Remediation of trichloroethylene by zero-valent iron permeable reactive barriers

Transcription

1 Remediation of trichloroethylene by zero-valent iron permeable reactive barriers M. C. Zanetti, S. Fiore & G. Genon Department of Georesources and Territory, Polytechnic of Turin, Italy Abstract Trichloroethylene (TCE), belonging to the class of Dense Non-Aqueous Phase Liquids (DNAPLs), is a contaminant which is very persistent and widespread in groundwater because of its massive industrial use in past decades (i.e. in the metallurgical, textile and dye production industries). Among the in situ remediation technologies for the recovery of groundwaters polluted by chlorinated solvents, zero-valent iron Permeable Reactive Barriers (PRBs) have a primary importance; this technique, developed in Canada in the nineties, is based on the reductive action of zero-valent iron in a dechlorination process. The PRB is set perpendicularly to the groundwater flow direction, and chlorinated solvents are converted to non-toxic products, such as alkanes and alkenes, through various reaction pathways. In this work a zero-valent iron PRB is proposed as a cleanup methodology for the polluted groundwater of a site near Turin (Northern Italy), used in the past as an industrial landfill for the waste coming from a cast iron foundry. A zero-valent iron (Connelly iron, purchased from Environmental Technologies Inc., Canada), after characterization by means of particle-size and chemical analyses, was evaluated as reactive material. A leaching test, to verify the environmental impact of the material on the groundwater, was also performed on Connelly iron. Batch and column laboratory tests were performed using at first distilled water, and then an aqueous phase with a chemical composition similar to the polluted groundwater. The degradation mechanisms of TCE, hypothesizing a first order kinetic, were discussed, and the values of the kinetic constant and the necessary residence time in the PRB (the period necessary to lower the pollutant concentration below the Italian law limits for groundwater) obtained from the batch and column tests were compared and discussed. Keywords: zero-valent iron, permeable reactive barrier, perchloroethylene, trichloroethylene, chlorinated solvent, DNAPL, in situ remediation, groundwater.

2 126 Brownfield Sites II 1 Introduction Halogenated volatile organic compounds (VOCs), including chlorinated aliphatic hydrocarbons, due to their massive industrial use are the most frequently occurring type of contaminant in soil and groundwater in hazardous waste sites in the United States [1] and all over the world. The major environmental releases of trichloroethylene (TCE), a typical Dense Non-Aqueous Phase Liquid (DNAPL), are due to wastewater from metal finishing and degreasing plants, paint and ink formulation, electrical/electronic components, and rubber processing industries. These releases were primarily from steel pipe and tube manufacturing industries [2]. Among the in situ remediation technologies for the recovery of groundwater polluted by chlorinated organics, zero-valent iron Permeable Reactive Barriers (PRBs) have a primary importance. A PRB is essentially made of a reactive wall, which is set perpendicularly to the groundwater flow direction. The reactive material, characterized by a high permeability, may perform its remediation action by means of chemical, sorptive or biological processes. Zero-valent iron PRBs, developed in Canada in the Nineties, are based on a chemical process of dechlorination: zero-valent iron plays a reductive action on chlorinated solvents, which are converted by means of hydrolysis reactions to non toxic products (such as alkanes, alkenes and chloride ions), according to reaction (1), where RCl is a chlorinated compound and RH is the correspondent alkane/alkene [3-5]. Fe 0 + RCl + H + Fe 2+ + RH + Cl - (1) The above cited dechlorination process may be described by means of a first order kinetic model [3, 5, 6], expressed in equation (2), where C 0 and C are the initial/final concentration values of the pollutant and k is the first order kinetic constant. C kt = e (2) C 0 The degradation of TCE may follow two different pathways: sequential hydrogenolysis, through cis 1,2-dichloroethene (cis 1,2-DCE), vinyl chloride, ethene and ethane, and β-elimination, through chloroacethylene, acethylene, alkanes and alkenes. Some researchers proved that the fastest and predominant mechanism is the β-elimination [7, 8]. The most common configurations of zero-valent iron PRBs are the continuous barrier, for small plumes, and funnel and gate, for larger polluted areas. The maximum depth of remediation in the aquifer by means of a PRB is equal to about m [9]. The groundwater composition has a strong influence on the performances of a PRB, playing a corrosive action on the reagent material: actually water itself, dissolved oxygen, sulphates and bicarbonates are able to oxidize zero-valent iron increasing aqueous phase ph [5, 10]. Moreover, a zero-valent iron PRB modifies groundwater chemistry because of the decholorination process, increasing ph

