the more abundant (lighter) isotope 1765

dx.doi.org/10.1021/es203386a | Environ. Sci. Technol. 2012, 46, 1764−1773


Article

Figure 1. (a) CFC-113 and (b) CFC-11 degradation curves during batch test with ZVI material (● = control vials). First-order degradation plot for (c) CFC-113 (n = 8) and (d) CFC-11 (n = 7 samples). Eh-pH diagrams for (e) CFC-113 and (f) CFC-11 batch tests. Error bars are 2 SE. N.B. The first-order kinetics are determined using the solid symbols data only. CFC samples lying off the main trends (a−e) are highlighted as open symbols; i.e., for CFC-113 all samples at or beyond the limits of its analytical determination (>115 h); for CFC-11 all samples after 20 h which appear to exhibit some rate-limiting behavior possibly related to a shift to less reduced redox conditions (see f); see text for full explanation.

(grid location 54.63° N, −5.91° W) during the period March to May 2004. Full details are given in the Supporting Information. Two parts of the landfill site where no active gas extraction was taking place were sampled and monitored. Flux chamber bases were emplaced on both sites one week prior to the first sampling round.22 For sampling, a clear Perspex chamber (50 cm height × 40 cm diam.) was fitted into the base. At t = 0 a glass lid was placed on top to give a closed flux chamber. The 60-mL sample line was evacuated using a 100-mL glass syringe and a pre-evacuated 0.5-L electropolished stainless steel canister, then attached to the stainless steel line using a

X0 = initial (t = 0) concentration of the more abundant isotope α = isotopic fractionation factor which in standard isotopic δ-notation (δ(‰) = 1000·(R − Rstd)/Rstd, where Rstd is the isotopic ratio with respect to the reference standard VPDB) and defining the isotopic enrichment factor ε(‰) = 103(α − 1) yields (cf. Clark and Fritz27)

1000· ln[(δ + 1000)/[δ0 + 1000)] = ε·lnf

(2)

Landfill Field Study. Flux chamber studies were conducted on a landfill site on the Dargan Road, Belfast, Northern Ireland 1766



Article

the identified products are reported in SI (Batch Degradation Products) and Figure S6. Dissolved chloride concentrations increased in solution (from 3 to 35 mg/L) over the course of the batch test, however fluoride could not be identified under the circumstances (see Supporting Information for further discussion). For the batch tests, ZVI material indeed is capable of abiotically degrading both CFC-113 and CFC-11, as has been reported previously.8,12,13 The degradation of CFC-113 and CFC-11 under (measured) reducing conditions (Figure 1e, f) is most likely to be attributed to dehalogenation. CFC-113 appears to transform to both HCFC-123a and CFC-1113, the former product through reductive hydrogenolyis and dechlorination; both CFC-113 (by direct dichloroelimination) and HCFC-123a (by dehydrochlorination) abiotic degradation pathway to CFC-1113 have been reported previously for anaerobic media.24 Only trace amounts of CFC-21 (after 20-h batch test) and CFC-31 (after 40-h batch test) appear as transformation products of CFC-11. For the CFC-11 transformation pathway, Vidumsky and Thomson12 also report a second, competing pathway that resulted in the formation of an unstable dihalocarbene radical which rapidly underwent hydrolysis to yield HCl and HF. Although it was not possible to identify these acids in this study (as the column used in the GC-MS was not suitable) this might explain the different pH evolution for the CFC-11 batch test against that measured for CFC-113 (Figure 1e, f). The pH trend for CFC-113 degradation follows a characteristic evolution for unbuffered systems and reactions using chlorinated solvents and zero-valent iron,21,25 showing generally a gradual rise in pH from around neutral to a more basic pH (details are given in SI − Batch pH and Eh Evolution and Figures S3 and S4). By contrast, the pH curve of the batch test using CFC-11 and ZVI initially increased after 7.45 h but then decreased and remained “buffered” at around a neutral pH for the duration of the rest of the batch test. It is conceivable that hydrolytic reaction and production of HCl and HF after 7.45 h might account for this phenomenon. It is also conceivable that if CFC-11 degradation products transformed to acids the acid may leach iron hydroxide from the surface of the ZVI, which could therefore decrease both pH and increase Eh. Typically acids are used to pre-treat the surface of commercial ZVI material specifically to remove iron hydroxide, which is also true for laboratory batch or column test materials if the iron is being reused7, and so acid production could change the surface reactivity. Alternatively, the evidenced shift to less reducing conditions of the later batch samples (from ca. −500 mV to −300 mV after 20 h, Figure S5b) itself might shuggest a redox threshold operating, although it is not known whether this then simply reflects a problem with the batch sample integrity during experiment. The proposed reduction scheme for CFC-11 is therefore that this compound transforms to HCFC-21 and is then further reduced to HCFC-31 by single-electron reactions. One of three different pathways may be responsible for this transformation. (Supporting Information, Figure S2). CFC-113 degrades by a different reaction pathway. It potentially transforms to HCFC123a by a single-electron reaction pathway (similar to schemes outlined in Figure S2) but transforms to CFC-1113 by a double-electron (β-elimination) reaction pathway. Lesage et al.24 report that CFC-113 abiotically degraded to HCFC-123a and CFC-1113 using reduced hematin as the reductant.

