Reductive Dehalogenation of Hexachloroethane, Carbon

examined at 50 °C in aqueous solutionscontaining ei- ..... 0 Conditions: 250 µ Fe2+, 250 µ HS™, 25 mg of C/L of humic acid, 0.1 mM hematin, and 5...
0 downloads 0 Views 1MB Size
Environ. Sci. Technol. 1994,28,2393-2401

Reductive Dehalogenation of Hexachloroethane, Carbon Tetrachloride, and Bromoform by Anthrahydroquinone Disulfonate and Humic Acid Gary P. Curtis’ and Martin Reinhard

Department of Civil Engineering, Stanford University, Stanford, California 94306 The reductive dehalogenation of hexachloroethane (CzCLj), carbon tetrachloride (CC14),and bromoform (CHBr3) was examined at 50 “C in aqueous solutions containing either (1) 500 pM of 2,6-anthrahydroquinone disulfonate (AHQDS), (2) 250 pM Fe2+,or (3) 250 pM HS-. The pH ranged from 4.5 to 11.5 for AHQDS solutions and was 7.2 in the Fez+ solutions and 7.8 in the HS- solutions. The observed disappearance of C&16 in the presence of AHQDS was pseudo-first-order and fit k’ccb = ko[A(OH)zI + kdA(0H)O-I + kz[A(O)~~-l whereA(OH)2,A(0H)O-, and A(0)n2- represent the concentrations of the three forms of the AHQDS in solution. The values of ko,kl, and kz were -0,0.031, and 0.24 M-l s-l, respectively. The addition of 25 mg of C/L of humic acid or organic matter extracted from Borden aquifer solids to aqueous solutions containing HS- or Fe2+increased the reduction rate by factors 250 I~.M of up to 10. The logarithms of the rate constants for the disappearance of CzCl6 and C C 4 in seven different experimental systems were significantly correlated; log - 0.83 with r2 = 0.80. The observed k’cc14 = 0.64 log k’czc~ trend in reaction rates of C&l6 > CCl4 > CHBr3 is consistent with a decreasing trend in one-electron reduction potentials.

Introduction Polyhalogenated hydrocarbons are susceptible to reduction reactions in environmentssuch as soils, sediments, and aquifers (1,2).The electron donors in these reactions and the factors that affect the reaction rates are not well understood. Although iron porphyrins and other transition metal complexes reduce halogenated hydrocarbons in well-defined laboratory studies (3-5),it is not known if these reactants reduce halogenated hydrocarbons in sediment-water systems. In sediment-water systems, the rate of halogenated hydrocarbon reduction increases with increasing natural organic matter (NOM) content of the sediment (6, 7). However, hexachloroethane (CzCl6) is also reduced in mineralogic fractions containing silicates bearing ferrous iron (6)and tetrachloromethane (CCld) is reduced either in the presence of sheet silicates when HSis present (8)or in the presence of pyrite (9). Therefore, the role of organic matter in the reduction of halogenated hydrocarbons remains unclear and is considered in greater detail in this work. Studies of nitroaromatic reduction by reduced iron porphyrins and hydroquinones suggest possible reaction pathways for the reduction of halogenated hydrocarbons. It has been hypothesized that iron porphyrins and hydroquinones are electron transfer mediators that form highly reactive, reduced intermediates that reduce the oxidized contaminant as shown in Scheme 1(10,141. Once

Scheme 1

bulk donor ox

e- mediatorox

1\ [/

-ne-

subsusreox +ne-

the electron transfer mediator is oxidized by the contaminant, it is assumed to be reduced back to the initial state by a “bulk”electron donor, which may be present in excess of the mediator. If the concentration of reduced mediator is large relative to the concentration of the oxidized substrate, the extent of cycling of the mediator may be small. Some hydroquinones have been found to be efficient electron transfer mediators with either Fez+ or HS- as the bulk electron donor (10). However, hydroquinones have not been shown to mediate reduction of halogenated hydrocarbons. NOM has recently been shown to increase the rate of reduction of nitroaromatic compounds in the presence of HS-, and it was proposed that the hydroquinone/quinonetype couple was the electron transfer mediator (11). This is consistent with the known properties of model hydroquinones, as discussed above, and the redox properties of NOM. This hydroquinone/quinone-typecouple is believed to dominate the redox properties of humic acid (13). Spectroscopic and stoichiometric evidence indicate that Fez+ quantitatively and selectively reduces quinones in humic acid to hydroquinones (14-16). Similarly, HSreduces quinone groups to hydroquinones in simple model compounds (17), and it has been suggested that HS- also reduces quinone groups to the corresponding hydroquinone group in humic acid (18). In this work, it was hypothesized that both simple hydroquinones and hydroquinone groups in NOM can mediate the reduction of halogenated hydrocarbons. Specifically, the objectives were to (1)test the hypothesis that CzCl6 could be reduced by a simple hydroquinone, 2,6-anthrahydroquinone disulfonate (AHQDS), which is known to reduce nitroaromatic compounds to amines (19), (2) determine if NOM can mediate the reduction of CZCl6 with Fez+ or HS- as the bulk electron donor, and (3) correlate relative rates of the reduction of three halogenated hydrocarbons in different experimental systems.

Reaction Pathways and Theoretical Considerations

* Address correspondence to this author at his present address: Water Resources Division, Mail Stop 421, U.S. Geological Survey, 345 Middlefield Road, Menlo Park, CA 94025.

Reduction of C&16 and CClI. The overall reduction of c2c16 to Czc14 requires the transfer of two electrons and may involve the formation of one or more intermediates (Figure 1). Transfer of a single electron (k1) leads to the dissociation of a C-C1 bond and the formation of the pentachloroethyl radical (I). Formation of I is presumably the rate-limitingstep since the transfer of the first electron is typically rate limiting (20). I can react further by several

0 1994 American Chemical Soclety

Environ. Sci. Technol., Vol. 28, No. 13, 1994 2393

0013-936X/94/0928-2393$04.50/0

Coupling Products R,X

+ A(0H)O-

R,X

+ A(0H)O-

-

k'R1X

k'RZX

+ A(0H)O'

(la)

R2X'- + A(0H)O'

(lb)

R,X'-

where the alkyl radical anion dissociates to form X- and R', which continues to react to form the final products as indicated in Figure 1. A(0H)O' is ultimately reduced by the bulk electron donor to A(0H)O- if it is a mediator. The second-order rate constant for reaction l a is often related to the reduction potential of the electron acceptor E ~ Rand ~ xthe reduction potential of the electron donor EIDby (31)

