Unwanted Loss of Volatile Organic Compounds (VOCs) During in Situ Chemical Oxidation Sample Preservation: Mechanisms and Solutions

Abstract

To assess the performance of hazardous waste sites remediation technologies, like in situ chemical oxidation (ISCO) with persulfate (${\text{S}}_2{\text{O}}_8^{2-}$) researchers must periodically measure concentrations of target contaminants. Due to the presence of relatively high concentrations of the residual oxidant expected in many samples, the standard analytical method requires the addition of a relatively high concentration of ascorbic acid to prevent the oxidation process from continuing after sample collection. We discovered that addition of ascorbic acid quencher results in a radical chain reaction that transforms two common halogenated solvents (i.e., tetrachloroethene and hexachloroethane). To avoid the artifact associated with the radical chain reaction, a small quantity of n-hexane can be added to aqueous samples to extract target compounds and protect them from the radical chain reaction initiated by addition of the quencher. We recommend the use of this alternative sample preservation method whenever high concentrations of residual ${\text{S}}_2{\text{O}}_8^{2-}$ are expected to be present in water samples that are contaminated with halogenated solvents.

1. Introduction

In situ chemical oxidation (ISCO) is employed for the remediation soil, aquifer solids and groundwater contaminated with a variety of organic contaminants, including halogenated solvents such as tetrachloroethene (C₂Cl₄; PCE) (U.S. EPA, 2020; Siegrist et al., 2011). In one of the most popular applications of ISCO, the oxidant solution consists of a high concentration of a sodium or potassium salt of persulfate (${\text{S}}_2{\text{O}}_8^{2-}$), with concentrations ranging from about 400 to 1000 mM. In this process, aquifer heating or the addition of a strong base is used to convert ${\text{S}}_2{\text{O}}_8^{2-}$ into sulfate radical (${\text{SO}}_4^{•-}$) and hydroxyl radical (^•OH) (Tsitonaki et al., 2010; Lee et al., 2020). To assess the progress of the remediation process, samples are routinely collected and analyzed for target compounds. During the period when oxidants are being applied, relatively high concentrations of residual ${\text{S}}_2{\text{O}}_8^{2-}$ is often present in the samples. To minimize the potential for continued oxidation of analytes after sample collection, a chemical that is readily oxidized (i.e., a quencher) is often added to the after sample collection, filtration, and residual oxidant measurement (Ko et al., 2012).

The U.S. EPA guideline (EPA/600/R-12/049) that is recommended for analysis of volatile organic compounds (VOCs) includes the addition of ascorbic acid (C₆H₈O₆) as a quencher (Ko et al., 2012). At circumneutral pH values, the deprotonated form of ascorbic acid, ascorbate (${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$) readily undergoes a one-electron transfer reaction with most oxidants (Buettner and Jurkiewicz, 1993; Iyanagi et al., 1985). This guideline has been widely used over a decade and has been extensively utilized for the removal of residual ${\text{S}}_2{\text{O}}_8^{2-}$ in both real-world ISCO site sample preservation and laboratory-scale ISCO experiments. However, while we were conducting laboratory experiments on tetrachloroethene and hexachloroethane (C₂Cl₆; HCA) remediation under realistic ISCO site conditions (i.e., $\left[{\text{S}}_2{\text{O}}_8^{2-}\right]0$= 450 mM), we observed artifacts when the ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$ used as a quencher for ${\text{S}}_2{\text{O}}_8^{2-}$.

We observed prompt loss of VOCs when the ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$ was applied as a quencher in our experimental setup. To address the phenomenon, the following hypothesis is proposed: The U.S. EPA guideline (Ko et al., 2012) for preventing the loss of VOCs after sampling is designed for cases where the initial ${\text{S}}_2{\text{O}}_8^{2-}$ concentration is 10 mM. However, samples from actual ISCO sites often show much higher ${\text{S}}_2{\text{O}}_8^{2-}$ concentrations because the initial concentration of ${\text{S}}_2{\text{O}}_8^{2-}$ at these sites usually ranges from 400 to 1000 mM. The interaction between specific reactants at these higher concentrations potentially leads to a chain reaction. This might result in the transformation of VOCs during sample preservation, potentially leading to an overestimation of treatment efficiency in ISCO processes.