3 Brownfield Sites II 127 and chloride concentration, decreasing redox potential to negative values, typical of reducing conditions, and eventually introducing some ionic species, as a consequence of the reactive material oxidation. A correct PRB dimensioning takes place after a complete characterization of the polluted site (aquifer physical and geological characteristics, water geochemistry, pollutant plume characteristics), and a careful evaluation of the pollutant degradation kinetics considering different zero-valent materials by means of batch and column tests [11, 12]. In this work a zero-valent iron PRB is proposed as a cleanup methodology for the polluted groundwater of a site near Turin (Northern Italy) The authors analyzed the groundwater of the considered site to determine its geochemistry. A zero-valent iron material (Connelly iron, purchased from Environmental Technologies Inc., Canada), after a characterization by means of particle-size and chemical analyses, was evaluated as reactive medium by means of batch and column tests. The pollutant concentration was monitored and the degradation mechanisms of TCE were discussed. Batch and column laboratory tests were performed using at first distilled water, and then an aqueous phase with a chemical composition similar to the polluted groundwater (named artificial groundwater), to evaluate the influence of the groundwater geochemistry on the reactivity and stability of the reagent material. A leaching test, to verify the environmental impact of the material on the groundwater, was also performed on Connelly iron. 2 The polluted site The considered site, located about 25 km from Turin (Northern Italy), belongs to a cast iron foundry and in the past was used as an industrial landfill. The main pollutant detected in soil and groundwater is TCE, with a maximum concentration equal to about 150 ppb in groundwater (Italian regulation foresees a limit equal to 1.5 ppb), and a maximum concentration in the soil, equal to 0.07 mg/kg, that is widely below Italian law limit (10 mg/kg). The data obtained from the groundwater analysis, performed by the authors, are reported in Table 1. These results show a relevant presence of dissolved substances (especially calcium, magnesium, sulphates and chlorides) that enhances the consideration about the interaction of these ions and the zero-valent iron material in the present study. 3 The zero-valent iron material The authors characterized a zero-valent iron material (Connelly iron, purchased from Environmental Technologies Inc., Canada) by means of particle-size and chemical analyses and a leaching test. A visual exam of Connelly iron showed the presence of chips, shavings, spherical particles and dust. The particle-size analysis outlined that about the 80% b.w. of the material has a size between 0.3 and 1.4 mm, and that about the 86% of the material has dimensions below 1.4 mm. The chemical analyses

4 128 Brownfield Sites II revealed a composition typical of a cast iron: 91% b.w. of iron, 2.7% b.w. of carbon, 0.4% b.w. of manganese and minor contents of chromium, copper, nickel, lead and zinc. Table 1: Physico-chemical analysis of the polluted groundwater. ph 7.3 Electric conductivity (µs/cm) 1000 Dissolved oxygen (mg/l) 0.6 Alkalinity (mg CaCO 3 /l) 38.5 Total hardness ( F) 75 Na (mg/l) 27.5 K (mg/l) 9.3 Ca (mg/l) Mg (mg/l) 32.4 Fe (mg/l) 0.45 Mn (mg/l) 0.64 Zn (mg/l) <0.1 = SO 4 (mg/l) Cl - (mg/l) NO 3 (mg/l) 0.49 NO - 2 (mg/l) <0.03 NH + 4 (mg/l) 0.29 The leaching test, performed in distilled water for a period equal to 16 days (Italian law leaching test) to evaluate the environmental impact of the material on groundwater, produced the following results: the amount of heavy metals (Cd, Co, Cr, Ni, Pb, Cu, V, Zn) transferred to the aqueous phase is neglectable; the ph value is constant and about equal to 6.4 during all the considered period, and it doesn t increase as a consequence of iron oxidation by water and dissolved oxygen, indicating that the material has probably been activated with hydrochloric acid to increase its reactivity [13]. 4 Experimental Batch and column tests were performed on Connelly iron and a polluted aqueous phase containing 500 ppb of TCE. The aqueous phase, having a ph equal to 6.8, was made of distilled water in the first batch and column tests, and then of an aqueous solution of NaCl and MgSO 4. 7H 2 O, to obtain a chemical composition almost equal to the polluted groundwater (see Table 1), named artificial groundwater. The batch tests were performed in anaerobic conditions in 20 ml glass headspace vials, using a solid/liquid ratio equal to 1:4 and considering agitation times varying between 2 hours and 13 days (2h, 6h, 1d, 2d, 3d, 5d, 7d, 10d, 13d). For each period three samples and two blanks were analyzed. The solid/liquid contact was granted by means of a STR 4/1 Stuart Scientific rotator drive rotating at 10 rpm.

