- Email: [email protected]
Is it possible to remediate a BTEX contaminated chalky aquifer by in situ chemical oxidation?
Chemosphere 84 (2011) 1181–1187
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Is it possible to remediate a BTEX contaminated chalky aquifer by in situ chemical oxidation? Julien Lemaire a, Véronique Croze b, Joachim Maier b, Marie-Odile Simonnot a,⇑ a b
Laboratoire Réactions et Génie des Procédés, Nancy Université – INPL – CNRS, 1 rue Grandville BP 20451, 54001 Nancy Cedex, France ICF Environnement, 14 à 30 rue Alexandre, 92635 Gennevilliers, France
a r t i c l e
i n f o
Article history: Received 8 April 2011 Received in revised form 10 June 2011 Accepted 13 June 2011 Available online 5 July 2011 Keywords: BTEX ISCO Permanganate Persulfate Percarbonate
a b s t r a c t An industrial coating site in activity located on a chalky plateau, contaminated by BTEX (mainly xylenes, no benzene), is currently remediated by in situ chemical oxidation (ISCO). We present the bench scale study that was conducted to select the most appropriate oxidant. Ozone and catalyzed hydrogen peroxide (Fenton’s reaction) were discarded since they were incompatible with plant activity. Permanganate, activated percarbonate and activated persulfate were tested. Batch experiments were run with groundwater and groundwater-chalk slurries with these three oxidants. Total BTEX degradation in groundwater was reached with all the oxidants. The molar ratios [oxidant]:[Fe2+]:[BTEX] were 100:0:1 with permanganate, 100:100:1 with persulfate and 25:100:1 with percarbonate. Precipitation of either manganese dioxide or iron carbonate (siderite) occurred. The best results with chalk slurries were obtained with permanganate at the molar ratio 110:0:1 and activated persulfate at the molar ratio 110:110:1. To avoid precipitation, persulfate was also used without activation at the molar ratio 140:1. Natural Oxidant Demand measured with both oxidants was lower than 5% of initial oxidant contents. Activated percarbonate was not appropriate because of radical scavenging by carbonated media. Permanganate and persulfate were both effective at oxidant concentrations of ca 1 g kg1 with permanganate and 1.8 g kg1 with persulfate and adapted to site conditions. Activation of persulfate was not mandatory. This bench scale study proved that ISCO remediation of a chalky aquifer contaminated by mainly xylenes was possible with permanganate and activated or unactivated persulfate. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction BTEX (benzene, toluene, ethylbenzene and xylenes) are widespread contaminants of groundwater and soils. They have been widely used in industrial processes and are constituents of fossil fuels. These volatile and water soluble chemicals may be dispersed in the environment because of accidents and storage tank leakages (Pawlowski, 1998); they represent a threat to human health and ecosystems because of their toxicity (ATSDR, 2004). According to WHO, xylene content limit is 500 lg L1 in drinking water. In situ chemical oxidation (ISCO) is an increasingly popular method for the remediation of groundwater and soils contaminated by organic chemicals such as chlorinated VOCs (volatile organic compounds), MTBE (methyl tert-butyl ether), BTEX, PAHs (polycyclic aromatic hydrocarbons), pesticides and explosives (ITRC, 2005; Huling and Pivetz, 2006). Used for two decades in the USA, this technique is still emerging in Europe and mainly used for chlorinated VOCs. ISCO consists of injecting a chemical oxidant
⇑ Corresponding author. Tel.: +33 (0)383 175 260; fax: +33 (0)383 322 975. E-mail address: [email protected] (M.-O. Simonnot). 0045-6535/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2011.06.052
into groundwater or soil to degrade contaminants into less toxic substances. The most frequently used oxidants are potassium or sodium permanganate, hydrogen peroxide combined with Fe(II) (Fenton’s reaction), sodium persulfate, and ozone. Permanganate has received considerable attention and has been frequently used in remediation projects because of its relative ease of use and its persistence in the environment (Bennett, 2002; Li and Schwartz, 2004a,b; Waldemer and Tratnyek, 2006; Heiderscheidt et al., 2008; Crimi et al., 2009; Woo, 2009; Krembs et al., 2010). Fenton’s reaction has been extensively studied as well (Neyens and Baeyens, 2003; Huling and Pivetz, 2006; Crimi and Taylor, 2007; De Souza e Silva et al., 2009; Sutton et al., 2011). Hydroxyl radicals generated by this reaction are very powerful (redox potential E° = 2.7 V) but their persistence is rather low. The use of concentrated hydrogen peroxide solutions may be hazardous, thus for safety reasons, solid reagents such as sodium percarbonate, slowly releasing hydrogen peroxide, are of interest (Birnstingl et al., 2006; Ruffing et al., 2009). Persulfate, a relatively new ISCO reagent, is a strong and very persistent oxidant (E° = 2.0 V). Its chemical activation with Fe(II) leads to the generation of sulfate radical SO 4 comparable to hydroxyl radical (Liang and Lee, 2008; Liang et al., 2008, 2009, 2011; Tsitonaki et al., 2010). Hydrogen peroxide and sodium
1182
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187
persulfate have been compared in different conditions for BTEX destruction (Crimi and Taylor, 2007). An ISCO project begins with a laboratory study aiming at selecting the most appropriate oxidant and measuring parameters such as the Natural Oxidant Demand (NOD). NOD is the mass of oxidant consumed per mass of dry solid by the reaction of a variety of nontarget reduced species (organic matter, reduced forms of metals, etc.) (Mumford et al., 2005). NOD is a key design parameter and one must keep in mind that ISCO is not technically and/or economically feasible at very high NOD. In the present case, we consider an industrial coating site located in the North of France on a chalk plateau contaminated by a wide xylene plume, with concentrations up to saturation in some areas and around 70 mg L1 in the central area of the plume. A hydraulic barrier was installed to stop potential propagation towards municipal drinking water wells as an emergency measure. Then, the issue of determining an efficient remediation technique was addressed and ISCO has been envisioned as a fast and potentially efficient method for mass reduction. The present bench scale study was conducted to qualify suitable technology to be applied to the plant in current activity. Therefore, ozone was discarded for industrial production constraint; Fenton’s reaction was discarded as well for safety reasons and because of the high buffering capacity of the aquifer matrix (chalk). Three oxidants were investigated: permanganate, activated percarbonate and activated/unactivated persulfate. Our objective was to test their efficiency in order to reach xylene degradation in a reasonable time frame. Batch experiments were performed with groundwater and with groundwater-chalk slurries to approximate real site conditions. NOD was measured as well. 2. Materials and methods 2.1. Groundwater and chalk samples Groundwater sample, called GW, was collected at the industrial area of a coating plant located in the Picardie region, in the North of France. Groundwater contamination has been extended on a 100 m2 area, 2 m height at ca 12 m depth in the capillary fringe. It contained about 10 t BTEX, mainly xylenes. The collected sample contained 15.2 mg L1 xylenes, representing 80 wt.