Thermally activated persulfate oxidation of NAPL chlorinated organic compounds: effect of soil composition on oxidant demand in different soil-persulfate systems

Chlorinated organic compounds such as trichloroethylene (TCE) and 1,1,1-trichloroethane (1,1,1-TCA) can be commonly detected in contaminated field sites, posing a serious threat to ecological and human health due to their toxicity and carcinogenicity (Liang et al. 2003, 2007; Ko et al. 2012). Due to their high density and low water solubility, these substances tend to sink down through soils into the water table, potentially causing free non-aqueous phase liquid (NAPL) contamination. Once the subsurface media is contaminated by NAPL chlorinated organic compounds, it can be difficult to reduce concentrations back to a safe level using natural attenuation.

In-situ chemical oxidation (ISCO) exhibits great potential for the remediation of contaminated sites, a process that involves the insertion of strong oxidants into the subsurface media (Yan & Lo 2013; Hu et al. 2015; Peluffo et al. 2016). Recently, persulfate (S₂O₈²⁻, PS) has emerged as a useful strong oxidant (E₀ = 2.01 V) for in-situ soil remediation (An et al. 2015; Girit et al. 2015; Hilles et al. 2016). Persulfate can be activated by heat, metal or UV irradiation to generate sulfate radicals (⁠⁠) with a stronger oxidation potential (E₀ = 2.6 V) than S₂O₈²⁻ (Le et al. 2011; Ko et al. 2012; Yan & Lo 2013; Ahmadi et al. 2015; Zhou et al. 2016). Thermal activation has been established as a highly effective method to activate persulfate for the degradation of organic contaminants such as methyl tert-butyl ether, TCE, tetracycline and polycyclic aromatic hydrocarbons (Huang et al. 2002; Liang et al. 2003; Eslami et al. 2016; Peluffo et al. 2016).

The removal of organic contaminants from soil using thermally activated persulfate oxidation has been investigated in previous literature (Haselow et al. 2003; Liang et al. 2003, 2008b; Deng et al. 2013), although many of these studies focus on the treatment of single contaminants despite a high percentage of contaminated sites being characterized as complex mixtures, sometimes with the presence of NAPL contamination. The presence of contaminant mixtures in soils makes remediation more complex, as processes involve not only interactions between oxidants and the subsurface material, but also competition between contaminants for oxidants. Some previous studies have reported that removal rates for target contaminants were often remarkably reduced using activated persulfate in soil systems, compared to liquid based systems (Liang et al. 2003; Crimi & Taylor 2007). These studies suggest that the removal of target contaminants might be limited by oxidant consumption by a variety of non-target substances associated with subsurface matter, such as natural minerals (e.g., iron and manganese) or organic substances (soil organic carbon, f_oc) (Huang et al. 2002; Haselow et al. 2003; Liang et al. 2007, 2008b; Deng et al. 2013). Therefore, the consumption of oxidants in oxidation reactions with soil components must be fully characterized and quantified to understand the impact on oxidant availability for target contaminants, as well as to establish the limitations of in-situ remediation systems.

Much previous literature focusses on the levels of oxidant consumed by contaminants rather than the contribution of soil composition when designing field scale remediation, therefore this study investigates the interaction of persulfate with target contaminants as well as soil components. The purpose of this study is to establish the influence of soil composition on the consumption of persulfate and the removal of NAPL chlorinated organic compounds by thermally activated persulfate oxidation (e.g. 30 °C and 50 °C), in different soil types with varying mineral and f_oc content (sand soil, paddy soil and clay soil). The main objectives are: (1) to determine the rates of chlorinated organic compound oxidation and persulfate consumption in the varying soil-water systems; (2) to evaluate the competition between chlorinated organic compounds and soil components for oxidation and hence persulfate consumption; (3) to quantify the level of persulfate consumption by soil components to determine the relative importance of mineral and f_oc content on the rate of persulfate consumption.

Sodium persulfate (Na₂S₂O₈, >99.7%) and methanol (CH₃OH, >99.5%) were purchased from Sinopharm Chemical Reagent Co. Ltd (Shanghai, China). TCE (≥97%), cis-1,2-dichloroethene (cis-DCE, ≥97%), 1,1,1-trichloroethane (1,1,1-TCA, ≥97%) and 1,2-dichloroethane (1,2-DCA, ≥98%) were purchased from Sigma-Aldrich Chemical Co. Ltd (USA). Water was purified using a Millipore reverse osmosis purification system.

