Mechanism of Persulfate Activation by Phenols - Environmental

Zongsu Wei , Frederick A. Villamena , and Linda K. Weavers. Environmental Science ..... Richard J. Watts , Mushtaque Ahmad , Amanda K. Hohner , Amy L...
0 downloads 0 Views 829KB Size
Article pubs.acs.org/est

Mechanism of Persulfate Activation by Phenols Mushtaque Ahmad,† Amy L. Teel, and Richard J. Watts* Department of Civil & Environmental Engineering, Washington State University, Pullman, Washington 99163-2910, United States S Supporting Information *

ABSTRACT: The activation of persulfate by phenols was investigated to further the understanding of persulfate chemistry for in situ chemical oxidation (ISCO). Phenol (pKa = 10.0) activated persulfate at pH 12 but not at pH 8, suggesting activation occurred only via the phenoxide form. Evaluation of the phenoxide activation mechanism was complicated by the concurrent activation of persulfate by hydroperoxide anion, which is generated by the base catalyzed hydrolysis of persulfate. Therefore, phenoxide activation was investigated using pentachlorophenoxide at pH 8.3, midway between the pKa of pentachlorophenol (pKa = 4.8) and that of hydrogen peroxide (pKa = 11.8). Of the two possible mechanisms for phenoxide activation of persulfate (reduction or nucleophilic attack) the results were consistent with reduction of persulfate by phenoxide with oxidation of the phenoxide. The concentration of phenoxide required for maximum persulfate activation was low (1 mM). The results of this research document that phenoxides activate persulfate via reduction; phenolic moieties ubiquitous to soil organic matter in the subsurface may have a significant role in the activation of persulfate during its injection into the subsurface for ISCO. Furthermore, the results provide the foundation for activation of persulfate by other organic anions without the toxicity of phenols, such as keto acids.



INTRODUCTION In situ chemical oxidation (ISCO) is a group of technologies that have been developed over the past 20 years to treat contaminated soils and groundwater. The most recent process developed for ISCO is activated persulfate, which has the advantages of wide reactivity with contaminants and high longevity in the subsurface.1 The most common activators of persulfate for ISCO include soluble iron, iron chelates, and base.2−4 The mechanisms of persulfate activation by iron and base are well understood. Activation of persulfate by iron or other transition metals proceeds through reduction of persulfate to generate sulfate radical (SO•− 4 ), a reaction parallel to the Fenton initiation reaction:5 −

− 3OS−O−O−SO3



− 2− SO•− + HOO• 4 + OH → SO4

(4)

Recent results suggest that organic compounds may also activate persulfate during ISCO; for example, Ahmad et al.9 demonstrated that soil organic matter (SOM) activates persulfate. Activation of persulfate by SOM has important implications for persulfate ISCO, but in contrast to iron and base activation, activation by SOM and other organic compounds is not well understood. SOM contains numerous phenolic moieties,10 and it may be that phenols or their ionized form, phenoxides, function as activators of persulfate. Ahmad et al.9 found that SOM activation of persulfate occurred significantly more rapidly at basic pH than at neutral pH, suggesting that phenoxide moieties may be the activating species. One proposed mechanism for phenoxide activation of persulfate is reduction of persulfate by phenoxide, similar to reaction 1:11

3+ + Fe 2 + → SO4 2 − + SO•− 4 + Fe

OH−

+ H 2O ⎯⎯⎯⎯→ 2SO4 2 − + HO2− + H+ (2)

Received: February 15, 2013 Revised: May 8, 2013 Accepted: May 10, 2013

The hydroperoxide anion generated in reaction 2 then reduces another persulfate molecule in a manner parallel to reaction 1, forming sulfate radical and superoxide radical anion (O•− 2 ): © XXXX American Chemical Society

(3)

Hydroxyl radical (HO•) is then generated through the reaction of sulfate radical with hydroxide:5

Other reducing agents can potentially activate persulfate, including zerovalent iron.6,7 In contrast, base activation of persulfate proceeds initially through base catalyzed hydrolysis, decomposing persulfate to sulfate and hydroperoxide anion (HO2−, the conjugate base of hydrogen peroxide):8 − 3OS−O−O−SO3

+ HO2−

+ •− → SO4 2 − + SO•− 4 + O2 + H

(1)



− 3OS−O−O−SO3

A

dx.doi.org/10.1021/es400728c | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology PhO− +