% of total BTEX content, but no benzene (Table 1). The major anion was hydrogen carbonate (275 mg L1) and pH was 6.9 (Table 1). Two contaminated chalk samples, called CCh1 and CCh2, were collected in the capillary fringe. Total BTEX content was ca 6 mg kg1 in CCh1 and ca 1 mg kg1 in CCh2 (85% xylenes, no benzene). Non-contaminated chalk samples, called NCCh1 to NCCh3, were collected from the same industrial site at different depths. 2.2. Chemicals BTEX standard (1 mg L1) and sulfuric acid (95–97 wt.%) were supplied by Sigma–Aldrich, sodium percarbonate (OXIPER-S131) Table 1 Groundwater characteristics. Major ions
2+
Ca Mg2+ SO2 4 HCO 3 Cl
BTEX
mg L1
mM
180.0 12.4 2.76
4.50 0.52 0.03
Benzene Toluene Ethylbenzene
<0.2 1.16 2.75
<2.6 12.6 26.0
275.0 11.3
4.50 0.32
o-xylene m-xylene p-xylene Total xylene Total BTEX
1.71 11.40 2.13 15.24 19.35
16.1 107.6 20.1 143.8 191.5
Na+, K+, NO 3 not measured pH = 6.9
mg L1
lM
by Solvay Chemicals, potassium persulfate (>99 wt.%) by Acros Organics, ortho-phosphoric acid (85 wt.%) by Merck, potassium permanganate (>99 wt.%) and sodium thiosulfate (>99 wt.%) by Fluka Chemika, hydrogen peroxide (50 wt.%), ferrous sulfate heptahydrate Rectapur and sodium carbonate Normapur by VWR Prolabo. 2.3. Oxidation experiments Batch experiments were performed at the bench scale to select the most appropriate one for this project. Oxidation experiments were run with site groundwater and with chalk with the three oxidants. NOD was measured with permanganate and persulfate. 2.3.1. Groundwater oxidation In the first series, four preliminary experiments were run with 200 mL of GW in 500 mL capped flasks (Table 2, exp. #1–4). Oxidants were potassium permanganate, potassium persulfate activated by Fe(II) and sodium percarbonate, the molar ratio [oxidant]:[BTEX] was 100:1. Flasks were shaken at 150 rpm on a rotating table for 160 h. In the second series, eight experiments were run in stirred 500 mL triple neck flasks with temperature and pH monitoring for 48 h (Table 2, exp. #5–12). Experiments #5 and 6 used permanganate and persulfate at a molar ratio [oxidant]:[BTEX] of 10:1, 10 times lower than previously. Five experiments (#7–11) were run with percarbonate activated by Fe(II) at different [oxidant]:[catalyst]:[BTEX] molar ratios (Table 2). In experiment #11, Fe(II) was introduced in 10 sequences with a time interval of 10 min between them. A control experiment was performed as well (exp. #12). Both series were carried out at 20 °C, the flasks being covered in aluminum foil to prevent photocatalytic degradation of BTEX. Samples of GW (5 mL) were periodically taken; they were placed in 20 mL closed sampling vessel with 4 g sodium carbonate and 1 g sodium thiosulfate to stop oxidation. BTEX concentration was measured. For the second series, residual permanganate or hydrogen peroxide was analysed. 2.3.2. Contaminated chalk oxidation Oxidation was run with the contaminated chalk sample CCh1 with potassium permanganate, activated potassium persulfate and activated sodium percarbonate (Table 2, exp. #13–15) at 20 °C. Chalk samples (250 g) and ultra pure water (125 mL) were placed in 500 mL capped flasks and stirred at 150 rpm. During experiment #13, potassium permanganate was introduced at the molar ratio [KMnO4]:[BTEX] 10:1; after 144 h, a second addition was done to reach the global ratio 110:1. The same procedure was followed with activated potassium persulfate, at the initial molar ratio [persulfate]:[Fe2+]:[BTEX] 10:10:1 and, after 144 h, at the global ratio 110:110:1. Activated sodium percarbonate was directly introduced at the molar ratio [percarbonate]:[Fe2+]:[BTEX] 100:100:1. Additional experiments were performed with potassium persulfate without activation (Table 2, exp. #16). Chalk samples CCh2 (100 g) and ultra pure water (50 mL) were placed in 250 mL flasks and stirred at 150 rpm at 20 °C. Potassium persulfate was introduced at the molar ratio [persulfate]:[BTEX] 140:1. In all experiments, 5 mL samples were periodically taken and analysed. 2.3.3. NOD NOD was measured with three samples of non contaminated chalk (NCCh1 to NCCh3) (Frasco et al., 2005). Each chalk sample was divided into four or five samples of 50 g that were introduced in 250 mL flasks with hermetic capping. The first one enabled us to measure humidity, the three other ones to measure NOD after 48 h at three initial concentrations. The fifth one was used to measure long term NOD.
1183
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187 Table 2 Conditions of the oxidation experiments performed with contaminated groundwater (#1–12) and chalk (#13–16). Experimental conditions
Monitored parameters 2+
#
Oxidant
Molar ratio [Ox]:[Fe ]:[BTEX]
Ratio to SMR
Oxidant mg L1 (1–12)
FeSO47H2O mg kg1 (13–16)
BTEX
Oxidant
Temperature
pH
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
KMnO4 K2S2O8 2Na2CO33H2O2 – KMnO4 K2S2O8 2Na2CO33H2O2 2Na2CO33H2O2 2Na2CO33H2O2 2Na2CO33H2O2 2Na2CO33H2O2 – KMnO4 K2S2O8 2Na2CO33H2O2 K2S2O8
100:0:1 100:100:1 100:100:1 – 10:0:1 10:10:1 10:10:1 50:50:1 25:100:1 10:100:1 10:10:1 – 110:0:1 110:110:1 100:100:1 140:0:1
7.2 4.8 14.5
2370 4050 4710 – 237 405 471 2355 1178 471 471 – 1043 1784 1884 420
– 4170 4170 – – 417 417 2085 4170 4170 417 – – 1835 1668 –
x x x x x x x x x x x x x x x x
– – – – x – x x x x x – x – x –
– – – – x x x x x x x x – – – –
– – – – x x x x x x x x – – – –
0.72 0.48 1.45 7.25 3.62 1.45 1.45 7.92 5.28 14.5 6.72
Ultra pure water and potassium permanganate or persulfate (3 wt.%) were added into the flask and chalk was added to reach a mass ratio (water):(dry chalk) of 3 (Table 3). Control experiments were done as well. Flasks were capped and kept at ambient temperature (20 °C) in the dark. They were rotated at the beginning of the experiment and then twice a day. At the end, supernatant was sampled and residual oxidant concentration was measured, enabling us to calculate NOD.
