Article pubs.acs.org/est
Cosolvent Effects on Dechlorination of Soil-Sorbed Polychlorinated Biphenyls Using Bentonite Clay-Templated Nanoscale Zero Valent Iron Kai Yu,† G. Daniel Sheng,*,‡ and Wesley McCall§ †
Shanghai Academy of Environmental Sciences, Shanghai 200233, China State Key Laboratory of Pollution Control and Resource Reuse, College of Environmental Science and Engineering, Tongji University, Shanghai 200092, China § Geoprobe Syst Inc, Salina, Kansas 67401, United States ‡
S Supporting Information *
ABSTRACT: Zero-valent iron synthesized using bentonite clay as a template (CZVI) was tested for its reactivity toward polychlorinated biphenyl (PCB) dechlorination in soil slurries. Aqueous-phase decachlorobiphenyl (PCB209) was rapidly dechlorinated by CZVI with a reaction rate 10 times greater than that by conventional nanoscale zerovalent iron. This superior reactivity was due largely to the nanoscale size (∼0.5 nm) of the ZVI particles located in the clay galleries. In soil slurries where PCB209 was strongly soil-bound, adding ethanol as an organic cosolvent led to increased PCB209 desorption into the liquid phase, thereby enhancing the PCB209 dechlorination with CZVI. The more effective PCB209 dechlorination in such a cosolvent system also promoted the subsequent stepwise dechlorinative process, leading to a relatively more removal of chlorine in the product mixture. The dechlorination became more rapid as the ethanol fraction increased from 10% to 50%, due apparently to the increasingly greater PCB209 desorption and thus facilitated contact with CZVI. Further increase in ethanol fraction above 50% led to an insignificant enhancement in degradation rate, due partially to attenuated contact of PCB209 with CZVI and reduced proton source from limited water content in the liquid. It is suggested that addition of organic cosolvents may make CZVI potentially useful for remediation of soils containing halogenated organic contaminants.
■
INTRODUCTION Given their detrimental effects on both humans and wildlife,1,2 remediation of soils contaminated by PCBs is one of the top priorities. Due to potential risk concern and cost consideration, zerovalent iron (ZVI) has the advantage over bimetals for their remediation applications. Nanoscale ZVI (NZVI) has shown much higher reaction efficiencies than microscale ZVI (MZVI) due to its small size. This provides a great potential for remediation of soils and water contaminated with chlorinated solvents.3,4 A major obstacle for environmental remediation using NZVI particles is their strong tendency to agglomerate during preparation and application, thus substantially diminishing their efficiencies.5 To solve this problem, smectite clay was recently used as a template to synthesize a new form of ZVI (CZVI, particle sizes of ∼0.5 nm) with iron particles evenly dispersed on the smectite external and internal surfaces.6 Compared to other existing forms of ZVI, CZVI has demonstrated a superior reductive reactivity toward halogenated organic contaminants such as polybrominated diphenyl ethers.7 The fact that PCBs are sparsely water-soluble and strongly bound to soil particles creates another major challenge for applying ZVI for remediation of PCB-polluted soils. 8 © XXXX American Chemical Society
Developing a mechanism to provide sufficient contact between soil-bound PCBs and reactive iron is in need. Water-miscible organic solvents have been found to enhance the dissolution of soil-sorbed hydrophobic contaminants. Most of these solvents, with relatively low costs, do not pose a significant risk of ancillary harm to the environment.9 Research effort for such contaminated soils has accordingly been put forth by testing the combination of in situ cosolvent soil washing and subsequent chemical remediation.10 This two-phase approach uses organic cosolvents to promote the desorption of target pollutants from soil. Agents like ZVI were subsequently employed to convert the pollutants to their dechlorinated products.11,12 Although such combinations may lead to improved decontamination, the attendant drawbacks of this technique, for example, prolonged treatment times, ancillary equipment, and transportation of extracts from washing sites to treatment facilities, could complicate their implementation. Previous work aiming at improving this technique has been devoted to developing Received: June 11, 2016 Revised: November 7, 2016 Accepted: November 10, 2016