− 3OS−O−O−SO3

Article

→ SO4 2 − + SO•− 4 + PhOox

every 30 min, and additional sulfuric acid or sodium hydroxide was added as necessary to maintain the pH. The pH did not vary by more than 0.2 pH units. A relatively constant pH was maintained in the reactions in order to isolate the effect of phenols vs phenoxides on persulfate activation. At selected time points, a triplicate set of reactors was sacrificed for analysis. Sodium persulfate was measured by iodometric titration, and hydrogen peroxide was measured by visible spectrophotometry. Nitrobenzene concentrations were analyzed by gas chromatography (GC) after extracting the contents of the reactors with hexane. Positive control reactions containing persulfate but no phenol were conducted in parallel with each of the phenol− persulfate systems. Deionized water control systems, containing no persulfate and no phenol, were also evaluated in parallel for each system. All reactions were performed in triplicate, and the data were reported as the mean of the three replicates, with the standard error of the mean included as error bars. Persulfate Activation by Phenol vs Phenoxides. Reactions consisted of 0.5 M sodium persulfate, 1 mM phenol, and 1 mM nitrobenzene at pH 8 and pH 12. Additional control systems containing 1 mM phenol but no persulfate were conducted for these reactions to confirm that phenol and phenoxide have no direct effect on nitrobenzene degradation. To confirm the role of hydroxyl radical in phenoxide-activated persulfate systems, parallel reactions were conducted at pH 12 with the addition of 1 M 2-methyl-2-propanol to scavenge hydroxyl radical but not sulfate radical (kHO• = 5.2 × 108 M−1 s−1; kSO4•− = 8.4 × 105 M−1 s−1),15,16 for a 1000:1 molar ratio of scavenger to probe compound. Effect of Phenol Ring Substitution. The effect of ring substitution on phenoxide activation of persulfate was investigated using catechol, phenol, and the chlorophenols 2chlorophenol, 2,3-dichlorophenol, 2,4,6-trichlorophenol, 2,3,4,6-tetrachlorophenol, and pentachlorophenol. The pKa values for phenol, catechol, and 2-chlorophenol are 10.0, 9.3, and 8.6, respectively.17 The pKa values for 2,3-dichlorophenol, 2,4,6-trichlorophenol, 2,3,4,6-tetrachlorophenol, and pentachlorophenol are 7.6, 6.2, 5.4, and 4.8, respectively.18 Reactions consisted of 0.5 M persulfate, 2 M sodium hydroxide, 1 mM of one of the phenols, and 1 mM nitrobenzene at pH 13. Persulfate Activation by Pentachlorophenol at Varying pH. Persulfate activation by pentachlorophenol was studied at a range of pH regimes: 2.5, 3.5, 4.5, 5.5, 6.5, 7.5, 8.5, 9.5, 10.5, 11.5, and 12.5. Reactions consisted of 0.5 M sodium persulfate, 1 mM pentachlorophenol, and 1 mM nitrobenzene at pH 2.5−12.5. Mechanism of Phenoxide Activation of Persulfate. Reductive and nucleophilic mechanisms were studied as possible mechanisms for phenoxide activation of persulfate. Pentachlorophenol was used in this study as the activator because it is the most acidic of the chlorophenols used, and its activation of persulfate could be separated from base activation of persulfate. Reactions were conducted at pH 8.3, which is midway between the pKa of pentachlorophenol (4.8) and hydrogen peroxide (11.8). Reactions consisted of 0.5 M sodium persulfate, 1 mM pentachlorophenol, and 1 mM nitrobenzene at pH 8.3. Additional triplicate sets of reactors were established to measure pentachlorophenoxide concentrations; at selected time points, the reactor contents were adjusted to pH < 2, extracted with toluene, and analyzed by GC. To confirm the persulfate activation pathway, pentachlorophenoxide degradation products were analyzed by gas chromatography/mass spectrometry (GC/MS) within 2 h after extracting the contents

(5)

A second possible mechanism for persulfate activation by phenoxides is the Elbs persulfate oxidation, a nucleophilic attack on persulfate by the phenoxide, similar to reaction 2:11,12 PhO− +



− 3OS−O−O−SO3

→ HO2− + 2SO4 2 − + PhOproduct

(6)

Hydroperoxide formed in reaction 6 would then generate sulfate radical through reaction 3. In either mechanism, hydroxyl radical would be produced from sulfate radical through reaction 4. Elucidation of the mechanism of persulfate activation by organic compounds, including phenoxide-like groups in SOM and possibly some organic contaminants, is important because such an activation pathway is likely to play an important role in persulfate ISCO. However, a difficulty in investigating the mechanism is the competing activation of persulfate by hydroperoxide generated through the alkaline hydrolysis of persulfate at basic pH (reaction 2) vs activation by phenoxide at basic pH. The purpose of this research was to 1) develop a method for distinguishing between the effects of phenoxide and hydroperoxide on persulfate activation and 2) elucidate the mechanism of phenoxide activation of persulfate.