2.4. Analytical methods 2.4.1. BTEX analysis BTEX were analyzed following ISO 11423-1 procedure. Head space method consisted of heating at 80 °C for 1 h a given volume of non filtered water in a 20 mL glass vial fitted with crimped aluminum caps lined with PTFE-coated butyl rubber septa. An aliquot
of the gas phase was then taken and injected into a GC equipped with a FID. We used a Varian 3900 GC with a capillary column DB-WAX (30 m 0.54 mm 0.85 lm, Grace). Injector and detector temperature was 250 °C; oven temperature was programmed to increase from 70 to 115 °C at 5 °C min1. The carrier gas was helium, the flow rate was 3.9 mL min1, and the split ratio was 10%. BTEX analysis was run in 8.2 min. External standards of benzene, toluene, ethylbenzene, o-, m- and p-xylene were used. Standards and samples were prepared according to the ISO 11423-1 procedure. Septum glass vials of 20 mL were used in which 5 mL sample was introduced with 4 g sodium carbonate and, if necessary, 1 g sodium thiosulfate to stop oxidation reaction. 2.4.2. Residual oxidant analysis Residual oxidants were analysed by conventional titration. Permanganate was titrated by hydrogen peroxide (Brumblay, 1971)
Table 3 Experimental conditions for NOD determination. #
Sample
mwet (g)
mdry (g)
Potassium permanganate 1 NCCh1 2 NCCh1 3 NCCh1 4 NCCh1 5 NCCh2 6 NCCh2 7 NCCh2 8 NCCh3 9 NCCh3 10 NCCh3 11 – 12 – 13 –
50.2 50.5 50.9 51.6 51.8 51.5 50.4 50.9 51.3 50.4 – – –
38.8 39.0 39.3 39.9 35.9 35.6 34.9 39.1 39.5 38.8 – – –
Potassium persulfate 14 NCCh1 15 NCCh1 16 NCCh1 17 NCCh1 18 NCCh2 19 NCCh2 20 NCCh2 21 NCCh3 22 NCCh3 23 NCCh3 24 – 25 – 26 –
50.6 50.7 50.8 50.6 50.1 50.2 50.1 50.7 50.8 50.0 – – –
39.1 39.2 39.3 39.1 34.7 34.8 34.7 39.0 39.1 38.5 – – –
Vox (mL)
Vwater (mL)
Duration
mwater/mdry
Coxidant (g kgdry )
5 25 50 50 5 25 50 5 25 50 5 25 50
95 75 50 50 95 75 50 95 75 50 95 75 50
48 h 48 h 48 h 21 d 48 h 48 h 48 h 48 h 48 h 48 h 48 h/6 wk 48 h/6 wk 48 h/6 wk
2.9 2.9 2.8 2.8 3.2 3.3 3.3 2.9 2.8 2.9 – – –
4 19 38 38 4 21 43 4 19 39 – – –
10 50 100 100 10 50 100 10 50 100 10 50 100
90 50 0 0 90 50 0 90 50 0 90 50 0
48 h 48 h 48 h 13 d 48 h 48 h 48 h 48 h 48 h 48 h 48 h/6 wk 48 h/6 wk 48 h/6 wk
2.9 2.8 2.8 2.9 3.3 3.3 3.3 2.9 2.9 2.9 – – –
6 32 64 64 7 36 72 6 32 65 – – –
1
1184
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187
and hydrogen peroxide by permanganate (hydrogen peroxide is also a reducing agent). Persulfate was measured by indirect titration. Persulfate sample was acidified with sulfuric and phosphoric acids; Fe(II) added in excess was titrated by permanganate. 3. Results and discussion 3.1. BTEX oxidation in GW 3.1.1. With permanganate and activated persulfate BTEX oxidation in GW was carried out with permanganate and activated persulfate as outlined in Table 2. To compare our results, we expressed the oxidant quantities in terms of stoichiometric molar ratio (SMR). We defined this ratio as the number of moles of oxidant required to degrade one mole of a target compound to the final products CO2 and H2O. Stoichiometric balances were written for toluene, ethylbenzene and xylenes. For instance, xylene oxidation by permanganate yielded:
14MnO4 þ C8 H10 þ 2H2 O ! 8CO2 þ 14OH þ 14MnO2
ð1Þ
Then SMR [KMnO4]:[Xylenes] was equal to 14. The global SMR [KMnO4]:[BTEX] was calculated by weighing by BTEX composition (Table 1), it was equal to 13.8, close to 14 since xylenes were the major contaminants. BTEX oxidation by persulfate radicals obeys a series of radical reactions that are not accurately known. Following the same assumptions as previously, the stoichiometric balance of xylene oxidation yielded: 2 þ 21S2 O2 8 þ C8 H10 þ 16H2 O ! 8CO2 þ 42H þ 42SO4
ð2Þ
SMR for xylene oxidation by activated persulfate was 21 and 20.7 for BTEX oxidation. All our experiments were run with an excess of oxidants compared to BTEX, at molar ratios [Oxidant]:[BTEX] between 0.5-fold and 15-fold the SMR (Table 2). With activated persulfate, the molar ratio [K2S2O8]:[Fe2+] was set to unity as in other contributions (Liang et al., 2008). Residual BTEX ratios were plotted against time for oxidation by permanganate and persulfate (Fig. 1). At high permanganate concentration (Table 2, exp. #1), BTEX were slowly degraded and reached a residual concentration of 3% after 114 h (ca 5 d) (Fig. 1). At high persulfate concentration (Table 2, exp. #2), complete degradation was attained in less than 15 h (Fig. 1). Compared to chlorinated compounds, BTEX degradation is much slower. Indeed, trichloroethylene oxidation with persulfate occurred in a few minutes (Liang et al., 2004a,b). With both oxidants, BTEX degradation could be fairly well represented by pseudo-first order rates with effective constants of 0.06 h1 for permanganate and 2.8 h1 for activated persulfate. Residual permanganate analysis (exp. #5) showed that the molar ratio [KMnO4 consumed]:[BTEX degraded] was equal to 20 close to the SMR equal to 14 (Eq. 1). Then permanganate was almost totally consumed by BTEX oxidation, suggesting a low depletion by oxidation of non-targeted compounds. The final objective of this research was to select effective conditions for ISCO. The above conditions enabled us to reach this goal but we wanted to use lower oxidant concentrations to reduce costs. We used the molar ratios [MnO 4 ]:[BTEX] 10:1 (Table 2, exp. #5) and [K2S2O8]:[Fe2+]:[BTEX] 10:10:1 (Table 2, exp. #6) that were lower than SMR. Results are plotted in Fig. 1. A slow and incomplete degradation of BTEX was observed: ca 60% residual BTEX were measured after 160 h with permanganate (exp. #5) and ca 40% with persulfate (exp. #6). These results clearly showed that these concentrations were not sufficient. Intermediate concentrations should be tested to complete this study.
At this level, we can conclude that both oxidants were effective at high concentrations. Permanganate was expected to work well since there was no benzene. Benzene oxidation would have been much slower. However, significant precipitations occurred. Permanganate oxidation produced a brown precipitate of manganese dioxide. When formed in large amounts, manganese dioxide particles may be a drawback since they deposit in the subsurface and may reduce permeability (Sirguey et al., 2008). Anyway recent researches have proved that MnO2 particle growth and transport could be controlled and suspensions stabilized (Crimi and Ko, 2009). Similarly, activated persulfate induced iron carbonate (siderite) precipitation (solutions turned quickly to dark orange). Speciation in solution was calculated with the computer code CHESS (van der Lee and de Windt, 2000). Results predicted siderite precipitation and the related pH variations. In this case, precipitation involved Fe(II), suppressing activation. The question of activation necessity may be addressed since oxidation succeeded despite Fe(II) depletion caused by precipitation. 3.1.2. With activated percarbonate Xylene oxidation by activated percarbonate can be described by:
72Na2 CO3 3H2 O2 þ C8 H10 ! 28Naþ þ 18H2 O þ 16HCO3 þ 6CO2 3 ð3Þ Giving a SMR [percarbonate]:[xylene] of 7 and a SMR [percarbonate]:[BTEX] of 6.9 for this GW. Three experiments were run at the molar ratio [percarbonate]:[BTEX] 10:1 corresponding to 1.4-fold the SMR (Table 2, exp. #7, 10 and 11). Two experiments were performed at the conventional molar ratio [percarbonate]:[Fe2+] 1:1, Fe(II) being added at the beginning (exp. #7) or gradually (exp. #11). Then activation was enhanced by setting the molar ratio [percarbonate]:[Fe2+] at 1:10. Three experiments were run at higher percarbonate concentrations with molar ratios [percarbonate]:[BETX] 25:1 (exp. #9), 50:1 (exp. #8) and 100:1 (exp. #3) at different [percarbonate]:[Fe2+] ratios (Table 2). The evolution of residual BTEX is plotted in Fig. 2. We observed a rather poor degradation rate without activation despite a high percarbonate concentration equal to 14.5-fold the SMR (exp. #3) which was attributed to the absence of catalyst. The high concentration of percarbonate may be a drawback as well since high concentrations of hydroxyl radicals lead to radical scavenging (ITRC, 2005; Huling and Pivetz, 2006). At the conventional molar ratio [percarbonate]:[Fe2+] 1:1 (exp. #7 and 11), degradation was low, ca 50% of initial BTEX were still present after 48 h (Fig. 2). Gradual Fe(II) addition (exp. #11) had no effect compared to direct addition (exp. #7). Increasing 5-fold both oxidant and catalyst concentrations (exp. #8) did not lead to better results: ca 20% of initial BTEX were still detected after 20 h. Raising activation 10-fold while maintaining [percarbonate]:[BETX] ratio 10:1 (exp. #10) had no effect, ca 18% of initial BTEX were present after 40 h. The last conditions, consisting of setting [percarbonate]:[Fe2+]:[BETX] at 25:100:1 led to a complete BTEX degradation in less than 1 h, the major fraction being degraded in 10 min. These experiments enabled us to determine oxidant and catalyst concentrations to degrade BTEX, resulting in high percarbonate concentration (4-fold the SMR given in Eq. 3) and very high Fe(II) concentration. Percarbonate and Fe(II) addition affected GW pH (6.9 initially). Percarbonate addition increased pH because of carbonate addition while Fe(II) addition made it decrease because of iron carbonate precipitation. These results were consistent with the predictions obtained with the speciation modeling program CHESS. In our experiments, pH did not change at the molar ratio [percarbon-
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187
1185
Fig. 1. Evolution of the residual BTEX ratio during batch oxidation in groundwater with potassium permanganate and potassium activated persulfate. [oxidant]:[Fe2+]:[BTEX] were 100:0:1 (s), 10:0:1 (j) for permanganate, 100:100:1 (), 10:10:1 (4) for persulfate and 0:0:1 (x) for control experiment. The dotted lines are construction lines and the curves represent first-order rates for the highest oxidant concentrations.