Swagelok UltraTorr fitting (Swagelok Ireland, UK). Following closely the field protocol developed by Redeker:22 at t = 1 min the canister was opened slowly and equilibrated with the chamber to obtain a baseline sample; the canister was closed after 1 min and sampled; another sample was taken at 16 min in a 0.5-L canister, and finally at 31 min in a 1-L canister. Where positive fluxes of CFCs were recorded in the chamber the final sample represents the FLUX value; conversely where positive fluxes were not recorded (i.e., either no emissions or dynamic movement and consumption of gases into the landfill) the samples reflect NO FLUX values. Landfill ambient air (LAA) samples were also taken at the same location and time on the landfill in a 3-L pre-evacuated canister. Every day that flux sampling was conducted at the landfill site, an urban ambient air (UAA) sample also was taken at Queen’s University Belfast (∼6 miles away; grid location 54.58° N, −5.94° W17,18,23). For source comparison, a remote (rural) ambient air (RAA) sample from Mace Head (Republic of Ireland; 53.20° N, −09.54° W) also was collected by Mr. Gerry Spain (ALE-GAGE) under clean air conditions in a 15-L evacuated canister. A small number of samples also were collected from small township (∼3000−3500 population) rural locations at Crossgar (54.40° N, −5.76° W) and Hillsborough (in 54.46° N, −6.08° W) in Northern Ireland during August−November 2004. Lastly, on one of the sites the soil flux chambers were located in close proximity to a gas monitoring well. LFG samples (3-L stainless steel canister) therefore were collected from the gas vent of this monitoring well on three separate occasions and analyzed to characterize the δ13C values for CFCs within the landfill. For VOC analyses, a 100 μL subsample of headspace or source gas was injected into a GC-MS-IRMS (ThermoFinnigan Voyager MS coupled with a Micromass Isoprime IRMS) using a gas-tight syringe; full calibration details are given elsewhere.17,23 For the ambient air samples carbon stable isotope concentrations of CFC-11 and CFC-12 exceeded the detection limits of the IRMS and allowed for comparison of the δ13C signatures from the three locations. All stable carbon isotope values are subsequently reported as δ13C per mille (‰) relative to the Vienna PeeDee Belemnite (VPDB) reference standard.23

■

RESULTS AND DISCUSSION Batch Degradation Kinetics for CFC-113 and CFC-11. Over 90% of CFC-113 and CFC-11 was degraded within 88 and 116 h of the start of the batch tests, respectively (Figure 1). Control vials showed no significant decrease over the duration of the tests (ref 17 Appendix 6; Figure1a, b ). For CFC-113, degradation followed a simple, characteristic exponential decay (Figure 1a) with ∼95% of the mass concentration having degraded after 88 h. For CFC-11 (Figure 1b), rapid degradation apparently occurred within 7.45 h with concentrations decreasing an average 14 ppm (ca. 46% mass degradation from initial concentration; solid symbols, Figure 1b, d, f); however, after 20 h there appeared to be some ratelimiting behavior as concentrations decrease by a further 11 ppm only over the next 93 h (open symbols, Figure 1b, d, f). Control samples remained around pH = 6.2 ± 0.7 (2 SD; N = 12) and (oxidizing) Eh = +102 ± 129 mV (2 SD; N = 12) for CFC-113 batch tests, and pH = 6.4 ± 0.8 (2 SD; N = 8) and (oxidizing) Eh = +86 ± 90 mV (2 SD; N = 8) for CFC-11. The detailed pH and Eh evolution of the batch experiments is given in SI (Batch pH and Eh Evolution) and Figures S3 and S4, and 1767



Article

Figure 2. Plots of δ13C evolution for (a) CFC-113 and (b) CFC-11 (● = control vials). Rayleigh type plots for (c) CFC-113 (all data) and (d) CFC11 (first 20 h only). N.B. in Figure 1d (only) CFC-11 samples lying off the main linear trend are highlighted as open squares reflecting concentration samples apparently exhibiting rate-limiting behavior; see text for explanation. (e) Plot of δ13C evolution against time for CFC-113 degradation and its identified products (HCFC-123a, CFC-1113). Average fractionation factors (Δ13C) between the parent (CFC-113) and its daughter compounds (HCFC-123a, CFC-1113) are also indicated. Note that beyond ca. 100 h the isotopic trends for HCFc-123a and CFC-1113 flatten as the reaction appears to stop.

reductant.25 Both degradation curves (Figure 1a, b) have been transformed to a log-normal plot of (C/C0) versus time (Figure 1c, d) allowing characterization of (pseudo-) first-order rate constants (k) of −0.034 ± 0.002 h−1 (N = 7; R2 = 0.996)