EXPERIMENTAL SECTION Materials. Sulfuric acid, nitrobenzene, potato starch, hexane (>98%), phenol (98%), and sodium hydroxide (reagent grade, 98%) were obtained from J.T. Baker (Phillipsburg, NJ). Sodium hydroxide was purified by adding 10 mM magnesium chloride to 1 L of 8 M NaOH, stirring for 8 h, and passing through a 0.45 μm membrane filter. Sodium persulfate (reagent grade), magnesium chloride (99.6%), 2-methyl-2-propanol (reagent grade), catechol (98%), 2-chlorophenol (>99%), 2,3-dichlorophenol (98%), 2,4,6-trichlorophenol (98%), 2,3,4,6-tetrachlorophenol (>99%), and pentachlorophenol (98%) were purchased from Sigma Aldrich (St. Louis, MO). Sodium thiosulfate (99%), potassium iodide, and methylene chloride were purchased from Fisher Scientific (Fair Lawn, NJ). Deionized water was purified to >18 MΩ•cm with a Barnstead Nanopure II ultrapure system (Dubuque, Iowa). Probe Compound. Although sulfate radical is generated in either of the possible mechanisms for phenoxide activation of persulfate, its detection is problematic because every potential probe compound that reacts with sulfate radical also reacts with hydroxyl radical.3 Sulfate radical reacts with hydroxide to form hydroxyl radical at alkaline pH (reaction 4) and also reacts with water at lower pH regimes to form hydroxyl radical:13 2− SO•− + HO• + H+ 4 + H 2O → SO4

(7)

Therefore, nitrobenzene was used as a hydroxyl radical-specific probe to evaluate persulfate activation because of its high reactivity with hydroxyl radical (k•HO = 3.9 × 109 M−1 s−1) and negligible reactivity with sulfate radical (kSO4•− ≤ 106 M−1 s−1).14−16 General Reaction Procedures. All reactions were conducted in 20 mL borosilicate vials capped with polytetrafluoroethylene (PTFE) lined septa with a total reaction volume of 15 mL and a temperature of 20 ± 2 °C. The reaction pH was adjusted using 0.1 N sulfuric acid and 0.1 or 1 N sodium hydroxide. As the reactions proceeded, the pH was monitored B

dx.doi.org/10.1021/es400728c | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

of the reactors with methylene chloride, and compared to authentic standards. Effect of Phenoxide Concentration on Persulfate Activation. Persulfate activation by pentachlorophenoxide was studied at a range of pentachlorophenol concentrations: 0, 0.01, 0.05, 0.025, 0.05, 0.1, 0.2, 0.4, 0.6, 0.8, 1.0, 2.0, 3.0, and 4.0 mM. Reactions consisted of 0.5 M sodium persulfate, 0−4 mM pentachlorophenol, and 1 mM nitrobenzene at pH 8.3. Analysis. Hexane extracts were analyzed for nitrobenzene using a Hewlett-Packard Series 5890 GC with a 0.53 mm (i.d.) × 15 m SPB-5 capillary column and flame ionization detector (FID). Chromatographic parameters included an injector temperature of 200 °C, detector temperature of 250 °C, initial oven temperature of 60 °C, program rate of 30 °C/min, and a final temperature of 180 °C. Toluene extracts were analyzed for pentachlorophenol using the same GC fitted with an electron capture detector (ECD). Chromatographic parameters included an injector temperature of 220 °C, detector temperature of 240 °C, initial oven temperature of 120 °C, program rate of 20 °C/min, and a final temperature of 240 °C. Phenolic compounds and their degradation products were analyzed by GC/MS. Samples were acidified to a pH of 1−2 with sulfuric acid and then extracted with methylene chloride. The extracts were analyzed by GC/MS using a Hewlett-Packard model 7890A GC/5975C mass spectrometer; the column used was a 30 m MDB-5ms Agilent column (Santa Clara, CA) with a 0.5 μm i.d. and 250 μm film thickness. Chromatographic parameters included an injector temperature of 250 °C; an initial oven temperature of 40 °C for 2 min, then programmed at a rate of 40 °C/min to 100 °C and held for 0.5 min, and finally raised to 300 °C at a rate of 10 °C/min and held for 3 min. Sodium persulfate concentrations were determined by iodometric titration with 0.01 N sodium thiosulfate.19 Hydrogen peroxide was measured spectrophotometrically after complexation with titanium sulfate at 407 nm using a Spectronic 20 Genesys spectrophotometer.20 Solution pH was monitored using a Fisher Accumet 900 pH meter (Fisher Scientific, Hampton, NH). The Statistical Analysis System package S.A.S version 9.1 was used to calculate the variances between the experimental data sets and 95% confidence intervals for rate constants.