Fig. 2. Evolution of the residual BTEX ratio during batch oxidation in groundwater with activated sodium percarbonate. [oxidant]:[Fe2+]:[BTEX] were 100:0:1 (}), 50:50:1 (x), 10:10:1 (j), 10:10:1 with gradual supply of Fe2+ (s), 10:100:1 (), 25:100:1 (4) and 0:0:1 () for control experiment. Dotted lines are construction lines.
ate]:[Fe2+]:[BTEX] 10:10:1 (exp. #7 and 11). It increased to 7.6 at 50:50:1 (exp. #8) and, at high Fe(II) concentrations, it drastically decreased to 5.3 at 10:100:1 (exp. #10) and 3.9 at 25:100:1 (exp. #11). Again precipitation occurred as with persulfate, solutions turned from colorless to dark orange. Dramatic pH decrease and precipitation could be a major drawback for ISCO feasibility at the field scale.
3.1.3. Conclusions on BTEX oxidation in GW Considering all the experiments run with GW, the three ones leading to the best degradation results were (i) exp. #1 with permanganate at [oxidant]:[Fe2+]:[BTEX] set at 100:0:1, (ii) exp. #2 with persulfate at 100:100:1 and (iii) exp. #9 with percarbonate at 25:100:1. In the former case, manganese dioxide precipitation occurred and fine particles were formed. In the other cases, the
1186
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187
addition of Fe(II) led to the formation of flocs and precipitation of iron carbonate that was confirmed by speciation modelling. Other experiments (not presented here) run with citric acid as an iron chelatant (Tsitonaki et al., 2010) made precipitation decrease but did not improve significantly BTEX degradation. Then as a partial conclusion, in the objective of upscaling, we could chose potassium permanganate because of its persistence in solution as well as the formation of finer particles than iron carbonate precipitates. 3.2. BTEX degradation in chalk samples Oxidation experiments were performed with chalk slurries to approach field reality. They were run with a ratio (solid mass): (solution volume) of 2 gdry chalk mL1. Initial pH was 8.9. In the field, the accessible porosity of chalk was ca 5%. One must keep in mind that laboratory conditions were more favorable to BTEX oxidation than field conditions because of lower (solid mass):(solution volume) ratio, stirring and higher temperature (20 °C instead of ca 10 °C in the aquifer). Besides we used rather low oxidant concentrations in order to ensure feasibility at the field scale. 3.2.1. BTEX oxidation with the 3 oxidants Oxidation experiments with chalk were performed at molar ratios [oxidant]:[Fe2+]:[BTEX] 10:0:1 with permanganate (exp. #13), 10:10:1 with persulfate (exp. #14) and 100:100:1 (exp. #15) with percarbonate. Results are presented in Fig. 3. With permanganate and persulfate, we observed that BTEX residual ratio increased at first. This increase was attributed to BTEX desorption from the solid matrix occurring at a higher rate than oxidation. Exp. #15 with percarbonate was run afterward and we mixed the slurry for 2 h before oxidant addition; for this reason the initial BTEX concentration was higher than that of other. After 144 h (6 d), BTEX concentration was still high with the three oxidants. For permanganate and persulfate, a second addition was done at a higher concentration. The experiment with percarbonate was stopped since it had shown to be ineffective: after 2 d, percarbonate was completely depleted while BTEX were not
degraded. Permanganate led to rapid and almost complete BTEX degradation (exp. #13) and persulfate to complete degradation (exp. #14) after 2 d. Again manganese dioxide or iron carbonate precipitation occurred. Further experiments run with persulfate without activation showed that it was not required. Indeed, complete BTEX degradation was reached after 2 d with the molar ratio [K2S2O8]:[BTEX] 140:1. These results are not plotted for clarity reasons since they were obtained with the chalk sample CCh2 that was less contaminated than CCh1. However, they clearly showed the possibility of decontaminating with persulfate without activation. Our results showed that radical reactions involving hydroxyl radicals released by percarbonate according to Eq. 3 were effective in GW but not in chalk slurries. This result was attributed to the high carbonate concentration in chalk slurries responsible for radical scavenging. Carbonate was initially present in solution because of chalk-water equilibrium (initial pH: 8.9) and was added through percarbonate addition as expressed in Eq. 3. Moreover, Fe(II) failed to activate percarbonate since it was consumed by precipitation reaction. On the contrary, permanganate and persulfate with and without activation by Fe(II) enabled us to attain high BTEX degradation rates in both media (we recall that benzene was absent). Oxidation by permanganate and activated persulfate were accompanied by precipitation that may be a drawback especially with iron carbonate occurring with activated persulfate. This drawback was avoided by using persulfate without activation.
3.2.2. Oxidant persistence Residual permanganate and persulfate concentrations were measured after 48 h and after 6 wk in control experiments (Table 3, exp. #11–13, 24–26) to estimate oxidant loss. It was negligible compared to the amounts that were used (<3% after 6 wk). Permanganate and persulfate were very persistent in our experimental conditions. In field conditions, very low decomposition rates of permanganate and persulfate are expected, which is in agreement with reported data (Brown, 2003; Huling and Pivetz, 2006).
Fig. 3. Evolution of the residual BTEX concentration in chalk-groundwater slurries with permanganate (N), activated persulfate () and activated percarbonate (}). Dotted lines are construction lines.