5 Brownfield Sites II 129 Several works [3, 5, 11, 12] showed that batch tests are valuable for a comparative analysis of the employed reactive materials but not for a correct PRB dimensioning because the adopted operative conditions are very different from the real ones. The column tests were performed by means of a Plexiglas leaching column, characterized by the following dimensions: 1 m length, m external diameter, m internal diameter. The column was filled, starting from the bottom to the top, with m of sand, m of Connelly iron and m of sand. The liquid phase was fed to the column from the bottom to the top by means of a Tedlar bag (volume equal to 87.5 l) connected to a peristaltic pump, adopting a flow rate equal to 1 ml/min in order to simulate the groundwater flow conditions of the contaminated site. The calculated solid/liquid ratio in the column was equal to 4:1. The sampling operation was performed at 9 sampling ports (from the bottom to the top of the column, the distances were equal to 0, 0.025, 0.075, 0.125, 0.200, 0.275, 0.350, and m), by means of a glass syringe, after 5 weeks from the beginning of the leaching for the test in distilled water, and after 4 weeks for the test in artificial groundwater (stationary conditions were achieved, as verified in a previous study [12]). The collected volume for each sample was equal to 5 ml. The TCE and 1,2-DCE chemical analyses were performed using a static headspace method, by means of a Hewlett Packard gaschromatograph GCD1800C, equipped with a mass spectrometer detector and a J&W DB-624 column (30mx0.25mmx1.25 µm). The concentration of some metals (Na, K, Ca, Mg, Fe, Co, Mn, Zn, Cd, Cr, Pb, Ni) were analysed by means of a Perkin Elmer ICP OES Optima 2000DV. The chlorides, sulphates, nitrates, nitrites and ammonium contents were determined by means of standard methods [14], employing a UV-Visible spectrophotometer Unicam HEλIOS α. 5 Results The results of the batch tests performed in distilled water and in artificial groundwater allowed to determine the kinetic constant values reported in Table 2. The TCE and 1,2-DCE concentration values obtained in the two series of batch tests are shown in Figure 1 and the obtained kinetic parameters are compared in Table 2. The results of the column tests, shown in Figure 2, allowed the evaluation of the kinetic constant and of the required residence time in the PRB (the period necessary to lower the pollutant concentration under the Italian law limits for groundwater, equal to 1.5 ppb). The hydrodynamic dispersion coefficient and the effective porosity values, used in the calculation of the kinetic constant, were gathered by means of some tracer tests previously performed (η eff =0.59, D L = m 2 /s) [12]. The obtained kinetic parameters are compared in Table 2. To verify the possible hydrochloric acid treatment on Connelly iron, hypothesized considering the results of the leaching test (see section 3), the ph

6 130 Brownfield Sites II values were constantly monitored in all samples during batch and column tests (see Figure 3). distilled water C (ppb) TCE 1,2-DCE t (h) artificial groundwater C (ppb) TCE 1,2-DCE t (h) Figure 1: TCE and 1,2-DCE concentration values obtained from batch tests. Table 2: Kinetic parameters obtained from batch and column tests (k: first order kinetic constant, t: permanence time in the PRB). Batch tests Column tests Distilled water Artificial groundwater Distilled water Artificial groundwater k (s -1 ) t (min) / / 33 3 The initial ph decreasing, evident in batch tests, doesn t happen in the column test performed with artificial groundwater: in fact the presence of dissolved substances increases the ph, due to their degradative action on zero-valent iron [5, 10], and after about 4 hours the ph stabilizes near neutral values.

7 Brownfield Sites II 131 distilled water TCE 1,2 DCE C (ppb) t (h) artificial groundwater C (ppb) t (h) 6 7 TCE 8 1,2 DCE 9 Figure 2: TCE and 1,2-DCE concentration values obtained from column tests (numbers 1-9 indicate the sampling ports). 6 Conclusions On the grounds of the results obtained from the batch and column tests performed on Connelly iron with the described aqueous phases (distilled water and artificial groundwater), the following considerations can be formulated: - the kinetic constant values obtained from batch tests are in agreement with literature data (k= ) [3, 15, 16]. The values obtained from column tests are noticeably higher than the ones gathered from batch tests: obviously the two data sets can t be compared, because of the different experimental procedure (discontinous/continous reactor), but the results of