Reduction of halogenated compounds in an iron/water system generally is considered first-order with respect to not only the concentration of the halogenated compound but also the concentration of iron surface which can serve as a 1768



Article

for CFC-113 and −0.04 ± 0.02 h−1 (N = 4; R2 = 0.891) for CFC-11, and giving half-lives of reaction (τ1/2 = ln2/k) of around 20.5 h (CFC-113) and 17.3 h (CFC-11). Dividing the kinetic rate constants to the surface area concentration of Fe0 (see Supporting Information Experimental Set-up: Batch Studies) then gives normalized (pseudo-) first-order rate constants (kSA) of −(9.79 ± 0.46) × 10−5 L m−2 h−1 for CFC-113 and −(1.16 ± 0.46) × 10−4 L m−2 h−1 for CFC-11, giving normalized half-lives of reaction of around 7100 h m2 L−1 (CFC-113) and 5900 h m2 L−1 (CFC-11). The normalized rate constant for CFC-11 identified here is 2 orders of magnitude smaller than reported originally by Scheutz et al.8 for a watersaturated (anaerobic) batch test using an experimental setup similar to that described here but using coarse iron fillings supplied by BDH Laboratories, UK (in contrast to ConnellyGPM iron in the current study). Much of the difference however likely can be accounted for by the different ZVI surface areas (a factor ∼30) and the solids preparation/ reactivity. Such normalized rate constants identified herein provide a starting point for design calculations for treatability of CFC-113 and CFC-11 using Connelly-GPM (ETI CC-1004) ZVI material. However, in the current study, for CFC-11 there appears also to be some rate-limiting step in the reaction which potentially may be related to the production of acids passivating the reactive sites on the ZVI surface but seems to be more likely related to a redox threshold for activation. Scheutz et al.8 did not attempt to identify the degradation products for CFC-11; and given the (unspecified) greater ZVI mass packing achievable, the apparent half-life of reaction in flowthrough column tests using Connelly-GPM ZVI (20 ppm influent CFC113; 100 ppm influent CFC-11) are (as to be expected given the enhanced suface area to liquid ratios) significantly shorter than the half-lives of reaction in batch test, but nevertheless the resulting degradation products identified by Vidumsky13 are herein now confirmed also in batch testing. Isotope Values during ZVI Batch Degradation of CFC113 and CFC-11. The change in δ13C of CFC-113 during the batch test experiment is shown in Figure 2a. The initial δ13C value at t = 0 was −30.4 ± 0.5‰, within the standard deviation of the control values (−29.7 ± 1.3‰). The isotope values for CFC-113 then became more enriched in 13C resulting in a δ13C value of −16.7 ± 2.2‰ after 88 h and 95% degradation. The high standard deviation for this latter value is due to the concentrations of CFC-113 in the sample being on the cusp of detection limits of the IRMS estimated as 150 ng of carbon;26 it was not then possible to record the δ13C(CFC-113) for the remainder of the experiment (after 112 h) as the CFC-113 concentration was below detection. The more or less constant isotopic values for HCFC-123a and CFC-1113 after this period (Figure 2e) attest to the cessation of the parent compound reaction. A plot of δ13C against time shows a linear relationship, implying that initial degradation has followed a first-order reaction and should therefore fit a Rayleigh model (see below). The change in δ13C of CFC-11 during the batch test experiment showed an initial δ13C value (t = 0) of −39.8 ± 1.0‰ similar to the control values 40.4 ± 0.9‰ Figure 2b), which again then became more enriched in 13C over the course of degradation and yielding a δ13C value of −30.2 ± 1.1‰ after 20 h and 46% of the mass degradation. However, beyond 20 h of reaction the δ13C values of the residual CFC-11 appear to plateau and remain around −31.8 ± 1.3‰ for the remainder of the batch test. This is perhaps related to a shift in the redox condition of the later batch samples (from ca. −500 mV to