Figure 1. Nitrobenzene degradation in persulfate systems with or without the presence of phenol (reactors: 0.5 M sodium persulfate, 0 mM or 1 mM phenol, 0 or 1 M of the hydroxyl radical scavenger 2methyl-2-propanol, and 1 mM nitrobenzene at pH 8 or 12; 15 mL total volume; control reactors contained no persulfate and either 0 mM or 1 mM phenol). Error bars represent the standard error of the mean for three replicates. a) pH 8; b) pH 12.



RESULTS AND DISCUSSION Persulfate Activation by Phenol vs Phenoxide. To investigate the effect of phenol dissociation on persulfate activation, reactions using nitrobenzene as a probe for hydroxyl radical were conducted 2 pH units above and below the pKa of phenol (pKa = 10) (Figure 1a−b). At pH 8, with only 1% of the phenol as phenoxide, nitrobenzene loss was negligible in both persulfate-only and phenol−persulfate systems over 5 h, indicating minimal hydroxyl radical generation (Figure 1a). However, with 99% of the phenol as phenoxide at pH 12, 73% nitrobenzene loss was observed in the phenol−persulfate system (Figure 1b). Only 35% of the nitrobenzene was lost in the parallel persulfate-only system at pH 12 through base activation of persulfate, indicating that approximately half of the nitrobenzene loss in the phenol−persulfate system was due to phenoxide activation of persulfate. Nitrobenzene loss in the two control systems without persulfate (with and without phenol) was negligible at pH 8 and pH 12. To confirm the generation of hydroxyl radical in the phenoxide−persulfate system at pH 12, the hydroxyl radical scavenger 2-methyl-2-propanol was added

to parallel reactions. In the presence of the scavenger, no detectable nitrobenzene degradation occurred, demonstrating that hydroxyl radical was responsible for nitrobenzene degradation in the phenoxide−persulfate system at pH 12. The results of Figure 1 confirm that 1) phenol activates persulfate to generate hydroxyl radical, and 2) the only form of phenol that activates persulfate is the anionic form, phenoxide. The activation by phenoxide is parallel to the activation of persulfate by hydrogen peroxide; persulfate is not activated by hydrogen peroxide itself but instead by its conjugate base, hydroperoxide anion (HO2−):8 H 2O2 ↔ HO2− + H+

pK a = 11.8

(8)

Effect of Phenol Ring Substitution. The effect of ring substitution of phenols was investigated as an initial step in elucidating the mechanism of phenoxide activation. To evaluate the effect of ring substitution on persulfate activation, several chlorine-substituted phenols as well as catechol were added to C

dx.doi.org/10.1021/es400728c | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

least 75% of the nitrobenzene was degraded.21 The relationship between the rate constants derived from Figure 2a and the pKas of the various phenols is shown in Figure 2b. The result was a linear correlation between the pKas of the phenols and the rates of phenoxide-activated persulfate degradation of nitrobenzene, with an R2 of 0.997. Increasing chlorine substitution of phenols results in both increased acidity as well as deactivation of the ring (with less reducing potential). Therefore, the results of Figure 2 are consistent with both potential mechanisms (nucleophilic attack vs reduction) for the activation of persulfate by phenoxides. Persulfate Activation by Pentachlorophenol at Varying pH. In phenoxide activated persulfate reactions at basic pH, the effect of phenoxide and the effect of basicity cannot be separated because both phenoxide and hydroperoxide (generated through the base-catalyzed hydrolysis of persulfate) act as persulfate activators. The ability of pentachlorophenoxide to activate persulfate (Figure 2) provided a methodology to separate phenoxide activation of persulfate (reaction 5 or 6) from hydroperoxide activation of persulfate (reaction 3). The pKa of pentachlorophenol is 4.8, while the pKa of hydrogen peroxide is 11.8; therefore, across a wide range of pH regimes, pentachlorophenol will be present as pentachlorophenoxide, while the hydroperoxide anion generated through the basecatalyzed hydrolysis of persulfate will be present as hydrogen peroxide, which does not activate persulfate.8 Loss of the hydroxyl radical probe nitrobenzene in pentachlorophenol− persulfate systems and persulfate-only systems at pH regimes ranging from 2.5 to 12.5 is shown in Figure 3a−f. Generation of hydroxyl radical was negligible at pH 2.5, which is expected because 99% of the pentachlorophenol is in the protonated state. At pH 4.5 (near the pKa of pentachlorophenol), 12% loss of nitrobenzene occurred in the pentachlorophenol−persulfate system with no detectable nitrobenzene loss in the persulfateonly system. Hydroxyl radical was generated increasingly in the pentachlorophenoxide−persulfate systems at pH 6.5 and 8.5 with minimal hydroxyl radical generation in the persulfate-only systems. At pH 10.5, there was no subsequent increase in pentachlorophenoxide-activated generation of hydroxyl radical from persulfate. At pH 12.5, hydroxyl radical was generated in the persulfate-only system due to base activation of persulfate and was generated at a rate equal to that in the pentachlorophenoxide−persulfate system. To illustrate the different effects of pentachlorophenoxide activation and base activation, nitrobenzene degradation rate data from the persulfate-only and the pentachlorophenolpersulfate systems of Figure 3 along with data for additional pH regimes were fit to first-order rate equations; the pseudo-firstorder rate constants were then plotted as a function of pH (Figure 4). Pentachlorophenoxide activation of persulfate increased as the pH was increased from pH 2.5 to pH 7.5, while no base activation of persulfate occurred in the persulfate only systems at this pH range. At pH regimes above 7.5, there was minimal further increase in the rate of pentachlorophenoxide activation of persulfate, because the pentachlorophenol in the system was completely dissociated. Above pH 10.5, base activation of persulfate in the persulfate-only systems increased as hydrogen peroxide became progressively more dissociated to hydroperoxide. The pentachlorophenoxide activation of persulfate correlated with the degree of pentachlorophenol dissociation, and the base activation of persulfate correlated with the degree of hydrogen peroxide dissociation. The results of Figure 4