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187
3.2.3. NOD NOD was calculated according to the following equation:
NOD ¼
½Oxcontrol V control ½Oxsup V sup m
ð4Þ
where NOD is the Natural Oxidant Demand (g kg1), [Ox]sup the oxidant concentration in the supernatant (g L1), Vsup the supernatant volume (L), [Ox]control the oxidant concentration in the control (g L1), Vcontrol the control solution volume (L), m the mass of dry chalk (kg). NOD was measured after 48 h with three samples of non contaminated chalk with permanganate and with persulfate at three initial concentrations (Table 3). In each case, NOD was lower than 2% of the initial oxidant concentration (not statistically different from 0 according to error bars). The same was observed at longer times, after 13 and 21 d (NOD <5% of the initial oxidant content). The totality of the oxidant was consumed for BTEX degradation. This result was very important in the perspective of upscaling.
4. Conclusions This work showed that permanganate, activated persulfate and activated percarbonate (activated with Fe(II)) enabled us to attain BTEX destruction in the groundwater of a chalky aquifer. Activated percarbonate failed to degrade BTEX in a slurry composed of groundwater and chalk particles. This result was attributed to (i) radical scavenging in carbonated medium and (ii) inefficiency of unactivated percarbonate because of siderite precipitation involving Fe(II) and carbonate. Permanganate and activated persulfate were both effective in this slurry. Moreover, they were persistent for at least 6 wk and NOD was very low. However, both of them induced significant precipitation. Further experiments showed that persulfate could be used without activation, no precipitation occurred. In conclusion, this study enabled us to demonstrate ISCO feasibility to remediate a chalky aquifer contaminated by BTEX (that were mainly xylenes without benzene). Further experiments could be performed but we proved the efficacy of two possible oxidants at the bench scale. Permanganate and unactivated persulfate were selected for pilot tests that were conducted before implementing ISCO at the field scale. References ATSDR, 2004. Interaction Profile for Benzene, Toluene, Ethylbenzene, and Xylenes (BTEX). US Department of Health and Human Services, Public Health Service, Atlanta. Bennett, G.F., 2002. Principles and practices of in situ chemical oxidation using permanganate. Book Review of: Siegrist, R.L., Urynowicz, M.A., West, O.R., Crimi, M.L., Lowe, K.S., Battelle Press, Columbus, OH, USA, 001. J. Hazard. Mater. 90, 323–324. Birnstingl, J., Kelley, B., Koenigsberg, S., 2006. Field results with an alkaline in situ chemical oxidation process. Contam. Reclam. 14, 539. Brown, R.A., 2003. In Situ Chemical Oxidation: Performance, Practice, and Pitfalls. AFCEE Technology Transfer Workshop, San Antonio, Texas. Brumblay, R.U., 1971. Quantitative Analysis – College Outline, second ed. HarperCollins, New York. Crimi, M., Ko, S., 2009. Control of manganese dioxide particles resulting from in situ chemical oxidation using permanganate. Chemosphere 74, 847–853.
1187
Crimi, M.L., Taylor, J., 2007. Experimental evaluation of catalyzed hydrogen peroxide and sodium persulfate for destruction of BTEX contaminants. Soil Sediment Contam. 16, 29–45. Crimi, M., Quickel, M., Ko, S., 2009. Enhanced permanganate in situ chemical oxidation through MnO2 particle stabilization: evaluation in 1-D transport systems. J. Contam. Hydrol. 105, 69–79. De Souza e Silva, P.T., Da Silva, V.L., De Barros Neto, B., Simonnot, M.-O., 2009. Phenanthrene and pyrene oxidation in contaminated soils using Fenton’s reagent. J. Hazard. Mater. 161, 967–973. Frasco, K.A., Osborne, M.W., Vella, P.A., 2005. The permanganate soil oxidant demand: a critical parameter for ISCO remediation. In: The 21st Annual Conference on Soils, Sediments and Water. University of Massachusetts at Amherst, USA, pp. 32–33.. Krembs, F.J., Siegrist, R.L., Crimi, M.L., Furrer, R.F., Petri, B.G., 2010. ISCO for groundwater remediation: analysis of field applications and performance. Ground Water Monit. Remed. 30, 42–53. Li, X.D., Schwartz, F.W., 2004a. DNAPL remediation with in situ chemical oxidation using potassium permanganate: II. Increasing removal efficiency by dissolving Mn oxide precipitates. J. Contam. Hydrol. 68, 269–287. Li, X.D., Schwartz, F.W., 2004b. DNAPL remediation with in situ chemical oxidation using potassium permanganate: Part I. Mineralogy of Mn oxide and its dissolution in organic acids. J. Contam. Hydrol. 68, 39–53. Liang, C., Lee, I.L., 2008. In situ iron activated persulfate oxidative fluid sparging treatment of TCE contamination – a proof of concept study. J. Contam. Hydrol. 100, 91–100. Liang, C., Bruell, C.J., Marley, M.C., Sperry, K.L., 2004a. Persulfate oxidation for in situ remediation of TCE. I. Activated by ferrous ion with and without a persulfate– thiosulfate redox couple. Chemosphere 55, 1213–1223. Liang, C., Bruell, C.J., Marley, M.C., Sperry, K.L., 2004b. Persulfate oxidation for in situ remediation of TCE. II. Activated by chelated ferrous ion. Chemosphere 55, 1225–1233. Liang, C., Huang, C.-F., Chen, Y.-J., 2008. Potential for activated persulfate degradation of BTEX contamination. Water Res. 42, 4091–4100. Liang, C., Lin, Y.-T., Shih, W.-H., 2009. Treatment of trichloroethylene by adsorption and persulfate oxidation in batch studies. Ind. Eng. Chem. Res. 48, 8373–8380. Liang, S.H., Kao, C.M., Kuo, Y.C., Chen, K.F., 2011. Application of persulfate-releasing barrier to remediate MTBE and benzene contaminated groundwater. J. Hazard. Mater. 185, 1162–1168. Mumford, K.G., Thomson, N.R., Allen-King, R.M., 2005. Bench-scale investigation of permanganate Natural Oxidant Demand kinetics. Environ. Sci. Technol. 39, 2835–2840. Neyens, E., Baeyens, J., 2003. A review of classic Fenton’s peroxidation as an advanced oxidation technique. J. Hazard. Mater. 98, 33–50. Pawlowski, M.H., 1998. Analytical and Field Test Methods for Measuring BTEX Metabolite Occurrence and Transport in Groundwater. Oregon State University Corvallis Department of Chemistry, Defense Technical Information Center. Ruffing, S., Kline, R., Poling, B., 2009. When cleanup is rocket science: remediation technology spares sensitive underground infrastructure and cleanup dollars at the Kennedy space center. Pollut. Eng. 41 (2), 20–22. Sirguey, C., Souza, De., e Silva, P.T., Schwartz, C., Simonnot, M.-O., 2008. Impact of chemical oxidation on soil quality. Chemosphere 72, 282–289. Sutton, N., Grotenhuis, J., Langenhoff, A., Rijnaarts, H., 2011. Efforts to improve coupled in situ chemical oxidation with bioremediation: a review of optimization strategies. J. Soil Sediment 11, 129–140. Tsitonaki, A., Petri, B., Crimi, M., Mosbk, H., Siegrist, R.L., Bjerg, P.L., 2010. In situ chemical oxidation of contaminated soil and groundwater using persulfate: a review. Crit. Rev. Environ. Sci. Technol. 40, 55–91. Van der Lee, J., De Windt, L., 2000. Chess tutorial and cookbook, updated for version 2.5. Ecole Nationale Supérieure des Mines de Paris, Centre d’Informatique Géologique, Rep. Nr LHM/RD/00/13, Paris. Waldemer, R.H., Tratnyek, P.G., 2006. Kinetics of contaminant degradation by permanganate. Environ. Sci. Technol. 40, 1055–1061. Woo, N.C., 2009. Characteristics of permanganate oxidation of TCE at low reagent concentrations. Environ. Technol. 30, 1337–1342.