To gain insight into the mechanism through which halogenated solvents are transformed during the quenching process, we employed tetrachloroethene and hexachloroethane as representative compounds. Aqueous samples amended with known amounts of these compounds were analyzed after quenching samples containing concentrations of ${\text{S}}_2{\text{O}}_8^{2-}$ typical of those observed at sites undergoing ISCO remediation. By analyzing the transformation of target contaminants and ${\text{S}}_2{\text{O}}_8^{2-}$ during quenching reactions, along with the consumption of dissolved O₂, changes in pH, and the formation of radical species we elucidated the process that results in unexpected transformation of halogenated contaminants during sample preservation. To minimize the potential loss of contaminants after sampling, we developed and tested an alternative sample preservation process.

2. Materials and Methods

The chemicals used in the study (Text S1) were commercially available and were purchased at their highest level of purity. Purchased chemicals were used without additional purification.

A stock solution of 1 M Na₂S₂O₈ was prepared in deionized water from a Milli-Q ultrapure water system (18.2 MΩ·cm at 25°C). Stock solutions (223.3 mM) of quenchers were prepared by dissolving C₆H₈O₆ or sodium sulfite (Na₂SO₃) in deionized water. Stock solutions of 50 mM hexachloroethane and tetrachloroethene was prepared by dissolving them in 100% ethanol. Stock solutions of 500 mM and 50 mM of NaCl and NaNO₃ were prepared in deionized water, respectively.

To investigate conditions typical of those encountered at ISCO sites, 0.015 mL of each target compound from the 50 mM stock solution was added to the 44.785 mL of reducing agent solution (223.3 mM). At this point, the pH was 6.5 ± 0.3. Next, 0.1 mL of Cl⁻ (500 mM) and NO₃⁻ from 50 mM stock solutions were added prior to adding 5 mL of Na₂S₂O₈ solution from the 1 M solution were added to the reducing agent and target compound containing solution (total volume of 50 mL). The initial concentration of each component in this solution is summarized in Table S1.

After mixing, solutions were promptly transferred to 9 mL amber glass vials without headspace and sealed with a Teflon cap as described in U.S. EPA method 5021A (U.S. EPA, 2014). Scavenging experiments were conducted at 21°C for up to 1 hour. At each time point of sampling, 1 mL sample aliquots were transferred to 2 mL vials containing 1 mL of n-hexane and capped. Liquid-liquid extraction was conducted in the vial by manual shaking for 5 sec. Capped vials were stored at 4°C until analysis, which took place within 2 weeks of sample collection.

VOCs were analyzed by a GC-MS (8890 GC and 5977B MS, Agilent, Santa Clara, CA, US). The method detection limit (MDL) for hexachloroethane, pentachloroethane (C₂HCl₅), tetrachloroethene, and trichloroethene (CHCl₃) were 36 nM, 16 nM, 41 nM, and 28 nM, respectively. A modified version of U.S. EPA 8270E GC-MS analysis method was utilized (U.S. EPA, 2014). Details of the analytical method and an example chromatogram are shown in Text S2 and Figure S1. A representative calibration curve for the target analytes is presented in Figure S2.

Sulfate (${\text{SO}}_4^{2-}$) concentration was determined by U.S. EPA method 300.1 (U.S. EPA, 1997) with ion chromatography (Dionex™ Aquion™ IC System., ThermoFischer Scientific, Waltham, MA, US) performed with an Ion Pac column (AS23, 4 × 250 mm), and a guard column (AG23, 4 × 50 mm). AS23 eluent (0.8 mM of Na₂CO₃ and 4.5 mM of NaHCO₃) was used and the column temperature was set to 30° C. The current was 25 mA and the eluent flow rate was set to 1.0 mL min⁻¹. A representative ${\text{SO}}_4^{2-}$ calibration curve is depicted in Figure S3.

${\text{S}}_2{\text{O}}_8^{2-}$ was quantified by the iodine colorimetric method (Liang et al., 2008; Liu et al., 2021). The ${\text{S}}_2{\text{O}}_8^{2-}$ containing solutions were diluted with deionized water to achieve a final concentration of 1 mM or less. After that, 0.1 mL of the diluted sample was transferred to conical tubes containing 4.9 mL of KI and NaHCO₃ solution ([KI]=0.55M and [NaHCO₃] = 76 mM). Prepared samples were stored in a dark room for 30 min at room temperature (21° C) before measuring absorbance at 352 nm with an ultraviolet-visible spectrophotometer (UV-2600i, Shimadzu, Kyoto, Japan). Due to the reduction of I₂ to I⁻ in the presence of reducing agents, it is inappropriate to utilize the ${\text{S}}_2{\text{O}}_8^{2-}$ colorimetric analysis method in the presence of bisulfite ($\text{HS}{\text{O}}_3^{-}$) or ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$. Instead, the formation of ${\text{SO}}_4^{2-}$ was used as a means of assessing ${\text{S}}_2{\text{O}}_8^{2-}$ decomposition when these reducing agents were present. Details of the method and overall stoichiometry of the reactions are summarized in Text S3. The calibration curve for ${\text{S}}_2{\text{O}}_8^{2-}$ is shown in Figure S4.