Figure 1. Potential pathways for the reductive dehalogenation of hexachloroethane to tetrachloroethene illustrating the intermediate formation of pentachloroethane (adapted from ref 25).

different pathways depending on the solution conditions. I can abstract a hydrogen from a C-H, if present, to form C2HC15, which would subsequently undergo dehydrohalogenation to form CpC14 (21-25). The formation of C&14 from C2HC15 proceeds primarily by elimination of HC1 (reaction 8) (25). In the presence of astrong electron donor, I could accept another electron to form the pentachloroethyl carbanion (II), which could proceed to form CZC14 by the loss of C1- (reaction 9) or be protonated to form CzHClb. I could also be further reduced directly to form CzCl4 by the loss of C1- (reaction 7). An analogousreaction was observed for the bromoethyl radical formed by the reduction of 1,2-dibromoethane in aqueous dimethylformamide (26). Figure 1 also illustrates that I can form known or unknown addition and coupling products. The formation of I1 via reaction 2 as the first step appears unlikely because it involves the unlikely simultaneous transfer of two electrons (20). The reduction of C c 4 is similar to that of C2C16in that the first step of the reaction involves a dissociative electron transfer and the formation of an intermediate radical (20). Possible pathways have been presented previously (8,9, 27). The trichloromethyl radical formed in the reduction of C C 4 can be further reduced to form dichlorocarbene, which can hydrolyze to form CO or formate. The radical can also abstract a hydrogen to form CHCb or dimerize to form C2C16. Recent results demonstrate that, in the presence of HS- and either biotite or vermiculite, ccb is converted to C02 via CS2 by HS- with CHCl:, as a minor product and CO as a trace product (8). Evaluation of Relative Rates of Reduction. The rates of reductive dehalogenation reactions in chemically related systems can, for outer-sphere electron-transfer reactions, be correlated by means of the Marcus relation (20,28,29). More recently, the Marcus relation has been used as a basis for developing semiempirical relations between the observed reduction rate of an organic contaminant and the one-electronreduction potential (10, 30,31). For example, consider the reduction of two alkyl halides (R1X and R2X) by an electron donor which is postulated to be a deprotonated hydroquinone, A(0H)O-, since similar species reduce nitroaromatic compounds (10, 19). The rate-determining step for each reaction is the transfer of the first electron as indicated by 2394

Environ. Sci. Technol., Vol. 28, No. 13, 1994

where F is the Faradays constant, R is the gas constant, Tis the absolute temperature, n is the number of electrons transferred, and a and p are empirical constants. To correlate rate constants for two compounds by a common electron donor, eq 2 can be written for reaction lb. If a and P are the same in both reactions, the two rate constants are related by (3) Equation 3 suggests that the log of the rate constant for the reductive dehalogenation of CCL (k'ccq) in different reducing systems should be linearly related to log k ' c 2 c ~ (or any other reference alkyl halide), but should be offset by a factor dependent on the difference in E ~ R ~ x . The use of eq 3 depends on knowing the values of EIRix. Half-wave reduction potentials for alkyl halides do not necessarily relate to the one-electron reduction potential because the reaction at the electrode is slow and irreversible (32),and therefore, the values of Ei for most alkyl halides X in this work are not well-known. Values of E ~ R used were calculated using an approach described by Hush (33) and Eberson (32). With this approach, which is described can be in more detail in the Appendix, values of EIR~X estimated from

= - [mf(R')(g)- TAS,(R')(g) - AGf(X-)(aq)AGf(RX)(g)+ RT In (HR/HRx)I/F (4) where AHf(R*)is the enthalpy of formation of the alkyl radical, ASf(R') is the entropy of formation of the alkyl radical, AGf(X-) is the Gibbs free energy of formation of the halide ion, AGARX) is the Gibbs free energy of formation of the alkyl halide, H R and HRXare the Henry's constants for the alkyl radical and alkyl halide, respectively, (g) refers to the gas phase, and (as) refers to the aqueous phase. Since H R is unknown, it is approximated by using H R H the , Henry's constant for the corresponding alkane. Reduction potentials calculated from eq 4 are listed in Table 1for C2C16,CC4,and bromoform (CHBr3). These calculated reduction potentials are for standard x 3 should be for conditions whereas values of E ~ Rin~eq the experimental conditions used in measuring the rate. However,standard reduction potentials can be used in eq 3 if it is assumed that activity corrections for the two radical anions cancel each other. At 25 OC, the estimated standard reduction potentials correspond to relative reactivities calculated from eq 3 of 1:10-5:10-14for C2C16,Cc4, and

Table 1. Thermodynamic Data Used in Computing One-Electron Reduction Potentials for CzCle, CC14, and CHBrs RX

AGf(RX)”(kJ/mol) -57.7 CZC~ cc4 -60.59 CHBr3 8

H(RX)b(-) 0.16 1.1 0.02

(kJ/mol) 35.2 79.5 188.4

&(E) (kJ/mol K) 403.6d 302.lf 304.7f

H(RH) (-) 0.09e 0.16b 0.04e

AHf(X-P (kJ/mol) -131.23 -131.23 -103.96

EIRX(V) 0.37 0.11 -0.56

Wagman et al. (34). Munz (35). McMillen and Golden (36). Taylor et al. (37).e Estimated from the ratio of vapor pressure to aqueous solubility (38).f O’Neal and Benson (39).g Tschuikow-Roux and Paddison (40).

CHBr3, respectively, if CY is assumed to equal unity as has been observed for the reduction of nitroaromatic comx CC14 equal to 0.11 V pounds (IO). The value of E l ~ for is bracketed by two previous estimates of 0.25 and 0.05 V (32). This range in reduction potentials results primarily from older values of aHf(R)equal to 58.6 and 87.8 kJ/ mol. Our calculation is based on a more recent value of AHf(R’) equal to 79.5 kJ/mol tabulated by McMillen and Golden (36). This value also agrees well with other recently reported values of 77.4 (41) and 78.6 kJ/mol (42).