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Is it possible to remediate a BTEX contaminated chalky aquifer by in situ chemical oxidation? Julien Lemaire a, Véronique Croze b, Joachim Maier b, Marie-Odile Simonnot a,⇑ a b
Laboratoire Réactions et Génie des Procédés, Nancy Université – INPL – CNRS, 1 rue Grandville BP 20451, 54001 Nancy Cedex, France ICF Environnement, 14 à 30 rue Alexandre, 92635 Gennevilliers, France
a r t i c l e
i n f o
Article history: Received 8 April 2011 Received in revised form 10 June 2011 Accepted 13 June 2011 Available online 5 July 2011 Keywords: BTEX ISCO Permanganate Persulfate Percarbonate
a b s t r a c t An industrial coating site in activity located on a chalky plateau, contaminated by BTEX (mainly xylenes, no benzene), is currently remediated by in situ chemical oxidation (ISCO). We present the bench scale study that was conducted to select the most appropriate oxidant. Ozone and catalyzed hydrogen peroxide (Fenton’s reaction) were discarded since they were incompatible with plant activity. Permanganate, activated percarbonate and activated persulfate were tested. Batch experiments were run with groundwater and groundwater-chalk slurries with these three oxidants. Total BTEX degradation in groundwater was reached with all the oxidants. The molar ratios [oxidant]:[Fe2+]:[BTEX] were 100:0:1 with permanganate, 100:100:1 with persulfate and 25:100:1 with percarbonate. Precipitation of either manganese dioxide or iron carbonate (siderite) occurred. The best results with chalk slurries were obtained with permanganate at the molar ratio 110:0:1 and activated persulfate at the molar ratio 110:110:1. To avoid precipitation, persulfate was also used without activation at the molar ratio 140:1. Natural Oxidant Demand measured with both oxidants was lower than 5% of initial oxidant contents. Activated percarbonate was not appropriate because of radical scavenging by carbonated media. Permanganate and persulfate were both effective at oxidant concentrations of ca 1 g kg1 with permanganate and 1.8 g kg1 with persulfate and adapted to site conditions. Activation of persulfate was not mandatory. This bench scale study proved that ISCO remediation of a chalky aquifer contaminated by mainly xylenes was possible with permanganate and activated or unactivated persulfate. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction BTEX (benzene, toluene, ethylbenzene and xylenes) are widespread contaminants of groundwater and soils. They have been widely used in industrial processes and are constituents of fossil fuels. These volatile and water soluble chemicals may be dispersed in the environment because of accidents and storage tank leakages (Pawlowski, 1998); they represent a threat to human health and ecosystems because of their toxicity (ATSDR, 2004). According to WHO, xylene content limit is 500 lg L1 in drinking water. In situ chemical oxidation (ISCO) is an increasingly popular method for the remediation of groundwater and soils contaminated by organic chemicals such as chlorinated VOCs (volatile organic compounds), MTBE (methyl tert-butyl ether), BTEX, PAHs (polycyclic aromatic hydrocarbons), pesticides and explosives (ITRC, 2005; Huling and Pivetz, 2006). Used for two decades in the USA, this technique is still emerging in Europe and mainly used for chlorinated VOCs. ISCO consists of injecting a chemical oxidant
⇑ Corresponding author. Tel.: +33 (0)383 175 260; fax: +33 (0)383 322 975. E-mail address: [email protected] (M.-O. Simonnot). 0045-6535/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2011.06.052
into groundwater or soil to degrade contaminants into less toxic substances. The most frequently used oxidants are potassium or sodium permanganate, hydrogen peroxide combined with Fe(II) (Fenton’s reaction), sodium persulfate, and ozone. Permanganate has received considerable attention and has been frequently used in remediation projects because of its relative ease of use and its persistence in the environment (Bennett, 2002; Li and Schwartz, 2004a,b; Waldemer and Tratnyek, 2006; Heiderscheidt et al., 2008; Crimi et al., 2009; Woo, 2009; Krembs et al., 2010). Fenton’s reaction has been extensively studied as well (Neyens and Baeyens, 2003; Huling and Pivetz, 2006; Crimi and Taylor, 2007; De Souza e Silva et al., 2009; Sutton et al., 2011). Hydroxyl radicals generated by this reaction are very powerful (redox potential E° = 2.7 V) but their persistence is rather low. The use of concentrated hydrogen peroxide solutions may be hazardous, thus for safety reasons, solid reagents such as sodium percarbonate, slowly releasing hydrogen peroxide, are of interest (Birnstingl et al., 2006; Ruffing et al., 2009). Persulfate, a relatively new ISCO reagent, is a strong and very persistent oxidant (E° = 2.0 V). Its chemical activation with Fe(II) leads to the generation of sulfate radical SO 4 comparable to hydroxyl radical (Liang and Lee, 2008; Liang et al., 2008, 2009, 2011; Tsitonaki et al., 2010). Hydrogen peroxide and sodium
1182
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187
persulfate have been compared in different conditions for BTEX destruction (Crimi and Taylor, 2007). An ISCO project begins with a laboratory study aiming at selecting the most appropriate oxidant and measuring parameters such as the Natural Oxidant Demand (NOD). NOD is the mass of oxidant consumed per mass of dry solid by the reaction of a variety of nontarget reduced species (organic matter, reduced forms of metals, etc.) (Mumford et al., 2005). NOD is a key design parameter and one must keep in mind that ISCO is not technically and/or economically feasible at very high NOD. In the present case, we consider an industrial coating site located in the North of France on a chalk plateau contaminated by a wide xylene plume, with concentrations up to saturation in some areas and around 70 mg L1 in the central area of the plume. A hydraulic barrier was installed to stop potential propagation towards municipal drinking water wells as an emergency measure. Then, the issue of determining an efficient remediation technique was addressed and ISCO has been envisioned as a fast and potentially efficient method for mass reduction. The present bench scale study was conducted to qualify suitable technology to be applied to the plant in current activity. Therefore, ozone was discarded for industrial production constraint; Fenton’s reaction was discarded as well for safety reasons and because of the high buffering capacity of the aquifer matrix (chalk). Three oxidants were investigated: permanganate, activated percarbonate and activated/unactivated persulfate. Our objective was to test their efficiency in order to reach xylene degradation in a reasonable time frame. Batch experiments were performed with groundwater and with groundwater-chalk slurries to approximate real site conditions. NOD was measured as well. 2. Materials and methods 2.1. Groundwater and chalk samples Groundwater sample, called GW, was collected at the industrial area of a coating plant located in the Picardie region, in the North of France. Groundwater contamination has been extended on a 100 m2 area, 2 m height at ca 12 m depth in the capillary fringe. It contained about 10 t BTEX, mainly xylenes. The collected sample contained 15.2 mg L1 xylenes, representing 80 wt.