Reactive radical species were investigated using electron paramagnetic resonance (EPR) spectroscopy, employing 10 mM of 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as a spin trapping agent. A JES-X310 EPR spectrophotometer was utilized to obtain radical spin adducts (JEOL, Tokyo, Japan).

Temperature was measured with an alcohol thermometer (ThermoFisher Scientific, Waltham, MA, US). pH and dissolved O₂ were measured with an DB-10 pH meter (Denver Instruments, Bohemia, NY, US) and Orion STAR A213 meter coupled with the Orion dissolved oxygen probe (ThermoFisher Scientific, Waltham, MA, US), respectively.

3. Results and Discussion

3.1. Transformation of halogenated contaminants caused by reactions of ${\text{S}}_2{\text{O}}_8^{2-}$ with ${\text{C}}_6{\text{H}}_7{\text{O}}^{-}$ and $\text{HS}{\text{O}}_3^{-}$

As observed previously (Huling et al., 2011), the presence of a modest concentration of ${\text{S}}_2{\text{O}}_8^{2-}$ (i.e., 10 mM) relative to the amount of ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$ specified in the analytical method (i.e., 20 mM) did not result in the transformation of tetrachloroethene and hexachloroethane (Figure S5). However, under conditions representative of in-situ chemical oxidation (ISCO) (i.e., 100 mM of ${\text{S}}_2{\text{O}}_8^{2-}$), approximately 17% of the tetrachloroethene (Figure 1a) and 50% of hexachloroethane (Figure 1b) were lost at the first sampling point immediately after adding ${\text{S}}_2{\text{O}}_8^{2-}$ solution to the VOC and ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$ containing solution. Fully halogenated compounds (i.e., hexachloroethane), tend to react more quickly with reducing radicals relative to less halogenated compounds (i.e., tetrachloroethene and trichloroethene) (Kim and Sedlak, 2023).

Figure 1.

Loss of tetrachloroethane (a), hexachloroethane (b), and ${\text{S}}_2{\text{O}}_8^{2-}$ (c) during ${\text{S}}_2{\text{O}}_8^{2-}$ scavenging by ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$ or $\text{HS}{\text{O}}_3^{-}$. The control experiment was conducted without ${\text{S}}_2{\text{O}}_8^{2-}$ and reductants. Conditions: [Tetrachloroethene]₀ = 15 μM; [Hexachloroethane]₀ = 15 μM; ${[{\text{S}}_2{\text{O}}_8^{2-}]}_0$= 100 mM; ${[{\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}]}_0$= 200 mM; ${[{\text{Cl}}^{-}]}_0$= 1 mM; ${[{\text{NO}}_3^{-}]}_0$= 0.1 mM; [Ethanol]₀ = 263 μM; pH₀ = 6.5; Temp = 21±1 °C; ${[{\text{O}}_2]}_0$ = 187 μM; and n=2.

Most of the VOC loss occurred at the first sampling point (5 sec). No further significant amount of target compound loss was observed after that. This may be due to the accumulation of oxidation products of ascorbate, such as dehydroascorbic acid and oxalic acid. Results from revious studies indicate a similar termination of the radical chain reaction between ${\text{S}}_2{\text{O}}_8^{2-}$ and ethanol, which has been attributed to the accumulation of oxidative byproducts of alcohols (Beyerian and Khachatrian, 1984; Kim and Sedlak, 2023).

The reaction between ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$ and ${\text{S}}_2{\text{O}}_8^{2-}$ resulted in an immediate decrease in the concentration of ${\text{S}}_2{\text{O}}_8^{2-}$ by 70 mM, a decrease in concentration of dissolved O₂ by approximately 50 μM (Figure S6a) and a drop in pH from 6.8 to 1.9 (Figure S6b). The disappearance of halogenated compounds and consumption of oxygen upon ${\text{S}}_2{\text{O}}_8^{2-}$ decomposition were similar to results observed when ${\text{S}}_2{\text{O}}_8^{2-}$ initiated a radical chain reaction in the presence of excess ethanol (Kim and Sedlak, 2023). Therefore, we hypothesized that the one-electron oxidation of ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$ by ${\text{S}}_2{\text{O}}_8^{2-}$ and O₂ initiated a radical chain reaction (reaction 1 and 2) (Cao et al., 2019; Hou et al., 2020; Yin et al., 2022):

$$S_2{\text{O}}_8^{2-}+{\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}\to \text{S}{\text{O}}_4^{•-}+{\text{C}}_6{\text{H}}_6{\text{O}}_6^{•-}+{\text{SO}}_4^{2-}+{\text{H}}^{+}$$ (1)

$${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}+{\text{O}}_2\to {\text{C}}_6{\text{H}}_6{\text{O}}_6^{•-}+{\text{O}}_2^{•-}+{\text{H}}^{+}$$ (2)