8 132 Brownfield Sites II column tests, deriving from conditions more similar to the ones happening in a PRB, appear more reliable; 7,5 batch tests 7 6,5 ph 6 5,5 5 distilled water artificial groundwater 4, t (h) ,5 column tests 7 6,5 ph 6 5,5 5 distilled water artificial groundwater 4, t (h) Figure 3: ph values measured during batch and column tests. - in batch tests the aqueous phase composition didn t have any influence on Connelly iron degradative efficiency; on the other hand the kinetic constant values obtained from column tests depend strongly on the aqueous phase composition: in fact the presence of dissolved substances accelerates the TCE degradation; - considering TCE and 1,2-DCE concentration values obtained from batch and column tests, β-elimination seems the main degradation mechanism, according to literature data [7, 8]. In fact TCE degradation was in all cases very quick at the beginning of the tests, and 1,2-DCE concentration values,

9 Brownfield Sites II 133 deriving from sequential hydrogenolysis, always became relevant after a longer period. It s noticeable that 1,2-DCE concentration values were in all cases widely below Italian law limit (60 ppb) at the end of the tests; - the hypothesized pre-treatment with hydrochloric acid (see section 3) of the reagent medium was confirmed by the ph values measured during the performed tests; - about the impact of the zero-valent material on groundwater composition, besides the results of the leaching test (see section 3), the chlorides input due to TCE degradation must be taken in account. Considering the complete dechlorination of 500 ppb of TCE, a chloride concentration equal to 0.4 mg/l will be created, and it doesn t play a relevant influence on the studied aquifer ([Cl - ]=35.1 mg/l, as shown in Table 1). In conclusion Connelly iron showed a good application to TCE remediation by means of a high degradative efficiency in column tests (the most similar to real conditions), influenced by the considered concentration of dissolved salts, and a negligible impact on groundwater composition, therefore confirming its reliability in PRBs installation for the remediation of chlorinated solvents. References [1] United States Environmental Protection Agency (US EPA), Groundwater cleanup: overview of operating experience at 28 sites, EPA 542-R , [2] United States Environmental Protection Agency (US EPA), Federal remediation technologies roundtable, EPA 542-R , [3] Gillham, R.W., O Hannesin, S.F., Enhanced degradation of halogenated aliphatics by zero-valent iron, Ground Water 32 (6), pp , [4] Helland, B.R., Alvarez, P.J.J., Schnoor, J.L., Reductive dechlorination of carbon tetrachloride with elemental iron, Journal of Hazardous Materials 4I, pp , [5] Matheson, L.J., Tratnyek, P.G., Reductive dehalogenation of chlorinated methanes by iron metal, Environmental Science and Technology 28 (12), pp , [6] Scherer, M.M., Balko, B.A., Gallagher, D.A., Tratnyek, P.G., Correlation analysis of rate constants for dechlorination by zero-valent iron, Environmental Science and Technology 32 (19), pp , [7] Roberts, A.L., Totten, L.A., Arnold, W.A., Burris, D.R., Campbell, T.J., Reductive elimination of chlorinated ethylenes by zero-valent metals, Environmental Science and Technology 30 (8), pp , [8] Campbell, T.J., Burris, D.R., Roberts, A.L., Wells, J.R., Trichloroethylene and tetrachloroethylene reduction in metallic iron-water-vapour batch system, Environ. Toxicol. Chem. 16 (4), pp , [9] Gavaskar A. R., et. al., Permeable barriers for groundwater remediation. Design, construction and monitoring, Battelle Press, 1998.

10 134 Brownfield Sites II [10] Agrawal, A., Ferguson, W.J., Gardner, B.O., Christ, J.A., Bandstra, J.Z., Tratnyek, P.G., Effects of carbonate species on the kinetics of dechlorination of 1,1,1-trichloroethane by zero-valent iron, Environmental Science and Technology 36 (20), pp , [11] Poulsen, M.M., Kueper, B.H., A field experiment to study the behaviour of tetrachlorothylene in unsaturated porous media, Environmental Science and Technology 26 (5), pp , [12] Zanetti, M.C., Fiore, S., Sethi, R., Genon, G., Di Molfetta, A., In-situ remediation by means of zero-valent iron reactive barriers: laboratory tests, Proceedings of the IX International Waste Management and Landfill Symposium Sardinia, eds., pp., [13] Su, C., Puls, R.W., Kinetics of trichloroethylene reduction by zerovalent iron and tin: pre-treatment effect, apparent activation energy and intermediate products, Environmental Science and Technology 33 (1), pp , [14] APHA, AWWA, WPCF, Standard methods for the examination of water and wastewater, 20 th Ed., APHA, New York, [15] Gillham, R.W., O Hannesin, S.F., Long-term performance of an in situ iron wall for remediation of VOCs, Ground Water 36 (1), pp , [16] Johnson, T.L., Scherer, M.M., Tratnyek, P.G., Kinetics of halogenated organic compound degradation by iron metal, Environmental Science and Technology 30 (8), pp , 1996.