basic persulfate systems containing nitrobenzene as a probe compound to evaluate hydroxyl radical generation (Figure 2a).

Figure 2. Nitrobenzene degradation in persulfate systems activated by phenols with varying ring deactivation (reactors: 0.5 M sodium persulfate, 2 M NaOH, 1 mM phenol, and 1 mM nitrobenzene at pH 13; 15 mL total volume; control reactors contained water in place of phenol and persulfate). Error bars represent the standard error of the mean for three replicates (a) and 95% confidence intervals (b). a) Nitrobenzene degradation over time; b) Nitrobenzene degradation rate constants plotted against the pKas of the phenols.

Persulfate concentrations remained constant during the reactions (p < 0.05). Nitrobenzene loss in deionized water control systems was minimal. Relative rates of phenoxide activation of persulfate were inversely related to the degree of chlorine substitution, with pentachlorophenoxide promoting a slower hydroxyl radical generation rate and the more reduced phenoxides promoting faster hydroxyl radical generation (Figure 2a). When the data of Figure 2a were plotted as ln [nitrobenzene]/[nitrobenzene]0 vs time to calculate pseudofirst-order rate constants, the data for catechol, phenol, and the mono-, di-, tri-, and tetrachlorophenols did not fit first-order kinetics for the entire duration of the reactions. Therefore, firstorder rate constants were calculated from time zero until at D

dx.doi.org/10.1021/es400728c | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Figure 3. Nitrobenzene degradation at varying pH regimes in persulfate systems with or without pentachlorophenol (PCP) (reactors: 0.5 M sodium persulfate, 0 or 1 mM pentachlorophenol, and 1 mM nitrobenzene; 15 mL total volume). Error bars represent the standard error of the mean for three replicates. a) pH 2.5; b) pH 4.5; c) pH 6.5; d) pH 8.5; e) pH 10.5; f) pH 12.5.

Figure 4. First-order reaction rate of nitrobenzene loss in persulfate-only and pentachlorophenol (PCP)−persulfate systems at pH 2.5−12.5, with percent ionization of pentachlorophenol and hydrogen peroxide. Error bars represent 95% confidence intervals.

E

dx.doi.org/10.1021/es400728c | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

would be generated through this mechanism. In contrast, if activation proceeded through the nucleophilic mechanism of reaction 6, hydroperoxide would be generated, and because it would exist in the unreactive form of hydrogen peroxide, no sulfate radical or hydroxyl radical would be produced through reactions 3 and 4 and no nitrobenzene oxidation would be observed. Therefore, the results of Figure 5 indicate that the reductive activation step shown in reaction 5 is the mechanism of phenoxide activation of persulfate. This mechanism is consistent with the results of Figure 2, in which the more reduced phenoxides were the strongest persulfate activators. The loss of pentachlorophenoxide in pentachlorophenoxideactivated persulfate systems at pH 8.3 is shown in Figure 6a.