Three types of soils: high sand content, paddy field and high clay content soil were applied in this study. The sand soil-type and paddy field soil-type were obtained from a depth of 10 cm, from local farmland in southern Changchun, China. The clay soil-type was obtained from a depth of 3 m below the ground surface (Changchun, China). The soil samples were air-dried, then ground by hand and sieved through sieve #35 (0.5 mm) prior to use. The characteristics of the various soil materials used in this study are presented in Table 1.

All experiments were performed in triplicate in a series of 35 mL brown glass bottles equipped with Agilent caps and polytetrafluoroethylene liners. The solid-water ratio (m/v) of 1:2 was used with 15 g of dry soil and 30 mL of aqueous solution. 90 μL of mixed NAPLs containing TCE, 1,1,1-TCA, cis-DCE and 1,2-DCA (volume ratio, 1/1/1/1) were syringe injected into the test reactors, filling up reaction bottles completely to allow for no air bubbles. All the reaction bottles were shaken continuously, using a ZWY-240 thermostatic reciprocating shaker at 100 rpm at either 30 °C or 50 °C. Residual contaminants present in the soil slurry system were determined following methanol extraction. The extraction vials were shaken for 48 h using a reciprocating shaker, as specified previously. Following extraction, the mixtures were centrifuged (2,000 rpm, 5 min), with the supernatant collected for the quantification of the concentrations of four chlorinated volatile organic compounds (CVOCs), TCE; 1,1,1-TCA; cis-DCE; and 1,2-DCA. Control tests were performed in triplicate, assessing CVOC concentrations in the absence of persulfate.

Persulfate concentration was determined using UV spectrophotometry with potassium iodide (Liang et al. 2008a). The four representative CVOCs were quantified using an Agilent 7890A gas chromatograph equipped with a HP-5MS capillary column (30 m × 0.25 mm × 0.25 μm), using a micro-electron capture detector. Nitrogen was used as the carrier gas with a constant flow rate of 50 mL min⁻¹, with column, injector and detector temperatures of 40 °C (isothermal), 150 °C and 250 °C, respectively. The pH of soil samples was determined using a PHS-25C pH meter (NY-T 1377-2007, China). Soil organic matter mass was established based on the weight fraction of organic carbon in the soil (f_oc: with the units g^C per g^S, where g^C is the number of grams of carbon and g^S is the number of grams of soil). Total organic carbon (TOC, f_oc) of soil samples was measured using a TOC analyzer (Shimadzu SSM-5000A). Mineral content was quantified using X-ray fluorescence spectrometry (Shimadzu XRF-1800). Particle size distribution and specific surface area (SSA) of soil samples were characterized using a laser particle size distribution analyzer (Bettersize 2000).

Increased temperature conditions from 30 °C to 50 °C resulted in an increased overall treatment efficiency for all CVOCs assessed (Figure 2), as well as a greater level of consumption of persulfate (Figure 3). Treatment at 30 °C resulted in effective yet incomplete removal of the four CVOCs (50–66% removal, as shown in Figure 2), with lower levels of persulfate consumption (35–36% consumed, as shown in Figure 3). In the case of sand-soil-persulfate systems, when temperature conditions increased from 30 °C to 50 °C, an increase in persulfate consumption was also observed from 35% to 82%, resulting in an increased level of removal of total CVOCs from 50% to 78%, consistent with the findings of previous studies, establishing that higher levels of persulfate consumption result in a higher degree of chlorinated organic removal in aqueous systems (Gu et al. 2012; Deng et al. 2013). In comparison to the sandy-soil-persulfate system, the paddy-soil-persulfate system consumed almost equal amounts of persulfate (ranging from 35% to 84% depending upon other conditions), but achieved a lower degree of CVOC removal (from 60% to 80%) with an increase in operational temperature from 30 °C to 50 °C. The same case was found for clay-soil-persulfate systems, with persulfate consumption of between 36–76% and total CVOC removal of 66–87% at 30 °C–50 °C. The occurrence of significantly increased persulfate consumption with only limited removal of CVOC species, was observed with elevated reaction temperatures in paddy- or clay-soil-persulfate system, showing that the consumption of persulfate was not only due to CVOC oxidation.