−300 mV after 20 h; Figure S5b), although it is not known whether this reflects a problem with the batch sample integrity during experiment or reflects changing reactivity of the surface. Nevertheless, the initial degradation effects on isotopic values over the first 20 h (N = 4 samples) again followed a linear relationship. For CFC-113, δ13C values for its products HCFC-123a and CFC-1113 were also monitored alongside the parent compound (Figure 2e). Although both degradation products were detected by GC-MS after 4.75 h, the δ13C value for these compounds could not be detected with any precision by GCIRMS until 20 h. All degradation products followed a similar trend, becoming enriched in 13C over time. At 116 h the δ13C values fell below the detection limit of the IRMS system for CFC-113. As CFC-113 degrades to both HCFC-123a and CFC-1113, the latter products remain isotopically shifted (Δ13C fractionation) some −15‰ (HCFC-123a) and −25‰ (CFC-1113) lighter than the parent CFC-113 compound. The δ13C values of the degradation products, HCFC-123a and CFC-1113, subsequently continued to become more enriched in 13C until approximately 200 h where their δ13C values appear to plateau. The left-hand side values identified in eq 2 are plotted against the natural logarithm fraction of CFC remaining (Figure 2c, d). The slope of this line through the origin then defines the (kinetic) isotopic enrichment factor (ε(‰) = (α − 1)·1000). Kinetic isotope enrichment factors of ε = −5.0 ± 0.3‰ (CFC113; based on 6 samples) and −18 ± 5‰ (CFC-11; based on 4 samples over first 20 h only) are derived. Because bonds containing 13C energetically are much more stable than 12C, during reductive dehalogenation the 12C-halogen bonds are more susceptible to both single- and double- electron transfer steps than the 13C-halogen bond which infers a Rayleigh type model for fractionation.28 The measured isotopic enrichments for CFC-113 and CFC-11 are consistent then with a reaction scheme that degrades 12C-containing molecules at a faster rate than 13C-containing molecules, and both ε values are consistent with a primary kinetic isotope effect.27 However, the larger isotopic effect associated with CFC-11 indicates less limitation on the isotopic degradation of the molecule than for CFC-113 (the latter transforming through multiple steps in the reaction scheme to CFC-1113 and HCFC-123a). Armed with the knowledge of ε, the initial δ13C value of source CFCs, and following δ13C values at any other point of remediation then allows the fraction of the CFC remaining to be calculated.29 Measured fractionation factors (ε) also may be used to compare the kinetic isotope effects for CFC-11 or CFC-113 using different grades and types of ZVI; Slater et al.7 have used ε to compare the enrichment factors in published batch and column tests for TCE. Finally, an isotopic mass balance approach for CFC-113 was attempted using the peak area from the MS chromatograms and the corresponding δ13C values (Figure 2e) to determine whether the carbon isotopes of CFC-113 and transformation products balanced; the sum of CFC-113, HCFC-123a, and CFC-1113 contributions adds up to 95% implying that the major degradation compounds have been identified more or less quantitatively. Further details are given in SI − Batch Degradation Products. Landfill Field Study. Where positive fluxes were identified in the chambers, calculations suggest mass flux estimates through the soil cover typically ∼10−5 g/m2/day for CFC-11, CFC-12, and CFC-113.17 Further details are given in 1769



Article

Table 1. Averaged Isotopic Value Results (δ13C, ‰) for All CFC/HCFC/HFC Compounds Reported in This Paper, Sampled from Multiple Comparative Sources, and Single Standard Deviations (SD) Based on Temporal (and, in the Case of Landfill Samples, Spatial) Variability over the Entire Study Perioda δ13C (‰) avg

±SD

8 9 4 1 1 8 4

−42.7 −38.7 −14.6 −1.2 −28.2 −28.5 −12.6

3.9 2.6 4.9

31 41 13

−44.3 −35.9 −18.9 not detected −22.1 −28.9 −28.1

count N RAA HCFC-22, CHClF2 CFC-12,CCl2F2 HFC-134a, C2H2F4 HCFC-142b,C2H3ClF2 CFC-114, C2Cl2F4 CFC-11,CCl3F CFC-113,CCl2FCClF2 UAA HCFC-22, CHClF2 CFC-12,CCl2F2 HFC-134a, C2H2F4 HCFC-142b,C2H3ClF2 CFC-114,C2Cl2F4 CFC-11,CCl3F CFC-113,CCl2FCClF2 LAA HCFC-22, CHClF2 CFC-12,CCl2F2 HFC-134a, C2H2F4 HCFC-142b,C2H3CIF2 CFC-114,C2Cl2F4 CFC-11,CCl3F CFC-113,CCl2FClF2 published solvent sources (δ13 C, HCFC-22, CHClF2 CFC-12, CCl2F2 CFC-11, CCl3F CFC-113,CCl2FCClF2

2 34 15 8 9 1

9 ‰)

−37.6 −33.0 5.5 not detected not detected −20.1 not detected Ertl31

2.9 6.8 9.2 5.5 9.8 9.7 6.1 4.6 7.6 2.8

4.6

count N

δ13C (‰) avg

±SD

2 10

−22.3 −32.2 not detected not detected not detected -24.9 −29.5

5.6 5.4

NO FLUX HCFC-22, CHC1F2 CFC-12,CCl2F2 HFC-134a, C2H2F4 HCFC-142b, C2H3C1F2 CFC-114, C2Cl2F4 CFC-11,CCl3F CFC-113, CCl2FCCIF2 FLUX HCFC-22, CHC1F2 CFC-12,CCl2F2 HFC-134a, C2H2F4 HCFC-142b,C2H3ClF2 CFC-114, C2C12F4 CFC-11, CCl3F CFC-113, CCl2FCClF2 LFG HCFC-22, CHC1F2 CFC-12,CCl2F2 HFC-134a, C2H2F4 HCFC-142b,C2H3ClF2 CFC-114, C2C12F4 CFC-11, CCl3F CFC-113, CCl2FCClF2 Thompson et al.19 (2× std error)

−22.3 −8.0 3.3 1.5 not detected not detected 3 −10.7 4.3 1 −27.8 Archbold et al.23 (2× std error N = 5)