Materials and Methods Experimental Approach. The relative rates of reduction of C2C16, CCL, and CHBr3 by 500 pM AHQDS was investigated, and results from these studies were compared with experiments conducted with 25 mg of C/L of NOM. The effect of adding either 250 pM Fe2+or 250 pM HS- to the NOM was also investigated. Control experiments were conducted with solutions containing the halogenated organic and 250 pM Fe2+or 250 pM HS-. The initial concentrations of the halogenated substrate was 2.5 pM in all cases, and the initial molar ratio of electron donor to acceptor was 100 or greater. The influence of adding hematin to the Fe2+ and HS- solution was also explored. Sample and Solution Preparation. All experiments were conducted in flame-sealed borosilicate glass ampules prepared in an anaerobic glovebox as described previously (6,43).Aqueous solutions were filter-sterilized by passage through a sterile 0.2-km nylon filter (Nalge Co., Rochester, NY) immediately before being dispensed into autoclaved ampules. The samples were then covered with a small piece of poly(viny1idenechloride) film (Saran) held in place by latex tubing, placed in an Nz-filled container, and removed from the glovebox. The alkyl halide was added as a 2.5-pL spike of a methanol solution, and the ampule was sealed with a propane-oxygen flame while the top of the ampule was flushed with filter-sterilized N2. The 0 2 concentration in the methanol solution was minimized by diluting a stock solution that was saturated with atmospheric 0 2 by 1:250 (v/v) withmethanol that was degassed and stored in the anaerobic glovebox. Stock solutions of the AHQDS were prepared by reducing the sodium salt of 2,6-anthraquinone disulfonic acid (AQDS) (Aldrich Chemical Co., Milwaukee, WI) by a method adapted from Tratnyek and Macalady (19).The AQDS was added to boiled Milli-Qwater (Millipore Corp., Bedford, MA) that had been cooled under a stream of N2. The solution was purged with Hz gas (99.999% ) for 30 min, and then six palladium-coated, alumina catalyst pellets (Aldrich Chemical Co., Milwaukee, WI) were added to the bottle, which was then quickly recapped. The solution was sparged for 12 h with a stream of purified HZat 10 mL/min to reduce the AQDS to AHQDS and then placed in the glovebox. This reduced solution was first filtered

with a 0.2-pm nylon filter and then with a filter cartridge containing 300 mg of Cls-silica (Alltech Associates, Deerfield IL). The AHQDS was added to a buffer solution (described below) in the glovebox to give a final concentration of 500 pM AHQDS. The AHQDS stock solution was analyzed by ion-pair, high-pressure liquid chromatography (44)) UV-vis spectroscopy, and titration with Cr(V1). The results from these three different analyses were consistent with the complete reduction of AQDS to AHQDS. All buffer solutions contained 10 mM Na2S04 (Baker Chemical Co., Phillipsburg, NJ). A 5 mM buffer solution was prepared from the sodium and free-acid forms of (morpho1ino)propanesulfonic acid (MOPS) (pK, = 7.6) (Sigma Chemical Co, St. Louis, MO), and small adjustments to reach the desired pH were made by the addition of either 0.1 N H2SO4 or 0.1 N NaOH. The pH of the Fe2+ solutions was adjusted to 7.2 while the pH of the HSsolutions was 7.8. These values were selected so that the pH in the two different systems were as close as possible while not precipitating ferrous hydroxide at the higher pH in the Fez+ samples nor driving H2S into the headspace of the ampules at the lower pH in the HS- samples. A 5 mM tris(hydroxymethy1)aminomethane (Tris) (Sigma) buffer was used in investigations of the effect of pH on the reduction rate of humic acid. The MOPS and Tris buffers were used because both complex Fez+weakly relative to more common pH 7 buffers, such as phosphate. The study of the effects of pH on the rate of reduction of CzC16 by AHQDS used 5 mM solutions of the following buffers: potassium hydrogen phthalate, potassium hydrogen phosphate, sodium hydrogen carbonate, sodium carbonate, and sodium borate (all from Baker) and morpholino-ethane sulfonic acid (Sigma). Solutions containing Fe2+were prepared in the glovebox by adding a gravimetrically determined amount of solid FeSOp9H20 (Baker) to the buffer solution. A stock solution containing HS- was prepared from a single crystal of Na2S (Baker) that was washed with N2 sparged Milli-Q water to remove any opaque coatings from the crystal, dried with a tissue, and added to a preweighed, nitrogenflushed vial. After the crystal was placed in the vial, the vial was quickly reweighed and placed in the anaerobic glovebox. The hematin used was isolated from bovine blood and used as received (Aldrich). A stock solution of hematin was prepared in the anaerobic glovebox by dissolving 0.1 g of the hematin in 100 mL of deoxygenated water that contained 0.5 g of Na2C03. This solution was diluted with the buffer solution to give a final hematin concentration equal to 0.1 pM. The NOM used was obtained from several different sources. Humic acid was extracted from soils purchased from the International Humic Substance Society (IHSS) (Golden, CO) and is classified by the IHSS as “soil humic acid”. The soil humic acid was extracted and purified Envlron. Sci. Technol., Vol. 28, No. 13, 1994

2395

3

2'5

2.0

5 2 O3

0.4

3 1.5

X

1.0

8

d

h

r:

U

0.5

%

.......

v

CY

n "."n

0.0

l l Z 4 .............................................

2.0

4.0

8.0 Time (d)

6.0

10.0

12.0

14.0

Flgure 2. Disappearance of 2.5 MM c$& and appearance of C&l4 in the presence of 500 pM AHQDS at 50 OC and pH 7.2. The CzHC15 concentration (dotted line) is a hypothetical value obtained using the observed appearance of C2CI4and the elimination rate constants from Cooper et al. (22) assumingthat 100% of c2c16 is transformed to c2Cl4 via C2HC15.

accordingto the standard procedures outlined by the IHSS. This procedure included purifying the humic acid to obtain a low inorganic content by treatments with NaOH, HCl, and HF followedby dialysis against Milli-Q water. A single experiment was conducted with IHSS aquatic humic acid that was remaining from previous studies in our laboratory (45).