provide further support that persulfate is activated by phenoxides, rather than phenols. Furthermore, the results of Figure 4 demonstrate that base activation predominates over pentachlorophenoxide activation of persulfate at pH regimes >11.5. Mechanism of Phenoxide Activation of Persulfate. The results of Figures 1−4 demonstrate that persulfate is activated only by the dissociated phenoxide form of phenols. Persulfate activation by phenoxide may occur through 1) the Elbs persulfate oxidation reaction, in which the phenoxide decomposes persulfate through nucleophilic attack and hydroperoxide is generated (reaction 6), or 2) the reduction of persulfate by phenoxide, with subsequent oxidation of the phenoxide (reaction 5).11 These mechanisms were investigated by 1) quantifying the generation of hydroperoxide, the product of nucleophilic attack of pentachlorophenoxide on persulfate, and 2) evaluating the likely oxidation products of pentachlorophenoxide generated during its reduction of persulfate. The generation of hydroperoxide in phenoxide-activated persulfate reactions was investigated using pentachlorophenoxide to activate persulfate and nitrobenzene as a probe for hydroxyl radical activity (Figure 5). Because of the high rate of

Figure 5. Concentrations of nitrobenzene and hydrogen peroxide in a pentachlorophenoxide-activated persulfate system at pH 8.3 (reactors: 0.5 M sodium persulfate, 1 mM pentachlorophenol, and 1 mM nitrobenzene at pH 8.3; control contained no pentachlorophenol; 15 mL total volume). Error bars represent the standard error of the mean for three replicates.

Figure 6. a) Loss of pentachlorophenoxide in a pentachlorophenoxideactivated persulfate system at pH 8.3 (reactors: 0 or 0.5 M sodium persulfate and 1 mM pentachlorophenol at pH 8.3; 15 mL total volume). Error bars represent the standard error of the mean for three replicates; b) Structures of pentachlorophenol and two products identified by GC/MS.

reactivity of hydroperoxide with persulfate (reaction 3),8 any hydroperoxide generated through reaction 6 will not accumulate in measurable concentrations in reactions conducted at high pH. Therefore, reactions were conducted at pH 8.3; any hydroperoxide that is generated at this pH regime would exist only as hydrogen peroxide, which is not reactive with persulfate,8 and would therefore accumulate. The results of Figure 5 show that in the pentachlorophenoxide−persulfate system at pH 8.3, the hydroxyl radical probe nitrobenzene was oxidized while the hydrogen peroxide concentration was undetectable (detection limit: 0.0015 mM), demonstrating that hydroperoxide/hydrogen peroxide was not generated. These results are consistent with the reductive mechanism of reaction 5: sulfate radical and its propagation product hydroxyl radical would be generated and the probe compound nitrobenzene would be degraded, but no hydrogen peroxide

Pentachlorophenoxide was degraded by 69% over 5 h in the system containing persulfate, while no pentachlorophenoxide loss was observed in the control system with no persulfate. These results demonstrate that phenoxide was degraded during its activation of persulfate. Degradation products of pentachlorophenoxide during its reductive activation of persulfate were identified by GC/MS analysis compared to authentic standards. During the activation of persulfate, pentachlorophenoxide was transformed to hydroquinone byproducts, which are the documented oxidation products of pentachlorophenol;22 the compounds identified were 2,3,5,6-tetrachloro-1,4-benzenediol and 3,4,6-trichloro-1,2-benzenediol (Figure 6b). The F

dx.doi.org/10.1021/es400728c | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

initial pentachlorophenoxide concentration (Figure 7b). Rates of pentachlorophenoxide activation of persulfate increased nearly linearly up to a concentration of 1 mM pentachlorophenoxide and were zero-order above 1 mM. A small decline in hydroxyl radical activity was evident above 2 mM pentachlorophenoxide, which may be due to scavenging of hydroxyl radical by the pentachlorophenoxide. Pentachlorophenoxide is slightly more reactive with hydroxyl radical (kHO• = 5.0 × 109 M−1 s−1)25 than is nitrobenzene. The results of Figure 7 suggest that only minimal phenoxide concentrations are needed to activate persulfate. Therefore, even trace concentrations of SOM may potentially activate persulfate during persulfate ISCO. Furthermore, other hydroxylated contaminants, including contaminant degradation products, may potentially activate persulfate in the subsurface. The results of this research demonstrate that phenols, when present in their dissociated anionic form, activate persulfate. At highly basic pH regimes (pH > 10.5) both base and phenoxide activation occur simultaneously, with base activation proceeding by the recently elucidated pathway of base catalyzed hydrolysis of persulfate to generate hydroperoxide anion.8 Using the most acidic chlorophenol, pentachlorophenol (pKa = 4.8), persulfate activation by phenoxide was separated from persulfate activation by hydroperoxide (pKa = 11.8). In pentachlorophenoxide-activated persulfate reactions conducted at pH 8.3, phenoxide activation of persulfate was confirmed to proceed via reduction of persulfate by the phenoxide, rather than through nucleophilic attack by the phenoxide. The results of this research are in agreement with the findings of Ahmad et al.,9 who found that persulfate can be activated by SOM, which contains phenolic moieties. Other anionic low molecular weight organic compounds, such as keto acids, may also activate persulfate.26 Many of the common methods of activating persulfate, such as iron chelates and base, are problematic for ISCO,1 and further studies using a nontoxic organic anion may provide the basis for more effective pathways for the activation persulfate for ISCO.