−31.3 ± 0.5‰

−46.8 ± 0.2‰ −26.2 ± 0.6‰ −26.5 ± 0.8‰

−48 to −45‰ −45 to −33‰ −35 to −25‰

4 1

5

1 3 1

not detected −24.3 not detected not detected −26.7 -28.6 −20.8

2.2

1.7

6.5

1 3 1

a

Values in italics are from analyses where the corresponding sample concentrations are close to or at baseline detection making and may not be representative. Cited solvent source ranges and isotopic signatures are also presented.

reported value here of δ13C(‰) = −12.6 ± 6.8 for CFC-113 based on only 4 samples, however, is in fact isotopically heavier than that reported by ref 18 of δ13C(‰) = −28.1 ± 4.4 (N = 15) which is more in line with the reported given RAA value (Table 1). Nevertheless, overall the characteristic isotopic values for the UAA samples are therefore not significantly different from the RAA samples, and generally are within the characteristic isotopic ranges cited for solvent source terms for HCFC-22, CFC-12, CFC-11, CFC-113 given above but apparently therefore do not show any local impact at least at QUB of the Dargan Road landfill gas emissions. LFG samples taken from the gas vent (Old Site) show significantly heavier isotopic signatures than both the UAA and RAA samples for all gases excepting CFC-113 (Tables 1 and S3). With the exception of the singular CFC-113 sample, these LFG values are also significantly isotopically heavier than potential solvent sources. An isotopic difference (Δ13C) around −15‰ between a possible solvent source signature and the sampled LFG samples for CFC-11 (N = 3 samples) apparently is of a similar order of magnitude to the kinetic isotope effect shown for the abiotic degradation by ZVI material. Significant isotopic differences (Δ13C) of the order −13‰ (N = 1) and −35‰ (N = 3) are evidenced also for HCFC-22 and CFC-12 in the landfill gas (LFG) samples, respectively. The impact of LFG emissions at the landfill site is picked up then in the isotopic values of the monitored LAA and FLUX

Supporting Information - Emissions of CFCs through Landfill Soil Cover on the Old Landfill (Figure S7; Table S1). Characteristic average isotopic values measured for all of the different sampled sources (UAA, RAA, LAA, FLUX, NO FLUX, LFG) are reported in Table 1. Comparisons of differences between the mean values for the different sampled sources subsequently have been evaluated using the standard two-tail Student’s t-test30 for independent groups (SI, Table S3). In terms of potential solvent source characterization (Table 1), Ertl31 originally carried out an extensive study on the δ13C values of solvents sourced from a number of CFC manufacturers and has reported δ13C ranges for CFC-11, CFC-12, and HCFC-22. More recently, Thompson et al.19 has reported a value for a single commercially-produced CFC-113; Archbold et al.23 provide replicate δ13C analysis of individual industrially-produced standard gas samples having for CFC113, CFC-11, and CFC-12 (the latter value only slightly more depleted than the isotopically lightest end of the range given by ref 31; Table 1). The average isotopic value reported for the UAA samples (Table 1) taken close to Queen’s University Belfast in the South of the city and some six miles from the Dargan Road landfill site proves not to be significantly different (Student’s test, 95% confidence limit; Table S3) from that of the RAA samples across all identified gases, except CFC-113. The 1770



Article

Figure 3. Summary of δ13C values characterized in this study for CFC-12 in the landfill environment. Error bar estimates are one standard deviation, except Archbold et al.23 which is two times standard error (= 2× standard deviation/√N). Note how, e.g., measured soil flux chamber (FLUX) values lie between the signatures characterized for landfill ambient air (LAA) and landfill gas (LFG); similarly, the LAA samples themselves likely reflect end-member mixing of Belfast urban ambient air (UAA) and the characteristic FLUX emissions signature.

experiments which generally prove intermediate in values between the LFG samples and UAA/RAA samples. Significant isotopic differences (Tables 1 and S3) are particularly shown for CFC-12 and CFC-11 for the LAA and UAA samples, with the former samples isotopically heavier. The FLUX experiments data generally confirm significantly isotopically heavier values for CFC-12 again than both the landfill ambient air (LAA) samples and the NO FLUX the latter as to be expected proving similar in isotopic values to the LAA samples (i.e., under NO FLUX conditions the movement of gases will be into the landfill, and we should expect greater local air signatures in sampling chambers). Overall, the characterized relationships between sample sources particularly for CFC-12 are shown schematically in Figure 3. A similar set of source relationships can be shown also for CFC-11; although here the landfill FLUX values shown appear closer to urban/remote atmospheric signatures these are likely not representative and instead the Landfill Ambient Air