Organic matter was also extracted from the Borden aquifer solids (BOM) (46). This organic matter was extracted from a 2-kg sample of homogenized Borden sand with 0.1 N NaOH under an N2 atmosphere. The supernatant was transferred to 0.25-L polyethylene bottles and centrifuged (2000g, 45 min) to remove fine grain solids. The supernatant was decanted, and the organic matter was precipitated by lowering the pH to 3 with 5 N HC1 and chilling to 4 "C. In the case of BOM extraction, the pH was not lowered below 3.0, and the extract was not further purified because it was hypothesized that this purification could remove an electron transfer mediator. This extract was neutralized to pH 6.0 with NaOH and diluted with a 2.5 mM pH 5.8 phthalate buffer to give a concentration of 20 and 40 mg of C/L of BOM. The solution was filter-sterilized with a 0.2-pm nylon filter and added to autoclaved ampules. Sample Analysis. Ampules were periodically sacrificed, and 1 mL of the aqueous solution was extracted with 1mL of hexane in a 2.5-mL borosilicate vial equipped with a Teflon-lined silicone septum. Hexane extracts containing CC4, C2C16,and CHBr3 were analyzed using gas chromatography and electron capture detection (6). All pH measurements were made at the temperature of the reaction (50 "C) using calibration buffers maintained at the reaction temperature. Results and Discussion

Reduction by Hydroquinones. Figure 2 illustrates the transformation of C2C16to C2Cl4 in the presence of 500 pM AHQDS a t pH 7.2 and 50 OC. The observed disappearance of C2C16fit the first-order model (r2= 0.90), and the formation of the C2Cl4 appears to be quantitative. This is evident by comparing the CzC14 data with the CzC14 concentrations calculated by assuming quantitative conversion of C2C16to CzCl4. C2HC15 was not a significant intermediate. Throughout the experiment, the CzHC15 concentration never exceeded the limits of quantification of 0.005 pM. These nondetectable concentrations can be 2396

Environ. Sci. Technol., Vol. 28, No. 13, 1994

3

O,

0.75 0.50

t

2 0.25

1

0.00

PH Figure 3. The pH dependence of the pseudo-first-order rate constant for disappearance of c2c18 by 500 p M AHQDS at 50 OC. Error bars represent the standard error of the pseudo-first-order regression.

compared to intermediate C2HC15 concentrations calculated by assuming that all of the C2Cl6 was first reduced to CzHC16 (reactions 1, 5, and 8 of Figure 1). Assuming that reaction 5 is fast relative to reaction 1,the concentration of C2HC15 can be calculated from (47)

where [C&l6]0 is the initial C2C16concentration, t is the time, and k'C2Cb is the observed pseudo-first-order rate constant for the disappearance C2C16. Figure 2 illustrates the hypothetical C2HC15 concentrations calculated from eq 5 when ke, the elimination rate constant, was calculated from the Arrhenius parameters reported by Cooper et al. (22). CzHC15 would have reached a maximum concentration of 0.1 pM after approximately 0.4 d when the first sample was analyzed. Several other reported values of ks (21-25) result in slightly larger intermediate C2HC15 concentrations. This calculation ignores potential impacts of buffer catalysis or promotion of the dehydrohalogenation rate by deprotonated anthrahydroquinone nucleophiles. Nevertheless, the data suggest that if CzHCl5 formation occurs, it is only a minor reaction pathway in these experimental systems. The pH dependence was investigated over a pH range from 4.5 to 11.5, and the observed k'c2cbare plotted versus the solution pH in Figure 3. This figure illustrates that k'czck increased from approximately zero a t pH 4.5 to 1.2 X lo4 s-1 at pH 11. The overall rate expression for reaction in solution can be considered to be the sum of three independent, parallel, first-order reactions: k'c,cg = ko[A(OH)J

+ k1[A(OH)O-l + k,[A(O);-l

(6)

where [A(OH)2], [A(OH)O-I, and [A(0)t2-1 represent the concentrations of the fully protonated, the monophenolate, and the diphenolate forms of AHQDS and ko, ki, and k2 are the corresponding rate constants. The equation may be rewritten in terms of the total AHQDS concentration EA(OH)PIT: k'cp, = [A(OH),lT(koao + k1aI + k

2 4

(7)

~~

~~

Table 2. Rate Constants for Disappearance of C&16 in Several Systemsa experimental system

PH

k (d-l)*

r2

NO

% recoveredd

BOM BOM (40 of mg/L of DOC) IHSS aquatic humic acid IHSS soil humic acid Fez+ Fez++ BOM HSHS- + BOM Fez+ Fe2++ IHSS humic acid Fez++ hematin HSHS- IHSS humic acid HS- + hematin AHQDS HzOe

6.0 5.9 5.6 7.8 5.8 5.8 8.7 8.7 7.2 7.2 7.2 7.8 7.8 7.8 7.8 7.8

-0.00001 f 0.00002 0.00001 f 0.00001 0.0024 f 0.0011 0.0045 f 0.005 0.014 f 0.001 0.214 f 0.031 0.089 f 0.016 0.65 f 0.14 0.010 f 0.007 0.103 f 0.010 0.095 i 0.018 0.029 f 0.004 0.26 f 0.15 1.303 f 0.50 0.059 i 0.003 2.0 f 10-10

0.01 0.002 0.52 0.02 0.97 0.90 0.85 0.76 0.74 0.96 0.86 0.81 0.67 0.70 0.99

7 6 8 7 7 7 7 10 7 6 4 10 5 4 6

100 f 5 100 f 1 102 i 2 98 f 5 98 & 4 95f7 93 f 7 90 f 5 77 f 5 59 f 4 71 f 3 84 f 14 70 f 10 57 f 15 88 f 8

+

Conditions: 250 pM Fez+,250 pM HS-, 25 mg of C/L of humic acid, 0.1 mM hematin, and 500 pM AHQDS, 50 OC; BOM was present at 20 mg/L except as noted. Slope and standard error of the linear regression. N = the number of samples used in the regression. Represents the average and standard deviation of the observedmoles of CzC4 plus C2Cls relative to the initial CzCleadded. The only reductive dehalogenation products detected were CzC14. e The hydrolysis rate computed from ref 23.