The consumption of persulfate during the reaction with soil components is referred to as the natural oxidant demand (NOD) and is expressed in g Na₂S₂O₈/kg dry weight soil (dw). In situations where contaminants are present, the term TOD (total oxidant demand) is used, which is composed of NOD combined with contaminant demand (Haselow et al. 2003). Comparison of TOD values and NOD values shown in Table 2 highlights that the reactions which drive NOD occur in parallel with the reaction between soil contaminants and persulfate. The loss of natural soil minerals or f_oc content after an 11 d period in the soil-persulfate system in the absence of CVOCs, suggests that the NOD is composed of several reactions, including the reactions of persulfate with soil organics, minerals and other components. There is only a minor difference between the TOD and NOD values in sand-, paddy- or clay-soil-persulfate systems, suggesting that the reaction of soil components with persulfate (NOD), exerted a large relative contribution to the TOD. Therefore, the reactions of soil components with persulfate had a significant impact on the available oxidant demand and must be taken into account when estimating the oxidation efficiency of in-situ oxidation remediation systems.

For soil-persulfate systems in the absence of CVOCs, NOD values in paddy-field soil were the closest to TOD at 30 °C, further increasing at 50 °C (Table 2), suggesting the available oxidant demand in paddy-field soil was affected significantly by NOD, followed by sand soil types and finally clay soil types. The most significant decrease in OE values with increased temperature from 30 °C to 50 °C occurred in the paddy-soil-persulfate system (Figure 4), which also indicates that soil composition had the most significant effect on treatment efficiency in the paddy-soil-persulfate system. In comparison with the sand and clay-soil-persulfate systems, the paddy-soil-persulfate system had the most notable loss of f_oc and the lowest reduction in natural minerals at either 30 °C or 50 °C (Table 2). This phenomenon suggests that the most important factor affecting the available oxidant demand in paddy-field soil was natural organic species (f_oc) rather than natural mineral content. Moreover, as the temperature increased from 30 °C to 50 °C, the greatest increase in NOD (34.2 g Na₂S₂O₈/kg dw) and Δf_oc (1.03 g/kg dw) was observed, along with the lowest measured increase in CVOC species removal (21%) in paddy-field soil. These results suggest that natural organic species contribute to the additional consumption of persulfate, without any increase in the rate of CVOC species removal in paddy-field soil. Conversely, for the clay-soil-persulfate system, the maximal loss of natural minerals (20.40 g/kg dw at 30 °C and 21.17 g/kg dw at 50 °C, Table 2), in combination with the lowest observed NOD values (11.4 g Na₂S₂O₈/kg dw at 30 °C and 40.2 g Na₂S₂O₈/kg dw at 50 °C, Table 2) resulted in a highly efficient removal rate for the CVOC species assessed (65% at 30 °C and 87% at 50 °C, Table 2) with also the highest OE values observed (0.066 at 30 °C and 0.041 at 50 °C, Figure 4). These results show that the natural mineral content contributed significantly to CVOC removal in clay soil types. For the sand-soil-persulfate system, as the temperature increased from 30 °C to 50 °C, the significant increase in mineral content (2.66 g/kg dw, Table 2) and CVOC species removal (28%, Table 2) suggests that the changes in mineral content had a positive effect on the removal efficiency of CVOCs in sand soil types.

The mass removal of four individual CVOCs, using thermally activated persulfate oxidation, decreased in efficiency in the order: cis-DCE > TCE > 1,2-DCA > 1,1,1-TCA. Persulfate was consumed gradually over time in all soils and elevated reaction temperature resulted in an increase of persulfate consumption and CVOC removal, but no improvement of oxidation efficiency. The degree of persulfate consumption was found to not only be driven by CVOC species oxidation in soils, but also by various interactions between persulfate and soil components, leading to an increased NOD. The reaction of persulfate with soil components had varying effects on the available oxidant demand and rate of CVOC removal, depending on specific soil type composition of minerals and f_oc content. NOD exerted a large relative contribution to the TOD; therefore, the reactions of soil composition with persulfate must be taken into account when estimating the oxidation efficiency of in-situ oxidation remediation systems.

The present work was funded by National Natural Science Foundation of China (Grant No. 41472214).