A
DOI: 10.1021/acs.est.6b02933 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology improved cosolvents or degradation agents.9−12 However, the requirement for additional installation and setup is inevitable. This is an inherent limitation of this technique for contaminant degradation that cannot be fundamentally ameliorated by developing better cosolvents or agents. This limitation is especially problematic for the decontamination of sites with limited space, for example, with buildings or other required structures on site. To the best of our knowledge, few studies have focused on employing ZVI with organic cosolvents simultaneously for ex-situ treatment of contaminated soil. This method is proposed to endow ZVI with an enhanced reactivity to destroy pollutant molecules, at lower installation and operation expenses. Numerous studies have demonstrated the effectiveness of NZVI for pollutant dechlorination in aqueous phase, yet information has been lacking pertaining to the effectiveness of degrading soil-bound pollutants with the assistance of organic cosolvents and how the addition of various concentrations of organic cosolvents to the soil slurry affects treatment efficiency. The objective of this study was therefore to evaluate the feasibility of CZVI, combined with an organic cosolvent based treatment, to achieve effective dechlorination of soil-sorbed decachlorobiphenyl (PCB209). With regard to the unique structural nature of CZVI, a related objective was to elucidate the specific role of the organic cosolvent in the CZVI mediated PCB209 degradation. The purpose of the latter objective was to advance the mechanistic understanding of cosolvents in soil− cosolvent−CZVI systems. The experiments provided useful information for the design of ex-situ treatment systems for PCB contaminated soils, with high remediation efficiency at a relatively low installation and setup cost.
PCB209 Desorption and Sorption. Five g of PCB209 spiked soil (30 mg/kg) was mixed with 50 mL of ethanol/water solution at varying ethanol-to-water ratios in a 100 mL glass vial. After agitation for 13 h (same as the degradation time) to reach equilibrium, samples were taken and PCB209 in the liquid phase was analyzed to determine the desorption of PCB209. The sorption experiment was conducted by adding 2.5 g of bentonite clay (no ZVI attached) into a 100 mL glass vial containing 50 mL of ethanol/water solution of 0.2 mg/L PCB209 at varying ethanol fractions. Likewise, after agitation for 13 h (equilibrium was reached), samples were taken and PCB209 in the liquid phase was analyzed. The change in concentration was determined to calculate the sorption of PCB209 onto “CZVI”. Analytical Methods. Details on the analytical methods are included in Supporting Information.
■
RESULTS AND DISCUSSION Characterization. The size of commercial MZVI particles was measured using SEM and ranged in diameter between 2 and 4 μm (Figure S1a). The morphology of synthesized NZVI from TEM showed that these particles existed mostly as aggregates of spherical structures averaging 20−40 nm in diameter and 0.3−2 μm in aggregate chain length, similar to the morphologies observed in previous studies (Figure S1b).13,14 A TEM image of CZVI is shown in Figure S1c. At the lower right corner the dark clusters ranging 20−50 nm may have resulted from ZVI residing on the external surfaces of clay tactoids.7 No other significant signal for iron was observed in the image. It should be noted that not only did Fe(III) cations locate on the external clay surfaces during synthesis, but an abundance of them were spatially dispersed near the negatively charged sites within the interlayer, resulting in majority of ZVIs intercalated in the gallery regions.6 The intercalated ZVI particles were apparently much smaller, probably in an amorphous form that technically cannot be detected by TEM. The XRD patterns of Na-bentonite, Fe(III)-bentonite and Fe(0)-bentonite clays, along with calculated basal spacings, are shown in Figure S2. Sodium bentonite was generally hydrated with a shell of water resulting in a basal spacing of 1.11 nm, consistent with previous