hydroxyl groups of the identified diol products are ortho- and para-oriented, which is typical of electrophilic substitution phenols.23 No sulfonated phenols were detected as the reactions proceeded. The formation of oxidized products of pentachlorophenoxide provides additional evidence that the mechanism for phenoxide activation of persulfate is through a one-electron reduction of persulfate by the phenoxide, similar to initiation by a reduced metal (e.g., iron(II)). The pathway is parallel to the activation of persulfate by hydroperoxide8 as well as the initiation reaction for ozone decomposition by hydroperoxide.24 Effect of Phenoxide Concentration on Persulfate Activation. The effect of pentachlorophenoxide concentration in activating persulfate to generate hydroxyl radical is shown in Figure 7a. Nitrobenzene degradation increased with increasing concentrations of pentachlorophenoxide. Each of the nitrobenzene loss curves were fit to pseudo-first-order kinetics, and the first-order rate constants were plotted as a function of the



ASSOCIATED CONTENT

S Supporting Information *

Figure S1: Persulfate concentrations during the course of phenoxide-activated persulfate reactions (demonstrates that persulfate concentrations do not decrease significantly during the course of the reactions); Figure S2: Nitrobenzene degradation at pH regimes not shown in Figure 3 (presents data from which rates in Figure 4 were calculated); Figure S3: Scavenging of hydroxyl radical, but not sulfate radical, in a pentachlorophenoxide-activated persulfate system at pH 6.5 (demonstrates that significant levels of hydroxyl radical are generated in phenoxide-activated persulfate systems at pH < 9, in agreement with previous findings in iron-EDTA-activated persulfate systems at pH 5,4 thermally activated persulfate systems at pH 2−9,27 and mineral-activated persulfate systems at pH 4−5.28 This material is available free of charge via the Internet at http://pubs.acs.org.

Figure 7. Nitrobenzene degradation in 0.5 M persulfate systems activated by various concentrations of pentachlorophenol at pH 8.3 (reactors: 0.5 M sodium persulfate, 0−4.0 mM pentachlorophenol, and 1 mM nitrobenzene at pH 8.3; 15 mL total volume; control reactors contained water in place of pentachlorophenol and persulfate). Error bars represent the standard error of the mean for three replicates (a) and 95% confidence intervals (b). a) Nitrobenzene degradation at selected pentachlorophenol concentrations; b) Nitrobenzene degradation rate constants for all concentrations of pentachlorophenol.



AUTHOR INFORMATION

Corresponding Author

*Phone: 509-335-3761. Fax: 509-335-7632. E-mail: rjwatts@ wsu.edu. G

dx.doi.org/10.1021/es400728c | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Present Address

(20) Cohen, I. R.; Purcell, T. C. Spectrophotometric determination of hydrogen peroxide with 8-quinolinol. Anal. Chem. 1967, 39, 131− 132. (21) Glaze, W. H.; Kenneke, J. F.; Ferry, J. L. Chlorinated byproducts from the TiO2-mediated photodegradation of trichloroethylene and tetrachloroethylene in water. Environ. Sci. Technol. 1993, 27 (1), 177− 184. (22) Merz, J. H.; Waters, W. A. The oxidation of aromatic compounds by means of the free hydroxyl radicals. J. Chem. Soc. 1949, 2427−2433. (23) Metelitsa, D. I. Mechanisms of the hydroxylation of aromatic compounds. Russ. Chem. Rev. 1971, 40, 563−579. (24) Staehelin, J.; Hoigné, J. Decomposition of ozone in waterrate of initiation by hydroxide ions and hydrogen peroxide. Environ. Sci. Technol. 1982, 16, 676−681. (25) Peller, J.; Kamat, P. V. Radiolytic transformations of chlorinated phenols and chlorinated phenoxyacetic acids. J. Phys. Chem. A 2005, 109 (42), 9528−9535. (26) Ocampo, A. M. Persulfate Activation by Organic Compounds, Ph. D. Dissertation, Washington State University: Pullman, WA, 2009. (27) Liang, C. J.; Su, H.-W. Identification of sulfate and hydroxyl radicals in thermally activated persulfate. Ind. Eng. Chem. Res. 2009, 48, 5558−5562. (28) Teel, A. L.; Ahmad, M.; Watts, R. J. Persulfate activation by naturally occurring trace minerals. J. Hazard. Mater. 2011, 196, 153− 159.