Subsequent chain propagation reactions led to the rapid consumption of ${\text{S}}_2{\text{O}}_8^{2-}$ and release of proton, ${\text{SO}}_4^{2-}$, ${\text{SO}}_4^{•-}$, ascorbyl anion radical (${\text{C}}_6{\text{H}}_6{\text{O}}_6^{•-}$), and C₆H₆O₆ (reaction 3 and 4) (Cao et al., 2019; Hou et al., 2020):

$${\text{SO}}_4^{•-}+{\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}\to {\text{C}}_6{\text{H}}_6{\text{O}}_6^{•-}+{\text{SO}}_4^{2-}+{\text{H}}^{+}$$ (3)

$${\text{C}}_6{\text{H}}_6{\text{O}}_6^{•-}+{\text{S}}_2{\text{O}}_8^{2-}\to {\text{SO}}_4^{•-}+{\text{SO}}_4^{2-}+{\text{C}}_6{\text{H}}_6{\text{O}}_6$$ (4)

The loss of ${\text{S}}_2{\text{O}}_8^{2-}$ could not be attributed to thermolysis; the half-life of ${\text{S}}_2{\text{O}}_8^{2-}$ at this temperature (i.e., 21˚C) and pH value in the absence of ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$ is about 5 years (Johnson et al., 2008). During the chain reaction, tetrachloroethene can be oxidized by ${\text{SO}}_4^{•-}$ or reduced by the carbon-centered radical (reaction 5 and 6). Under the conditions used in the experiments, most of the ${\text{SO}}_4^{•-}$ would react with ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$. Thus, it is likely that reactions of tetrachloroethene with ${\text{C}}_6{\text{H}}_6{\text{O}}_6^{•-}$ (reaction 6) were of greater importance to loss of analyte than direct reactions with ${\text{SO}}_4^{•-}$ (reaction 5) (Kim and Sedlak, 2023):

$${\text{SO}}_4^{•-}+{\text{C}}_2{\text{Cl}}_4\to {\text{SO}}_4^{2-}+\text{product}$$ (5)

$${\text{C}}_6{\text{H}}_6{\text{O}}_6^{•-}+{\text{C}}_2{\text{Cl}}_4\to \text{producut}$$ (6)

Fully halogenated alkanes, such as hexachloroethane, are much less likely to be oxidized but can be reduced by reductive carbon-centered radicals (reaction 7) (Kim and Sedlak, 2023):

$${\text{C}}_6{\text{H}}_6{\text{O}}_6^{•-}+{\text{C}}_2{\text{Cl}}_6\to \text{product}$$ (7)

Formation of ${\text{SO}}_4^{•-}$ and ${\text{C}}_6{\text{H}}_6{\text{O}}_6^{•-}$ during the reaction between ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$ and ${\text{S}}_2{\text{O}}_8^{2-}$ was supported by previous EPR study (Hou et al., 2020).

In our previous study, we observed the transformation of hexachloroethane by alcohol carbon-centered radical, resulting in the formation of pentachloroethane and tetrachloroethene (Kim and Sedlak, 2023). However, during the ${\text{S}}_2{\text{O}}_8^{2-}$ quenching by ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$, neither product (trichloroethene MDL = 28 nM; pentachloroethane MDL = 16 nM) was detected. Therefore, we assumed that the ${\text{C}}_6{\text{H}}_6{\text{O}}_6^{•-}$ transforms hexachloroethane through a mechanism that is different from that of the alcohol carbon-centered radical. Further research is necessary to elaborate on this difference.

The proposed mechanism of the ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}/{\text{S}}_2{\text{O}}_8^{2-}$ chain reaction and VOCs transformation is summarized in Scheme 1:

Scheme 1.

Proposed radical chain reaction and transformation of tetrachloroethene and hexachloroethane initiated by the one-electron transfer between ${\text{C}}_6{\text{H}}_7{\text{O}}_6^{-}$ and ${\text{S}}_2{\text{O}}_8^{2-}$.