where ao, CY^, and a2 are the fractions of the three forms of AHQDS present in solutions. The values of ao, a1, and 0 2 were calculated from the solution pH and the first and second dissociation constants for the two phenolic protons of AHQDS (pK1 = 8.1 and pKz = 10.5) (48). Using the computed values of ao, al, and a2, a multivariate linear regression model (49)indicated that ho was not statistically different from zero (-0.1 X lo4 f 6.1 X lo4 M-l s-l). A second fit of the model that included only the A(0H)Oand A(0)zZ- species gave a value of 0.031 f 0.003 M-l s-l for k l and 0.24 f 0.02 for kz. Figure 3 illustrates that between the pH range from 6 to 8, the overall reaction is dominated by the reduction by A(0H)O-. However, at pH values greater than 8, the contribution of the diphenolate anion becomes increasingly important. The solid line in Figure 3 illustrates that over the entire pH range, the reduction of C2C16is well-described by the sum of the reactions with A(0H)O- and A(0)22-. The reaction with A(0)z2-is approximately 8 times faster than A(0H)O-while the rate constant for A(OH)2 is not statistically different from zero. These results follow the same trend observed for the reduction of the nitro group of 4-chloronitrobenzene by two naphthoquinone isomers where it was assumed that ko was zero and k2 was estimated to be 500 times greater than k l for both naphthohydroquinones (10). Reduction of C&16 by NOM. The rate of disappearance of C2C16in the presence of NOM was very slow but was increased markedly by the addition of a bulk electron donor. The rate constants were measured at 50 "C and under the conditions listed in Table 2. In a solution containing only the IHSS humic acid at pH 7.8, some CzCl4 was observed but the rate of CzC14 formation was statistically insignificant. Similar slow or insignificant rates were obtained with 25 mg of C/L of unreduced IHSS aquatic humic acid and with either 20 or 40 mg of C/L of BOM. This slow reaction with untreated NOM might be expected given that the redox active groups in NOM are probably oxidized during the extraction procedures if they are not already in the oxidized form in nature. Figure 4 and the data in Table 2 also illustrate the effect of combining the bulk electron donors HS- or Fez+with NOM on the disappearance of C2C16from solution. C2C16 disappeared at a rate of 0.029 d-l in the presence of 250 pM HS- and at rate of 0.01 d-l for 250 pM Fez+. The

0 Humic Acid - H S 0.1

15 20 25 Time (d) Flgure 4. Disappearance of C2Cleat 50 OC in the presence of 25 mg of C/L of IHSS humic acid without and with 250 pM Fez+ or 250 pM HS-. Lines indicate fit of first-order model to the data. 0

5

10

addition of 25 mg of C/L of IHSS humic acid to these solutions increasedthe CzC16disappearance rate by a factor of 9 for HS- and 10 for Fe2+. The effect of adding 20 mg of C/L of BOM to solutions containing Fez+or HS- was analogous to the results with the IHSS humic acid (Table 2). Specifically, the observed rates in systems containing BOM and the bulk electron donors Fe2+or HS- were faster by a factor of 8 and 10, respectively, relative to solutions containing only the bulk electron donors. The effect of pH on the CzCl6 reduction rate in the presence of IHSS humic acid was investigated in solutions containing 25 mg of C/L of IHSS humic acid and 250 pM HS-. The pH was varied from 7.2 to 8.0 using a Tris buffer, and the results are shown in Figure 5. The OH- concentration was calculated from the solution pH. This firstorder dependence of the reaction on OH- probably reflects the deprotonation of weak acids groups of the NOM since the reaction of OH- and C2C16is known to be slow (23). The observed increase in the pseudo-first-order rate constant resulting from the addition of NOM to Fe2+or HS- cannot be explained by assuming two parallel, independent reactions. If two reactions are parallel, the overall observed rate constant would equal the sum of the individual rate constants. Summing the independent rate constants does not account for the observed increase in the rate in solutions containing HS- or Fez+ and NOM Environ. Sci. Technoi., Vol. 28, No. 13, 1994

2397

Table 3. Rate Constants for Disappearance of CCld and CHBrs in Several Systems. 0.020 0.015

I

experimental system

v3

0.010

L 0.005

1

0.000 0.0

1

Q

2.0

4.0 6.0 8.0 OH- Concentration (M ~

10.0 12.0 0 ~ )

Figure 5. Pseudo-first-order rate constants for the disappearance of

CpCIBin the presence of 250 p M HS- and 25 mg of C/L of IHSS humic acid at 50

O C

versus OH- concentration.

since the rate in the presence of the NOM is approximately equal to zero (Table 2). These results indicate that the reductants interact or react with NOM by some pathway to form a more reactive reducing reagent. The observed increase in k'czck resulting from the addition of NOM to the bulk electron donor solutions suggest the NOM may act as electron transfer mediators as indicated in Scheme 1. One plausible reaction pathway is that quinone groups present in NOM are first reduced to hydroquinones, which then react with C2C16. The observed increase in the rate with increasing pH is consistent with more extensive dissociation of the hydroquinones in the organic matter. This pathway is analogous to that observed for C2C16reduction by AHQDS and for the reduction of nitroaromatic compounds by two naphthoquinones (10). The concept of reactive hydroquinone groups in NOM has more recently been extended to nitroaromatic reduction in the presence of NOM and HS- as the bulk electron donor (11). However, other reactive groups in the organic matter such as proteins may also be significant (12). The mass accounted for in these systems is also summarized in Table 3. When humic acid was used with either HS- or Fe2+as the bulk electron donor, the CzC14 yield was 59-70 mol 7%,respectively. Trichloroethene and dichloroethene, which are products of the microbially mediated reduction of C2C14(50),were not detected in the extracts. Relative Rates of Reduction of C2Cl6, CC14, a n d CHBr3. To generalize on the results presented above for C2C16,the disappearance of C C 4and CHBr3 was studied separately in several of the same reducing systems used for C2C16. When AHQDS was used as the reductant at pH 7.8, the relative reaction rates decreased in the following order: CzC16 > C c 4 >> CHBr3 (Table 3). The effect of IHSS humic acid on the disappearance of CCl4 was analogous to that for CzC16. In the presence of 250 pM Fe2+,the rate of CC14disappearance was 0.005 f 0.004 d-1 (1 standard deviation), which is not different from zero a t the 95% confidence level. The addition of 25 mg of C/L of IHSS humic acid to the Fe2+solution increased the observed rate constant to 0.092 d-l. The disappearance rate with 250 pM HS- present was 0.021 d-l, and the addition of 25 mg of C/L of IHSS humic acid increased the rate to 0.039 d-l. In the Fe2+-IHSS humic acid system, C C 4transformation was approximately three 2398

Envlron. Sci. Technol., Vol. 28, No. 13, 1994

CC4 Fez+ Fez+ + IHSS humic acid Fez+ + hematin HSHS- + IHSS humic acid HS- + hematin AHQDS H20e