(LAA) samples again are characterized by isotopically heavier values in line with a significantly heavier landfill gas (LFG) source (Table 1). It is likely then that the LAA samples reflect a mix of the FLUX signatures emanating from landfill cover and the local (Belfast) UAA signatures. A simple two-component mix of these isotopic signatures (assuming simply conservative mixing of characteristic end-members having similar concentrations) for CFC-12 would suggest that these LAA samples represent a mix ∼25% of the landfill emanations. Implications for Environmental and Engineering Systems Studies. During reductive dehalogenation reaction in water-saturated and anaerobic batch study with ZVI, relatively large changes in δ13C values of CFC-11 and CFC113 have been observed, characterized by significant kinetic isotope effects. These isotopic fractionations potentially have significant implications for monitoring the progress of ZVI remediation of these compounds. If the applied Rayleigh model 1771



Article

environmental and engineering systems studies. For CFC-11, the apparent similarity in magnitudes between the KIE for batch ZVI degradation (based however on only 4 degradation samples) and that evidenced between an assumed solvent source signatures and LFG in the landfill study does raise some question over the potential discriminative capacity for abiotic versus biotic processes if confirmed. However, even in that case further investigation could show the other CFC compounds giving better discrimination; moreover the magnitudes of fractionation themselves can be utilized simply to highlight degradative capacities for chlorofluorocarbon compounds. Potentially, CFC carbon isotope signatures also can be used in other environmental and also groundwater studies as a key indicator of environmental impacts,5 degradative concentration effects under anoxic conditions, or simply even to verify the conservative (or otherwise) nature of CFC compositional signatures used, e.g., in groundwater dating.34 Nevertheless, further investigation is now required both to resolve the discriminating power of isotopic fractionation magnitudes and effects evidenced, and also whether any possible temporal variabilities in the carbon isotope systematics for the CFCs (not looked at in the current work) also can be characterized and predicted/modeled.

accurately represents the pattern of isotopic fractionation during reductive dehalogenation, then knowledge of compound-specific carbon isotopic composition of the residual CFC contaminant can be used to quantify the degree of degradation of the compound in engineered ZVI remediation systems such as PRBs. Compared to ambiguities generally experienced in non-isotopic approaches, such as using concentrations of contaminant alone to estimate remediation efficiency, estimates based on stable carbon isotope measurements can help to determine whether a contaminant is being degraded or whether dilution and dispersion is affecting the contaminant.28 Potentially carbon-isotope fractionations of CFCs might also occur for unsaturated systems (untested in the current study), allowing isotopic monitoring of proposed abiotic remediation strategies for landfill gas emissions by placing ZVI material in top cover landfill material.8 Nevertheless, it is possible that the use of different reactive ZVI may result in varying values for the enrichment factor, ε. Moreover, dynamic column tests values need studying and evaluating for application to in situ monitoring or remediation technologies with PRBs. Because the current study has focused on abiotic reductive dehalogenation reaction mechanisms the isotopic effects for specific microbial and other chemical mechanisms remain to be evaluated in detail before compound-specific isotope analysis of CFCs necessarily can be applied to realworld contaminant and remediation problems. Although CFC emissions from landfills in the United States and UK have been assessed recently32 as posing only a minor source of ozone-depleting substances (ODSs) in the countries, the “potential remains for increased banking of ODSs in both MSW and no-MSW landfills in industrialized countries where landfilling is commonly practiced, such as the US or UK.” The greater significance of CFC-12 emissions in the UK is noted such that identifying degradation and attenuation mechanisms nevertheless are still of interest, and their effects in situ are only starting to be investigated (cf. ref 16). The landfill characterization exercise presented here provides further evidence of the potential of carbon isotope analyses for use in landfill (and other engineering and environmental) degradation studies involving biotic mechanisms. Potentially, the stable carbon isotopes of CFCs might be utilized alongside those of methane33 to help constrain regional atmospheric emission inventories if it can be shown that LFG temporal variabilities in stable isotopic signatures in fact are limited. The ZVI batch data itself directly may also be used to evaluate abiotic remediation strategies for (saturated) flowthrough PRBs, and possibly in landfill strategies such as ZVI as a top cover landfill material to attenuate CFCs under unsaturated but anaerobic (e.g., LFG atmosphere) conditions via the reductive dehalogenation mechanism8 if isotopic fractionation is shown to occur also. It appears that potentially natural attenuation processes indeed may be occurring or have occurred in the Dargan Road landfill, as CFCs sampled from gas monitoring wells significantly enriched in δ13C (isotopic differences (Δ13C) of the order −13‰ (HCFC-22), −35‰ (CFC-12) and −15‰ (CFC-11)) are suggested compared to solvent source valueslikely due to microbial transformation in the anaerobic LFG environment. Although the CFC-11 isotope fractionation found here is similar in magnitude to the kinetic isotope effect (KIE) evidenced for abiotic ZVI (chemical) degradation of dissolved CFC-11, CFC carbon isotope fractionations evidenced in the current work clearly suggest a potential for utilizing δ13C signatures in appropriate

■

ASSOCIATED CONTENT

S Supporting Information *

Details of laboratory batch study and field (landfill) site with background information on the potential degradation products of CFC-113 and CFC-11 using ZVI; CFC sources in atmospheric and landfill gases; methodology for the ZVI batch studies and sampling at the landfill site for the current study; results for the batch pH and Eh evolution and batch degradation products, for the emissions of CFCs through landfill soil cover on the old landfill site; results of applying the (Student’s) t-test between the characteristic average isotopic values of sample sources; and 7 figures, 2 plates, and 3 tables. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Present Address †

University of Strathclyde, David Livingstone Center for Sustainability, Level 6, Graham Hills Building, 50 Richmond Street, Glasgow, G1 1XN, Scotland, UK.