Persulfate-ISCO is often conducted under acidic conditions, so we focused on acidic pH conditions (pH₀=3) in our analysis of ${\text{S}}_2{\text{O}}_8^{2-}$ quenching reaction by ascorbic acid (pK_a=4.2). However, there were no significant differences in ${\text{S}}_2{\text{O}}_8^{2-}$ quenching and VOC loss between acidic (ascorbic acid) and neutral pH (ascorbate) conditions (data not shown).

As shown in Figure 1a and 1b, $\text{HS}{\text{O}}_3^{-}$, an alternative quencher, also resulted in tetrachloroethene and hexachloroethane transformation. In addition, $\text{HS}{\text{O}}_3^{-}$ only scavenged about 40% of the ${\text{S}}_2{\text{O}}_8^{2-}$ (Figure 1c). A possible explanation for the loss of halogenated contaminants when $\text{HS}{\text{O}}_3^{-}$ is used to quench ${\text{S}}_2{\text{O}}_8^{2-}$ is provided in Text S4. The addition of benzoic acid and short chain alcohols (i.e., methanol, ethanol, 2-propanol or t-butanol) to scavenge radicals produced during the quenching of ${\text{S}}_2{\text{O}}_8^{2-}$ by $\text{HS}{\text{O}}_3^{-}$ reduced the loss of analytes in the first few minutes of the reaction. However, eventually losses of up to 40% were observed (Figure S7).

3.2. Minimizing the unwanted loss of VOC during sample preservation by n-hexane extraction.

The most promising alternative approach for avoiding the loss of halogenated contaminants during sample preservation is to allow the compounds to partition into an organic solvent immediately after sample collection. Halogenated compounds, such as tetrachloroethane and hexachloroethane, readily partition into a small volume of a non-polar organic solvents, like n-hexane, while ${\text{S}}_2{\text{O}}_8^{2-}$ and other anions, remain in the aqueous phase (Oxtoby et al., 2015). Also, residual persulfate in the aqueous phase can be analyzed right after n-hexane extraction because it does not affect the residual concentration of persulfate. By adding a small amount of n-hexane (i.e., 1 mL to 1 mL of sample), no loss of halogenated analytes occurred as the VOCs were completely extracted during the 5 sec of manual shaking (Figure 2). This method eliminates the need for the addition of a quencher (e.g., ascorbate) that can contribute to loss of analytes. This approach, which can be used under field conditions, exhibits excellent recovery of target analytes (93–106%) at various initial concentration of tetrachloroethene and hexachloroethane (Table S2). Moreover, the n-hexane extracts remain stable for up to 14 days when stored at 4 °C for VOCs analysis (Figure 2), despite the presence of a small amount of headspace above the n-hexane.

Figure 2.

Stability of tetrachloroethene n-hexane extracts at 5 μM (plain bar) and 50 μM (striped bar) (a); as well as hexachloroethane n-hexane extracts at 5 μM (plain bar) and 50 μM (striped bar) (b). Conditions: [Tetrachloroethene]₀ = 5 and 50 μM; [Hexachloroethane]₀ = 5 and 50 μM; ${[{\text{S}}_2{\text{O}}_8^{2-}]}_0$= 100 mM; ${[{\text{Cl}}^{-}]}_0$= 1 mM; ${[{\text{NO}}_3^{-}]}_0$= 0.1 mM; [Ethanol]₀ = 175 (5 mM of VOCs solutions) and 877 μM (50 mM of VOCs solutions); pH₀ = 4.2; Temp = extracted at 21°C and stored at 4 °C; ${[{\text{O}}_2]}_0$ = 187 μM; and n=2. Reductants were not used.

4. Conclusions

The sample preservation approach employed in the EPA recommended guideline (EPA/600/R-12/049) and similar procedures, method preserves samples under conditions typically encountered in surface and groundwater samples. However, in the presence of relatively high concentrations of ${\text{S}}_2{\text{O}}_8^{2-}$ encountered during ISCO treatment (i.e., >100 mM) contaminants may be lost due to a chain reaction initiated when the scavenger is oxidized. By transferring a 1 mL sample of filtered groundwater sample to a 2 mL GC vial containing 1 mL of n-hexane, capping the vial shaking it for 5 sec, this artifact can be avoided. This relatively simple modification offers advantages in terms of labor, sampling time, and required ample volume. Additional research is needed to determine the potential for artifacts to occur for contaminants other than tetrachloroethene and hexachloroethane and to assess the efficacy of our proposed modification.