CHBr3 Fe2+ Fez+ + IHSS humic acid Fez+ + hematin HSHS- + IHSS humic acid HS- + hematin AHQDS H2Of

pH

k (d-1)*

r2

Ne yield

%d

CHC13 7.2 7.2 7.2 7.8 7.8 7.8 7.9 7.8

0.005 f 0.004 0.092 f 0.005 0.028 f 0.005 0.021 f 0.004 0.039 ic 0.005 0.155 f 0.025 0.021 f 0.003 0.0014

0.74 0.98 0.89 0.93 0.91 0.88 0.88

6 7 6 7 8 6 7

7.2 7.2 7.2 7.8 7.8 7.8 7.8 7.8

0.005 f 0.002 0.006 f 0.003 0.006 f 0.003 0.000 f 0.005 0.003 f 0.011 0.124 f 0.030 0.0001 f 0.003 0.0008

0.91 7 0.47 6 0.46 7 0.00 9 0.10 8 0.68 10 0.03 7

41 f 4 17 f 9 32 f 6 20 f 4 22 f 2 76 f 9 33 f 4

CHzBr2 nda nd nd nd nd 43 f 10

nd

Conditions: 250 pM Fe2+,250 pM HS-, 25 mg of CIL of humic acid, 0.1 pM hematin, and 500 pM AHQDS,50 O C . bSlope and standard error of the linear regression. Number of samples in regression. Calculated as the observed product concentration divided by the amount of reactant that disappeared times 100. e Jeffers et al. (23);the hydrolysis rate is slower at pH 7.2. f Mabey and Mill (24). g Not detected.

times faster than the observed rate in the corresponding' HS--1HSS humic acid system. This was the only case when a compound reacted faster in solutions containing Fe2+ than in solutions containing HS-. CHC13 was also formed in humic acid systems with a yield of only 2040%. Reaction products of CC14other than CHC18 remain unknown although dichloromethane (the dechlorination product of CHC13)and CzC16(the radical coupling product) were not detected by gas chromatography. CHBr3 was relatively nonreactive in all of the systems studied containing natural organic matter. The results for the reduction of C2C16,CC4,and CHBr3 in systems containing hematin and Fe2+ or HS- are also summarized in Table 3. The addition of 0.1 pM hematin to the Fez+ and HS- solutions increased the rate of reduction of C2C16 by factors of 10 and 45, respectively. Similarly, the rates of the disappearance of CC4 were factors of 6 and 7 times larger for the Fe2+and HS-systems containing hematin. CHBr3 did not react in systems containing Fe2+and hematin, but it did disappear in the HS-/hematin systems with an initial rate of 0.12 d-l. This was the only case where CHBr3 reacted at a significant rate. The initial rate observed in the hematin/HS-system for CHBr3 was faster than that of CC14. However, the pseudo-first-order plot for CHBr3 was nonlinear, so that after 24 d, the CC14 concentration was less than that of CHBr3. This nonlinearity may indicate deactivation of hematin by CHBr3 or the reaction products. The total concentration of the hematin was always a factor of 20 less than the concentration of the alkyl halide. It was observed that CC14 and C2Cl6 concentrations decreased in some cases to zero, while during the same time, less than 10% transformation would be expected in solutions containing only Fez+ or HS-. These results are consistent with hematin acting as an electron transfer mediator in the reduction of the cc14 and C2Cle by Fez+ and HS-.

Two NOM samples, IHSS humic acid and BOM, both transformed C2C16 and ccl4 more rapidly upon the addition of a bulk electron donor. The transformation of C2C16and CCl4 in NOM solutions with Fez+or HS- as the bulk electron donors is consistent with NOM acting as electron transfer mediator and hydroquinone groups being the reactive sites. This proposition is based on (1)the slow reduction observed with only the bulk electron donor, (2) the observed reduction of CC14 and C2Cl6 by ionized AHQDS in the model system studied here, (3) the observed increase in reaction rate with increasing pH in systems containing either AHQDS or NOM, and (4) previous reported reactions indicating that both HS- and Fez+can reduce the quinone groups in simple quinones (17,511 as well as in humic acid (14-16). Alternative reaction pathways could also play an important role in the transformations that we observed. For example, HS- can be incorporated into NOM (52), and it is plausible that the NOM-sulfur species could react with alkyl halides since alkyl halides do react with thiolates (53) and polysulfides (54). Complexation of Fez+ by NOM could make the ferrous iron a better reducing agent, resulting in a faster reduction of CCl, and C2Cl6. Therefore, detailed studies are still needed to confidently characterize NOM as an electron transfer mediator. A significant log-log correlation of the rates of C2C16 and CC14reduction in a variety of different solutions was observed. The relative trends in the reaction rates of CzCh and C C 4 agreed qualitatively with a linear free energy relation based on estimated one-electron reduction potentials. A more quantitative assessment was limited by the uncertainties in the computed values of the oneelectron reduction potentials and because the reactions in the different experimental systems may react by different mechanisms. Nevertheless, predicted trend of rates (C2Ch > CC14 > CHBr3) agreed with the observed data. The concentrations of dissolved NOM (25 mg of C/L) used here are low relative to typical organic carbon concentrations in the subsurface if sorbed organic carbon is also considered. For example, an aquifer containing solids with 0.01% organic carbon and a porosity of 0.3 would contain approximately 5000 mg of C/L of aquifer. If both sorbed and dissolved organic matter react with halogenated hydrocarbons at comparable rates in the presence of Fe2+or HS-, then the reduction in organicrich, reducing sediments would be expected to be more rapid that in less reducing systems with less organic carbon. Jafvert and Wolfe (55) studied the reduction of CZC16 in the presence of a pond sediment that was probably rich in organic carbon and reported rates that were 3 or 4 orders of magnitude faster than the rates we observed in this work. It is not known, however, if other forms of NOM such as fulvic acid or humins would also possess the capability of accelerating halogenated hydrocarbon reduction.

1

loo

t

10-3 1 1IIII 10-2

- 1 I

I

I I

IIIII

I

I

10" k' CZC16 (d-'1

1

I IIIII

I

1oo

Figure 6. Relation of the rate constant for the disappearance of C&Ie and CCI4 in reduced solutions at 50 OC. The solid line indicates the linear regression: log Kcc,, = 0.64 log Kc2cC- 0.83.