% of total BTEX content, but no benzene (Table 1). The major anion was hydrogen carbonate (275 mg L1) and pH was 6.9 (Table 1). Two contaminated chalk samples, called CCh1 and CCh2, were collected in the capillary fringe. Total BTEX content was ca 6 mg kg1 in CCh1 and ca 1 mg kg1 in CCh2 (85% xylenes, no benzene). Non-contaminated chalk samples, called NCCh1 to NCCh3, were collected from the same industrial site at different depths. 2.2. Chemicals BTEX standard (1 mg L1) and sulfuric acid (95–97 wt.%) were supplied by Sigma–Aldrich, sodium percarbonate (OXIPER-S131) Table 1 Groundwater characteristics. Major ions
2+
Ca Mg2+ SO2 4 HCO 3 Cl
BTEX
mg L1
mM
180.0 12.4 2.76
4.50 0.52 0.03
Benzene Toluene Ethylbenzene
<0.2 1.16 2.75
<2.6 12.6 26.0
275.0 11.3
4.50 0.32
o-xylene m-xylene p-xylene Total xylene Total BTEX
1.71 11.40 2.13 15.24 19.35
16.1 107.6 20.1 143.8 191.5
Na+, K+, NO 3 not measured pH = 6.9
mg L1
lM
by Solvay Chemicals, potassium persulfate (>99 wt.%) by Acros Organics, ortho-phosphoric acid (85 wt.%) by Merck, potassium permanganate (>99 wt.%) and sodium thiosulfate (>99 wt.%) by Fluka Chemika, hydrogen peroxide (50 wt.%), ferrous sulfate heptahydrate Rectapur and sodium carbonate Normapur by VWR Prolabo. 2.3. Oxidation experiments Batch experiments were performed at the bench scale to select the most appropriate one for this project. Oxidation experiments were run with site groundwater and with chalk with the three oxidants. NOD was measured with permanganate and persulfate. 2.3.1. Groundwater oxidation In the first series, four preliminary experiments were run with 200 mL of GW in 500 mL capped flasks (Table 2, exp. #1–4). Oxidants were potassium permanganate, potassium persulfate activated by Fe(II) and sodium percarbonate, the molar ratio [oxidant]:[BTEX] was 100:1. Flasks were shaken at 150 rpm on a rotating table for 160 h. In the second series, eight experiments were run in stirred 500 mL triple neck flasks with temperature and pH monitoring for 48 h (Table 2, exp. #5–12). Experiments #5 and 6 used permanganate and persulfate at a molar ratio [oxidant]:[BTEX] of 10:1, 10 times lower than previously. Five experiments (#7–11) were run with percarbonate activated by Fe(II) at different [oxidant]:[catalyst]:[BTEX] molar ratios (Table 2). In experiment #11, Fe(II) was introduced in 10 sequences with a time interval of 10 min between them. A control experiment was performed as well (exp. #12). Both series were carried out at 20 °C, the flasks being covered in aluminum foil to prevent photocatalytic degradation of BTEX. Samples of GW (5 mL) were periodically taken; they were placed in 20 mL closed sampling vessel with 4 g sodium carbonate and 1 g sodium thiosulfate to stop oxidation. BTEX concentration was measured. For the second series, residual permanganate or hydrogen peroxide was analysed. 2.3.2. Contaminated chalk oxidation Oxidation was run with the contaminated chalk sample CCh1 with potassium permanganate, activated potassium persulfate and activated sodium percarbonate (Table 2, exp. #13–15) at 20 °C. Chalk samples (250 g) and ultra pure water (125 mL) were placed in 500 mL capped flasks and stirred at 150 rpm. During experiment #13, potassium permanganate was introduced at the molar ratio [KMnO4]:[BTEX] 10:1; after 144 h, a second addition was done to reach the global ratio 110:1. The same procedure was followed with activated potassium persulfate, at the initial molar ratio [persulfate]:[Fe2+]:[BTEX] 10:10:1 and, after 144 h, at the global ratio 110:110:1. Activated sodium percarbonate was directly introduced at the molar ratio [percarbonate]:[Fe2+]:[BTEX] 100:100:1. Additional experiments were performed with potassium persulfate without activation (Table 2, exp. #16). Chalk samples CCh2 (100 g) and ultra pure water (50 mL) were placed in 250 mL flasks and stirred at 150 rpm at 20 °C. Potassium persulfate was introduced at the molar ratio [persulfate]:[BTEX] 140:1. In all experiments, 5 mL samples were periodically taken and analysed. 2.3.3. NOD NOD was measured with three samples of non contaminated chalk (NCCh1 to NCCh3) (Frasco et al., 2005). Each chalk sample was divided into four or five samples of 50 g that were introduced in 250 mL flasks with hermetic capping. The first one enabled us to measure humidity, the three other ones to measure NOD after 48 h at three initial concentrations. The fifth one was used to measure long term NOD.
1183
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187 Table 2 Conditions of the oxidation experiments performed with contaminated groundwater (#1–12) and chalk (#13–16). Experimental conditions
Monitored parameters 2+
#
Oxidant
Molar ratio [Ox]:[Fe ]:[BTEX]
Ratio to SMR
Oxidant mg L1 (1–12)
FeSO47H2O mg kg1 (13–16)
BTEX
Oxidant
Temperature
pH
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
KMnO4 K2S2O8 2Na2CO33H2O2 – KMnO4 K2S2O8 2Na2CO33H2O2 2Na2CO33H2O2 2Na2CO33H2O2 2Na2CO33H2O2 2Na2CO33H2O2 – KMnO4 K2S2O8 2Na2CO33H2O2 K2S2O8
100:0:1 100:100:1 100:100:1 – 10:0:1 10:10:1 10:10:1 50:50:1 25:100:1 10:100:1 10:10:1 – 110:0:1 110:110:1 100:100:1 140:0:1
7.2 4.8 14.5
2370 4050 4710 – 237 405 471 2355 1178 471 471 – 1043 1784 1884 420
– 4170 4170 – – 417 417 2085 4170 4170 417 – – 1835 1668 –
x x x x x x x x x x x x x x x x
– – – – x – x x x x x – x – x –
– – – – x x x x x x x x – – – –
– – – – x x x x x x x x – – – –
0.72 0.48 1.45 7.25 3.62 1.45 1.45 7.92 5.28 14.5 6.72
Ultra pure water and potassium permanganate or persulfate (3 wt.%) were added into the flask and chalk was added to reach a mass ratio (water):(dry chalk) of 3 (Table 3). Control experiments were done as well. Flasks were capped and kept at ambient temperature (20 °C) in the dark. They were rotated at the beginning of the experiment and then twice a day. At the end, supernatant was sampled and residual oxidant concentration was measured, enabling us to calculate NOD.