ARCADIS U.S., Inc. Ten Friends Lane, Newtown, PA 18940. Phone: 267-685-1800. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Funding for this research was provided by the Strategic Environmental Research and Development Program through Project No. ER-1489.



REFERENCES

(1) Watts, R. J.; Teel, A. L. Treatment of contaminated soils and groundwater using ISCO. Pract. Period. Hazard., Toxic, Radioact. Waste Manage. 2006, 10, 2−9. (2) Liang, C. J.; Bruell, C. J.; Marley, M. C.; Sperry, K. L. Persulfate oxidation for in situ remediation of TCE. I. Activated by ferrous ion with and without a persulfate-thiosulfate redox couple. Chemosphere 2004, 55, 1213−1223. (3) Furman, O.; Teel, A. L.; Watts, R. J. Effect of basicity on persulfate reactivity. J. Environ. Eng. 2011, 137, 241−247. (4) Ahmad, M.; Teel, A. L.; Furman, O.; Watts, R. J. Oxidative and reductive pathways in iron-ethylenediaminetetraacetic acid (EDTA) activated persulfate systems. J. Environ. Eng. 2012, 138, 411−418. (5) House, D. A. Kinetics and mechanism of oxidation by peroxydisulfate. Chem. Rev. 1962, 62, 185−200. (6) Liang, C.; Guo, Y. Mass transfer and chemical oxidation of naphthalene particles with zerovalent iron activated persulfate. Environ. Sci. Technol. 2010, 44, 8203−8208. (7) Le, C.; Wu, J.-H.; Li, P.; Wang, X.; Zhu, N. W.; Wu, P. X.; Yang, B. Decolorization of anthraquinone dye Reactive Blue 19 by the combination of persulfate and zero-valent iron. Water Sci. Technol. 2011, 64, 754−759. (8) Furman, O.; Teel, A. L.; Watts, R. J. Mechanism of base activation of persulfate. Environ. Sci. Technol. 2010, 44, 6423−6428. (9) Ahmad, M.; Teel, A. L.; Watts, R. J. Persulfate activation by subsurface minerals. J. Contam. Hydrol. 2010, 115, 34−45. (10) Sposito, G. The Chemistry of Soils; Oxford University Press: New York, NY, 2008. (11) Behrman, E. J. The Elbs and Boyland-Sims peroxydisulfate oxidations. Beilstein J. Org. Chem. 2006, 2(22), doi:10.1186/18605397-2-22. (12) Watson, K. G.; Serban, A. Evaluation of the Elbs persulfate oxidation reaction for the preparation of aryloxyphenoxypropionate herbicides. Aust. J. Chem. 1995, 48, 1503−1509. (13) Herrmann, H.; Reese, A.; Zellner, R. Time-resolved UV/VIS diode array absorption spectroscopy of SOx− (x = 3, 4, 5) radical anions in aqueous solution. J. Mol. Struct. 1995, 348, 183−186. (14) Neta, P.; Madhavan, V.; Zemel, H.; Fessenden, R. W. Rate constants and mechanism of reaction of SO4.− with aromatic compounds. J. Am. Chem. Soc. 1977, 99, 163−164. (15) Buxton, G. V.; Greenstock, C. L.; Helman, W. P.; Ross, A. B. Critical review of rate constants for reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (•OH/•O−) in aqueous solution. J. Phys. Chem. Ref. Data 1988, 17, 513−780. (16) Clifton, C. L.; Huie, R. E. Rate constants for hydrogen abstraction reactions of the sulfate radical, SO4•−. Alcohols. Int. J. Chem. Kinet. 1989, 21, 677−687. (17) Haynes, W. M. CRC Handbook of Chemistry and Physics, 93rd ed.; CRC Press: Cleveland, OH, 2012. (18) Schellenberg, K.; Leuenberger, C.; Schwarzenback, R. P. Sorption of chlorinated phenols by natural sediments and aquifer materials. Environ. Sci. Technol. 1984, 18, 652−657. (19) Kolthoff, I. M.; Stenger, V. A. Volumetric Analysis, Vol. II: Titration Methods: Acid-Base, Precipitation, and Complex Reactions, 2nd revised ed.; Interscience Publishers, Inc.: New York, 1947. H

dx.doi.org/10.1021/es400728c | Environ. Sci. Technol. XXXX, XXX, XXX−XXX