CC14 was also transformed in AHQDS solutions at pH 7.8 at a measurable rate but CHBr3 was not. Tables 2 and 3 also indicate the hydrolysis rates at 50 "C. A comparison of these values with the reported reduction rates suggests that the reduction of C C 4 and C2C16 may be more important in many cases since the hydrolysis rate is so slow. Evaluation of Relative Rates of Reduction. The observed relation between the rate constants for the disappearance of C2C16 and C C 4 in seven different experimental systems is illustrated in Figure 6. The regression equation relating the two rate constants is

which has an r2 value equal to 0.80. The observed trend indicates that (1)the rate of C C 4transformation increases with increases in the rate of C&16 transformation and (2) the rate of C C 4 transformation is slower than that for C2C16which has a higher reduction potential. One possible cause of the scatter in Figure 6 is that the rate constants represent a range of different experimental systems including NOM, AHQDS, and hematin, and therefore, the basic assumptions made in deriving eq 3 may be invalid. The constant -0.83 in eq 8 implies that the value of a in eq 3 is approximately 0.2. This value of a and the reduction potentials in Table 1 suggest transformation rates for CHBr3 that are several orders of magnitude slower than those for CC4. These slow rates are consistent with the small or immeasurable rate constants observed for CHBr3.

Summary and Conclusions This work demonstrates that AHQDS reduces C2Cl6 and CCl4 in aqueous solutions at 50 "C and pH 7.8. C&16 and CC4were transformed with half-lives of 12 and 31 d, respectively. CHBr3 did not react under identical condition. The rate of reduction of C2C16 increased with increasing pH and was modeled as three independent reactions with A(OH)2, A(0H)O-, and A(0)z2-. The reaction rate of A(OH)z was not significantly different from zero, whereas the rate constant for A(0H)O- equaled 0.031 M-l s-l and the rate constant for A(0)z2- equaled 0.24 M-' s-l.

Acknowledgments This work was supported by the R. S. Kerr Environmental Research Laboratory of U.S. Environmental Protection Agency in Ada, OK, through CR-812462 and CR-814823(Dermont Bouchard and Stephen R. Hutchins, Project Officers) and by a grant from the Shell Companies Foundation. Although the research described in this article has been funded in part by the U S . Environmental Protection Agency, it has not been subject to Agencyreview Envlron. Sci. Technol., Vol. 28, No. 13, 1994

2399

and therefore does not necessarily reflect the views of the agency, nor should any official endorsement be inferred. We thank Michelle Kreigman-King, Pei Chin Chiu and an anonymous reviewer for their helpful comments.

Appendix

Estimation of One-ElectronReduction Potentials. This appendix illustrates the method used for calculating the one-electron reduction potentials for halogenated hydrocarbons. The one-electron reduction potential for RX can be calculated from the standard Gibbs free energy of RX(aq)

+ 1/2H,(g)

-

R'(aq)

+ X-(aq) + H+(aq)

(A-1)

For the normal hydrogen electrode (AGf(Hz)(g) = AGp (H+)(aq) = 0 by convention), the standard Gibbs free energy of reaction A-1 is (32) AG- = Gf(R*)(aq)+ G,(X-)(aq) - G,(RX)(aq)

(A-2)

where Gf is the Gibbs free energy of formation. Hush (33) and Eberson (32) related the free energies of formation in the aqueous phase to that in the gas phase by Gf(RX)(aq)= Gf(RX)(g)+ RT In (HRx)

(A-3)

Gf(R')(aq) = Gf(R')(g) + RT In (HRH)

(A-4)

and

where HRXis the Henry's constant for RX and HRHis the Henry's constant for RH used as an approximation of the Henry's constant for R' (32). Values for AGf(RX)(g)and AGf(X-)(aq) have been tabulated (341, and Henry's constants for many alkyl halides have been experimentally determined (35, 38, 56). The value of AGf(R')(g) is obtained by first expanding the quantity into its enthalpic and entropic components: AGf(R')(g) = mf(R')(g) - TASf(R')(g)

(A-5)

where IWf(R')(g) is the enthalpy of formation and ASr (R')(g) is the entropy of formation of the alkyl radical. O'Neal and Benson (39) estimated ASf(R')(g) for most of the chloro- and bromomethyl radicals as well as the trichloro- and tribromomethyl radicals. Values of ASr (R*)(g)for other radicals can be estimated (39). The uncertainty in the estimated value of ASf(R')(g) is likely to be approximately 2.5 kcal/mol, which is less than the variations in the reported values of AHf(R*).The standard reduction potential can be calculated by combining eqs A-2-5 and the relation E, = -AG/F to yield

ElRX = - [aHf(R')(g) - TASf(R')(g)- AGf(X-)(aq)AGf(RX)(g)+ RT In (HR/HRx)l/F (A-6) A similar expression can be obtained in terms of the bond dissociation energy (DRx)rather than A"f(R*) since these latter two quantities are directly related to each other t 39).

Literature Cited (1) Macalady, D. L.; Tratnyek, P. G.; Grundl, T. J. J. Contam. Hydrol. 1986, 1, 1-28. 2400