research.6,7 After Na-bentonite was exposed to Fe(III) cations, the negative charges of the clay were presumably compensated by Fe(III) instead of Na+, and the basal spacing increased to 1.29 nm. The addition of NaBH4 reduced Fe(III) to Fe(0), and the cation-exchange sites on the clay were occupied by Na+ again. The XRD for Fe(0)-bentonite clay did not indicate the layer collapse back to 1.11 nm but an expansion to a basal spacing of 1.46 nm, suggesting that Na+ was not the only species remaining in the interlayer. It can be inferred that the formed ZVI clusters were intercalated in the interlayers and acted as pillars that filled the space between the clay layers. Subtracting the thickness of one bentonite layer (∼0.9 nm), the size of the intercalated ZVI was ∼0.5 nm perpendicular to the basal plane. Along with the TEM, the XRD implied that these specific ZVI clusters with small particle size were randomly distributed in the clay gallery regions. PCB209 Dechlorination in Aqueous Phase. Reaction of PCB209 (2 mg/L) in aqueous phase with MZVI (5 g/L), NZVI (5 g/L) and CZVI (1.5 g/L) was determined. Given the low solubility of PCB209 in water, ethanol was added as a cosolvent (50% by volume) to enhance the PCB209 dissolution in the liquid phase. As shown in Figure S3, no significant PCB209 degradation was observed after 4 h with MZVI,
■
EXPERIMENTAL SECTION Materials. Details on chemicals used and their suppliers, and preparation of spiked soils, are included in Supporting Information. Preparation and Characterization of ZVIs. CZVI and “conventional” NZVI were prepared using the procedures reported previously.6,7 Details on preparation and characterization of ZVIs are included in Supporting Information. Dechlorination of PCB209. Soil slurry experiments were conducted in a 100 mL glass vial. Typically, five g of PCB209spiked soil was first mixed with given amounts of different ZVIs, and then 50 mL of a deoxygenated ethanol/water mixture at varying ethanol-to-water ratios was added. The glass vials were immediately placed in an anaerobic chamber and constantly agitated in a thermostat water bath (60 °C). At predetermined time intervals, samples were taken and centrifuged at 7000 rpm for 4 min. The solid was extracted for PCB209 and its dechlorination products with a toluene/ hexane mixture (50/50, v/v) 3 times on a vortex orbital mixer. For effective extraction of PCBs in the supernatant, the solution was sacrificed first with a gentle N2 stream to remove ethanol, and then PCBs was extracted with the same method as that for solids. The extracts were combined, concentrated on a rotary evaporator (50 °C), followed by N2 blow to 1 mL for analysis. To evaluate the reactivities of different forms of ZVIs without soil or other possible matrix interference, MZVI (0.25 g), NZVI (0.25 g), or CZVI (0.075 g, in 2.5 g Fe(0)-clay) was mixed with 50 mL of ethanol/water (50/50, v/v) solution with an initial PCB209 concentration of 2 mg/L. At given sampling times, suspensions were extracted and analyzed for PCB209 and its products. B
DOI: 10.1021/acs.est.6b02933 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology whereas 28.7% and 57.4% of PCB209 were dechlorinated in 4 h with NZVI and CZVI, respectively. The reaction between PCB209 and ZVIs occurred through a surface mediated mechanism and can be written as RCl n + Fe0 + H+ → RCl n − 1H + Fe2 + + Cl−
(1)
The proton in this equation may have come from water. Since the amounts of water and ZVIs were significantly higher than PCB209 in the solution, it was plausible to assume that PCB209 dechlorination followed pseudo first-order kinetics: ⎛C⎞ ln⎜ ⎟ = kobst = k mρm t ⎝ C0 ⎠
(2)
where C (mg/L) and C0 (mg/L) refer to the PCB209 concentrations at time t (h) and initial t0, kobs is the observed rate constant, km is the specific rate constant normalized to the iron mass concentration (L/h·g), and ρm is the mass concentration of iron (g/L). The estimated specific rate constant (km) was 0.0003, 0.0168, and 0.145 (L/h·g) for PCB209 dechlorination by MZVI, NZVI, and CZVI, respectively. The low reactivity of MZVI was due mostly to its particle size being about 2 orders of magnitude larger than that of NZVI (Figure S1), as reported by Lowry et al.13 Wang and Zhang reported that NZVI (50 g/L) resulted in