■

ACKNOWLEDGMENTS Support from UK EPSRC Research Grants GR/L85183/01 (R.K.) and GR/R03099/01 (R.K., T.E.) is gratefully acknowledged. M.A. acknowledges a European Science Fellowship (ESF) award and a Queen’s University Belfast (QUB) William and Betty MacQuitty Scholarship. At EERC, IC analyses for batch tests were conducted by Arlene Jamieson; Karen McGeough gave valuable advice on and help with the ZVI experiments including the Connelly ZVI characterization; Michael McCartney gave assistance with fieldwork; Dr. Kelly Redeker provided advice and guidance on flux sampling. We are also grateful to Mr. G. Spain (Mace Head Research Station) for organizing remote ambient air samples, and Belfast City Council (Messrs. E. Mulligan, D. O’Hagan, and P. Douglas) 1772



Article

and marine locations. J. Geophys. Res. 2007, 112 (D16307), 23 DOI: 10.1029/2006JD007784. (19) Thompson, A. E.; Anderson, R. S.; Rudolph, J.; Huang, L. Stable isotope signatures of background tropospheric chloromethane and CFC-113. Biogeochemistry 2002, 60, 191−211. (20) Mead, M. I.; Khan, M. A. H.; Bull, I. D.; White, G.; Shalcross, D. E. Stable carbon isotope analysis of selected halocarbons at parts per trillion concentration in an urban location. Environ. Chem. 2008, 5, 340−346. (21) Gillham, R. W.; O'Hannesin, S. F. Enhanced degradation of halogenated alipathics by zero-valent iron. Ground Water 1994, 32, 958−967. (22) Redeker, K. R. Methyl halide emissions from rice paddies. Unpublished PhD Thesis. University of California, Irvine, CA, 2002. (23) Archbold, M.; Redeker, K. R.; Davis, S.; Elliot, T.; Kalin, R. M. A routine method for compound specific stable isotope measurements of methyl halides and chlorofluorocarbons at pptv concentrations. Rapid Commun. Mass Spectrom. 2005, 19 (3), 337−342. (24) Lesage, S.; Brown, S.; Holster, K. Degradation of CFC-113 under anaerobic conditions. Chemosphere 1992, 24, 1225−1243. (25) Johnson, T. L.; Scherer, M. M.; Tratnyek, P. G. Kinetics of halogenated organic compound degradation by iron metal. Environ. Sci. Technol. 1996, 30, 2634−2640. (26) Hall, J. A. Stable isotope methods for monitoring bioremediation of contaminated land. Unpublished PhD Thesis, Queen’s University of Belfast, Northern Ireland, 1999. (27) Clark, I.; Fritz, P. Environmental Isotopes in Hydrogeology; CRC Press LLC: New York, 1997. (28) Daylan, H.; Abrajano, T.; Sturchio, N. C.; Winsor, L. Carbon isotopic fractionation during reductive dehalogenation of chlorinated ethenes by metallic iron. Org. Geochem. 1999, 30, 755−763. (29) Hearty, L. J.; Fuller, M. E.; Huang, L.; Abrajano, T.; Sturchio, N. C. Isotopic fractionation of carbon and chlorine by microbial degradation of dichloromethane. Org. Chem 1999, 30, 793−799. (30) Miller, J. C.; Miller, J. N. Statistics for Analytical Chemistry, 3rd ed.; Ellis Horwood, 1993; 233 pp. (31) Ertl, S. J. Herkunftsbestimmung organischer schadstoffe durch untersuchung des naturlichen isotopengehalts. Unpublished PhD thesis, Technischen Universität München, 1997. (32) Hodson, E. L.; Martin, D.; Prinn, R. G. The municipal solid waste landfill as a source of ozone-depleting substances in the United States and United Kingdom. Atmos. Chem. Phys. 2010, 10, 1899−1910. (33) Lowry, D.; Holmes, C. W.; Rata, N. D.; O’Brien, P.; Nisbet, E. London methane emissions: Use of diurnal changes in concentration and delta C-13 to identify urban sources and verify inventories. J. Geophys. Res.-Atmos 2001, 106 (D7), 7427−7448. (34) Cook, P. G.; Plummer, L. N.; Solomon, D. K.; Busenberg, E.; Han L. F. Effects and processes that can modify apparent CFC age. In Plummer, L.N.; Busenberg, E.; Bö hlke, J. K., et al. Use of Chlorofluorocarbons in Hydrology: A Guidebook; STI/PUB/1238, 277; ISBN 92-0-100805-8; International Atomic Energy Agency (IAEA), 2006; pp 31−58.

for their co-operation. Three anonymous reviewers are particularly thanked for improvement of this manuscript.