Environ. Scl. Technol., Vol. 28, No. 13, 1994

(2) Wolfe, N. L.; Macalady, D. L. J. Contam. Hydrol. 1992,9, 17-34. (3) Wade, R. S.; Castro, C. E. J. Am. Chem. SOC.1973,95,22630. (4) Klecka, G. M.; Gonsior, S. J. Chemosphere 1984,13,391402. (5) Gantzer, C. J.; Wackett, L. P. Environ. Sci. Technol. 1991, 25, 715-21. (6) Curtis, G. P. Ph.D. Thesis, Stanford University, Stanford, CA, 1991, 228 pp. (7) Peijnenburg, W. J. G. M.; Hart, M. J.; den Hollander, H. A.; van de Meent, D.; Verboom, H. H.; Wolfe, N. L. Enuiron. Toxicol. Chem. 1992,11, 301-314. (8) Kriegman-King,M. R.; Reinhard, M. Enoiron. Sci. Technol. 1992,26, 2198-2206. (9) Kriegman-King, M. R.; Reinhard, M. Enuiron. Sci. Technol. 1994,28, 692-700. (10) Schwarzenbach,R.P.;Stierli,R.;Lanz, K.;Zeyer, J.Enuiron. Sci. Technol. 1990, 24, 1566-75. (11) Dunnivant, F. M.; Schwarzenbach, R. P.; Macalady, D. L. Enuiron. Sci. Technol. 1992, 26, 2133. (12) Carreira, L. H.; Wolfe, N. L. Preprint Extended Abstract, Presented before the Division of Environmental Chemistry, American Chemical Society, San Francisco, CA, Apr 5-10, 1992, pp 511-12. (13) Senesi, N.; Schnitzer, M. Soil Sci. 1977, 23, 224-34. (14) Glebko, L. I.; Ulkina, J. U.; Maximov, 0. B. Mikrochim. Acta 1970, 1247-54. (15) Schnitzer, M.; Riffaldi, R. Soil Sci. SOC.Am. Proc. 1972,36, 772-7. (16) Maximov, 0. B.; Glebko, L. I. Geoderma 1974, 11, 17-28. (17) Hudlicky, M. Reductions in Organic Chemistry; Halsted Press: New York, 1984; 304 pp. (18) Hayes, M. H. B; O'Callaghan, S. G. In Humic Substances 11, In Search of Structure; Hayes, M. H. B., et al., Eds.; J. Wiley: Chichester and New York, 1989; 764 pp. (19) Tratnyek, P. G.; Macalady, D. L. J . Agric. Food Chem. 1989, 37, 248-54. (20) Eberson, L. Electron Transfer Reactions in Organic Chemistry; Springer-Verlag: Berlin, 1987; 234 pp. (21) Walravens, R.; Trouillet, P.; Devos, A. Int. J . Chem. Kinel. 1974,6, 777-86. (22) Cooper, W. J.; Slifker, R. A.; Joens, J. A.; El-Shazly, 0. A. In Biohazards of Drinking Water Treatment; Larson, R. A,, Ed.; Lewis Publishers, Inc.: Chelsea, MI, 1989; pp 3746. (23) Jeffers, P. M.; Ward, L. M.; Woytowitch, L. M.; Wolfe, N. L. Enuiron. Sci. Technol. 1989, 23, 965-9. (24) Mabey, W. R.; Mill, T. J. Phys. Chem. Ref. Data 1978, 7, 383-415. (25) Roberts, A. L.; Gschwend, P. M. Enuiron. Sci. Technol. 1991,25, 76-86. (26) Singleton, D. M.; Kochi, J. K. J. Am. Chem. Soc. 1967,89, 6547-55. (27) Criddle, C. S.; McCarty, P. L. Enuiron. Sci. Technol. 1991, 25, 9753-8. (28) Marcus, R. In The Nature of Seawater; Goldberg, E. D., Ed.; Dahlem Konferenzen: Berlin, 1975; pp 477-504. (29) Eberson, L. Adu. Free Radical Biol. 1985, I , 19-90. (30) Schwarzenbach, R. P.; Gschwend, P. M. In Aquatic Chemical Kinetics; Stumm, W., Ed.; John Wiley & Sons, Inc.: New York, 1990; pp 199-233. (31) Tratnyek, P. G.; HoignB, J.; Zeyer, J.; Schwarzenbach, R. P. Sci. Total Environ. 1991, 109jll0, 327-41. (32) Eberson, L. Acta Chem. Scand. B 1982, 36, 533-43. (33) Hush, N. S. 2. Elektrochem. 1957, 61, 734-8. (34) Wagman, D. D.; Evans, W. H.; Parker, V. B.; Schumm, R. H.; Halow, I.; Bailey, S. M.; Churney, K. L.; Nuttall, R. L. J. Phys. Chem. Ref. Datu (Suppl.) 1982,2, 1-392. (35) Munz, C. Ph.D. Dissertation, Stanford University, Stanford, CA, 1985, 306 pp. (36) McMillen, D. F.; Golden, D. M. Annu. Reo. Phys. Chem. 1982,33,493-532.

(37) Taylor, P. H.; Dellinger, B.;Tirey,D. A.Znt. J. Chem. Kinet. 1991,23, 1051-74. (38) Mackay, D.; Shiu, W. Y. J . Phys. Chem. Ref.Data 1981,10, 1175-99. (39) O'Neal, H. E.; Benson, S. W. In Free Radicals; Kochi, J. K., Ed.; John Wiley & Sons: New York, 1973. (40) Tschuikow-Roux, E.; Paddison, S. Znt. J . Chem.Kinet. 1987, 19, 15-24. (41) Holmes, J. L.; Lowing, F. P. J. Am. Chem. SOC.1988,110, 7343-5. (42) Luke, B. T.; Loew, G. L.; McLean, A. D. J . Am. Chem. SOC. 1987, 109, 1307-11. (43) Barbash, J. E.; Reinhard, M. Enuiron. Sci. Technol. 1989, 23, 1349-58. (44) Hunter, A. T. J . Chromatogr. 1985, 39, 319-30. (45) Summers, R. S. Ph.D. Dissertation, Stanford University, Stanford, CA, 1986, 209 pp. (46) Curtis, G. P.; Roberts, P. V.; Reinhard, M. Water Resour. Res. 1986, 22, 2059-69. (47) Moore, J. W.;Pearson, R. G.Kineticsand Mechanism; John Wiley & Sons: New York, 1981; 455 pp.

(48) Clark, W. M. Oxidation-Reduction Potentials of Organic Systems; The Williams and Wilkins Co.: Baltimore, MD, 1960; 584 pp. (49) Draper, N. R.; Smith, H. Applied Regression Analysis, 2nd ed.; John Wiley & Sons: New York, 1981; 709 pp. (50) Vogel, T. M.; McCarty, P. L. Appl. Environ. Microbiol. 1985,49, 1080-3. (51) Baxendale, J. H.; Hardy, H. R.; Sutcliffe, L. H. Trans. Faraday SOC.1951,47, 963-73. (52) Francois, R. Geochim. Cosmochim. Acta 1987,51, 17-27. (53) Livesy, J. C.; Anders, M. W. Drug Metab. Disp. 1979, 7, 199-203. (54) Haag, W. R.; Mill, T. Environ. Toxicol. Chem. 1988,7,91724. (55) Jafvert, C. T.; Wolfe, N. L. Environ. Toxicol. Chem. 1987, 6, 827-37. (56) Gossett, J. M. Environ. Sci. Technol. 1987, 21, 202-8.

Received for review May 2, 1994. Accepted July 6, 1994." @

Abstract published in Advance ACS Abstracts, August 1,1994.

Envlron. Sci. Technol., Vol. 28, No. 13, 1994 2401