■

REFERENCES

(1) Höhener, P.; Werner, D.; Balsiger, C.; Pasteris, G. Worldwide occurrence and fate of chlorofluorocarbons in groundwater. Crit. Rev. Environ. Sci. Technol. 2003, 33 (1), 1−29. (2) Allen, M. R.; Braithwaite, A.; Hills, C. C. Trace organic compounds in Landfill Gas at Seven UK Waste Disposal Sites. Environ. Sci. Technol. 1997, 31, 1054−1061. (3) Deipser, A.; Poller, T.; Stegmann, R. Emissions of Volatile halogenated hydrocarbons from landfills. In Landfilling of Waste: Biogas; E & FN Spon: London, UK, 1996; pp59−72. (4) Kjeldsen, P.; Scheutz, C. Short- and long-term releases of fluorocarbons from disposal of polyurethane foam waste. Environ. Sci. Technol. 2003, 37 (21), 5071−5079. (5) Ho, D. T.; Schlosser, P.; Smethie, W. M.; Simpson, H. J. Variability in atmospheric chlorofluorocarbons (CCl3F and CCl2F2) near a large urban area: Implications for groundwater dating. Environ. Sci. Technol. 1998, 32, 2377−2382. (6) Hoffstetter, T. B.; Schwarzenbach, R. P.; Bernasconi, S. M. Assessing transformation processes of organic compounds using stable isotope fractionation. Environ. Sci. Technol. 2008, 42 (21), 7737−7743. (7) Slater, G. F.; Sherwood Lollar, B.; Allen King, R.; O'Hannesin, S. Isotopic fractionation during reductive dechlorination of trichloroethene by zero-valent iron: Influence of surface treatment. Chemosphere 2002, 49, 587−596. (8) Scheutz, C.; Winther, K.; Kjeldsen, P. Removal of halogenated organic compounds in landfill gas by top covers containing zero-valent iron. Environ. Sci. Technol. 2000, 34, 2557−2563. (9) Steele, J. R.; Griffith, R. W.; Randall, P. C. In Situ enhanced bioremediation of freon11/freon113 groundwater contamination using hydrogen release compound. In Proceedings of the Seventh International In Situ and On-Site Bioremediation Symposium, Orlando, Florida, June 2−5, 2003; Battelle Press: Columbus, OH, 2004, 7 pp. (10) Schüth, C.; Bill, M.; Barth, J. A. C.; Slater, G. F.; Kalin, R. M. Carbon isotope fractionation during reductive dechlorination of TCE in batch experiments with iron samples from reactive barriers. J. Contam. Hydrol. 2003, 66, 25−37. (11) Phillips, D. H.; Van Nooten, T.; Bastiaens, L.; Russell, M. I.; Dickson, K.; Plant, S.; Ahad, J.; Newton, T.; Elliot, T.; Kalin, R. M. Ten Year Performance Evaluation of a Field-Scale Fe0 Permeable Reactive Barrier Installed to Remediate TCE-Contaminated Groundwater. Environ. Sci. Technol. 2010, 44 (10), 3861−3869. (12) Vidumsky, J. E.; Thomson, M. Treatability of CFC-11 and CFC113 with Zero Valent Iron (Abstract). In Proceedings of Remediation of Chlorinated and Recalcitrant Compounds; The Fourth International Conference May 24−27, Monterey, CA, 2004. (13) Vidumsky, J. E. Treatability of CFC-11 and CFC-113 with ZVI. In Summary of the Remediation Technologies Development Forum Permeable Reactive Barriers Action Team Meeting, October 26−27, 2004, Best Western Winrock Inn, Albuquerque, NM, USA, 2004. http://www.rtdf.org/public/permbarr/minutes/102704/index.htm (accessed 13/09/2011). (14) Deipser, A.; Stegmann, R. Biological degradation of VCCs and CFCs under simulated anaerobic landfill conditions in laboratory test digesters. Environ. Sci. Pollut. Res. 1997, 4 (4), 209−216. (15) Scheutz, C.; Dote, Y.; Fredenslund, A. M.; Mosbæk, H.; Kjeldsen, P. Attenuation of fluorocarbons released from foam insulation in landfills. Environ. Sci. Technol. 2007, 41 (22), 7714−7722. (16) Scheutz, C.; Bogner, J. P.; Blake, D.; Morcet, M.; Aran, C.; Kjeldsen, P. Atmospheric emissions and attenuation of non-methane organic compounds in cover soils at a French landfill. Waste Manage. 2008, 28, 1892−1908. (17) Archbold, M. E. Carbon Isotopes of Volatile Organic Compounds for Environmental Tracing. Unpublished PhD Thesis, School of Civil Engineering, Queen’s University Belfast, UK, 2004. (18) Redeker., K. R.; Davis, S.; Kalin, R. M. Isotope values of atmospheric halocarbons and hydrocarbons from Irish urban, rural, 1773


Carbon Isotopic Fractionation of CFCs during ... - ACS Publications

Recommend Documents