2.4. Analytical methods 2.4.1. BTEX analysis BTEX were analyzed following ISO 11423-1 procedure. Head space method consisted of heating at 80 °C for 1 h a given volume of non filtered water in a 20 mL glass vial fitted with crimped aluminum caps lined with PTFE-coated butyl rubber septa. An aliquot
of the gas phase was then taken and injected into a GC equipped with a FID. We used a Varian 3900 GC with a capillary column DB-WAX (30 m 0.54 mm 0.85 lm, Grace). Injector and detector temperature was 250 °C; oven temperature was programmed to increase from 70 to 115 °C at 5 °C min1. The carrier gas was helium, the flow rate was 3.9 mL min1, and the split ratio was 10%. BTEX analysis was run in 8.2 min. External standards of benzene, toluene, ethylbenzene, o-, m- and p-xylene were used. Standards and samples were prepared according to the ISO 11423-1 procedure. Septum glass vials of 20 mL were used in which 5 mL sample was introduced with 4 g sodium carbonate and, if necessary, 1 g sodium thiosulfate to stop oxidation reaction. 2.4.2. Residual oxidant analysis Residual oxidants were analysed by conventional titration. Permanganate was titrated by hydrogen peroxide (Brumblay, 1971)
Table 3 Experimental conditions for NOD determination. #
Sample
mwet (g)
mdry (g)
Potassium permanganate 1 NCCh1 2 NCCh1 3 NCCh1 4 NCCh1 5 NCCh2 6 NCCh2 7 NCCh2 8 NCCh3 9 NCCh3 10 NCCh3 11 – 12 – 13 –
50.2 50.5 50.9 51.6 51.8 51.5 50.4 50.9 51.3 50.4 – – –
38.8 39.0 39.3 39.9 35.9 35.6 34.9 39.1 39.5 38.8 – – –
Potassium persulfate 14 NCCh1 15 NCCh1 16 NCCh1 17 NCCh1 18 NCCh2 19 NCCh2 20 NCCh2 21 NCCh3 22 NCCh3 23 NCCh3 24 – 25 – 26 –
50.6 50.7 50.8 50.6 50.1 50.2 50.1 50.7 50.8 50.0 – – –
39.1 39.2 39.3 39.1 34.7 34.8 34.7 39.0 39.1 38.5 – – –
Vox (mL)
Vwater (mL)
Duration
mwater/mdry
Coxidant (g kgdry )
5 25 50 50 5 25 50 5 25 50 5 25 50
95 75 50 50 95 75 50 95 75 50 95 75 50
48 h 48 h 48 h 21 d 48 h 48 h 48 h 48 h 48 h 48 h 48 h/6 wk 48 h/6 wk 48 h/6 wk
2.9 2.9 2.8 2.8 3.2 3.3 3.3 2.9 2.8 2.9 – – –
4 19 38 38 4 21 43 4 19 39 – – –
10 50 100 100 10 50 100 10 50 100 10 50 100
90 50 0 0 90 50 0 90 50 0 90 50 0
48 h 48 h 48 h 13 d 48 h 48 h 48 h 48 h 48 h 48 h 48 h/6 wk 48 h/6 wk 48 h/6 wk
2.9 2.8 2.8 2.9 3.3 3.3 3.3 2.9 2.9 2.9 – – –
6 32 64 64 7 36 72 6 32 65 – – –
1
1184
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187
and hydrogen peroxide by permanganate (hydrogen peroxide is also a reducing agent). Persulfate was measured by indirect titration. Persulfate sample was acidified with sulfuric and phosphoric acids; Fe(II) added in excess was titrated by permanganate. 3. Results and discussion 3.1. BTEX oxidation in GW 3.1.1. With permanganate and activated persulfate BTEX oxidation in GW was carried out with permanganate and activated persulfate as outlined in Table 2. To compare our results, we expressed the oxidant quantities in terms of stoichiometric molar ratio (SMR). We defined this ratio as the number of moles of oxidant required to degrade one mole of a target compound to the final products CO2 and H2O. Stoichiometric balances were written for toluene, ethylbenzene and xylenes. For instance, xylene oxidation by permanganate yielded:
14MnO4 þ C8 H10 þ 2H2 O ! 8CO2 þ 14OH þ 14MnO2
ð1Þ
Then SMR [KMnO4]:[Xylenes] was equal to 14. The global SMR [KMnO4]:[BTEX] was calculated by weighing by BTEX composition (Table 1), it was equal to 13.8, close to 14 since xylenes were the major contaminants. BTEX oxidation by persulfate radicals obeys a series of radical reactions that are not accurately known. Following the same assumptions as previously, the stoichiometric balance of xylene oxidation yielded: 2 þ 21S2 O2 8 þ C8 H10 þ 16H2 O ! 8CO2 þ 42H þ 42SO4
ð2Þ
SMR for xylene oxidation by activated persulfate was 21 and 20.7 for BTEX oxidation. All our experiments were run with an excess of oxidants compared to BTEX, at molar ratios [Oxidant]:[BTEX] between 0.5-fold and 15-fold the SMR (Table 2). With activated persulfate, the molar ratio [K2S2O8]:[Fe2+] was set to unity as in other contributions (Liang et al., 2008). Residual BTEX ratios were plotted against time for oxidation by permanganate and persulfate (Fig. 1). At high permanganate concentration (Table 2, exp. #1), BTEX were slowly degraded and reached a residual concentration of 3% after 114 h (ca 5 d) (Fig. 1). At high persulfate concentration (Table 2, exp. #2), complete degradation was attained in less than 15 h (Fig. 1). Compared to chlorinated compounds, BTEX degradation is much slower. Indeed, trichloroethylene oxidation with persulfate occurred in a few minutes (Liang et al., 2004a,b). With both oxidants, BTEX degradation could be fairly well represented by pseudo-first order rates with effective constants of 0.06 h1 for permanganate and 2.8 h1 for activated persulfate. Residual permanganate analysis (exp. #5) showed that the molar ratio [KMnO4 consumed]:[BTEX degraded] was equal to 20 close to the SMR equal to 14 (Eq. 1). Then permanganate was almost totally consumed by BTEX oxidation, suggesting a low depletion by oxidation of non-targeted compounds. The final objective of this research was to select effective conditions for ISCO. The above conditions enabled us to reach this goal but we wanted to use lower oxidant concentrations to reduce costs. We used the molar ratios [MnO 4 ]:[BTEX] 10:1 (Table 2, exp. #5) and [K2S2O8]:[Fe2+]:[BTEX] 10:10:1 (Table 2, exp. #6) that were lower than SMR. Results are plotted in Fig. 1. A slow and incomplete degradation of BTEX was observed: ca 60% residual BTEX were measured after 160 h with permanganate (exp. #5) and ca 40% with persulfate (exp. #6). These results clearly showed that these concentrations were not sufficient. Intermediate concentrations should be tested to complete this study.
At this level, we can conclude that both oxidants were effective at high concentrations. Permanganate was expected to work well since there was no benzene. Benzene oxidation would have been much slower. However, significant precipitations occurred. Permanganate oxidation produced a brown precipitate of manganese dioxide. When formed in large amounts, manganese dioxide particles may be a drawback since they deposit in the subsurface and may reduce permeability (Sirguey et al., 2008). Anyway recent researches have proved that MnO2 particle growth and transport could be controlled and suspensions stabilized (Crimi and Ko, 2009). Similarly, activated persulfate induced iron carbonate (siderite) precipitation (solutions turned quickly to dark orange). Speciation in solution was calculated with the computer code CHESS (van der Lee and de Windt, 2000). Results predicted siderite precipitation and the related pH variations. In this case, precipitation involved Fe(II), suppressing activation. The question of activation necessity may be addressed since oxidation succeeded despite Fe(II) depletion caused by precipitation. 3.1.2. With activated percarbonate Xylene oxidation by activated percarbonate can be described by:
72Na2 CO3 3H2 O2 þ C8 H10 ! 28Naþ þ 18H2 O þ 16HCO3 þ 6CO2 3 ð3Þ Giving a SMR [percarbonate]:[xylene] of 7 and a SMR [percarbonate]:[BTEX] of 6.9 for this GW. Three experiments were run at the molar ratio [percarbonate]:[BTEX] 10:1 corresponding to 1.4-fold the SMR (Table 2, exp. #7, 10 and 11). Two experiments were performed at the conventional molar ratio [percarbonate]:[Fe2+] 1:1, Fe(II) being added at the beginning (exp. #7) or gradually (exp. #11). Then activation was enhanced by setting the molar ratio [percarbonate]:[Fe2+] at 1:10. Three experiments were run at higher percarbonate concentrations with molar ratios [percarbonate]:[BETX] 25:1 (exp. #9), 50:1 (exp. #8) and 100:1 (exp. #3) at different [percarbonate]:[Fe2+] ratios (Table 2). The evolution of residual BTEX is plotted in Fig. 2. We observed a rather poor degradation rate without activation despite a high percarbonate concentration equal to 14.5-fold the SMR (exp. #3) which was attributed to the absence of catalyst. The high concentration of percarbonate may be a drawback as well since high concentrations of hydroxyl radicals lead to radical scavenging (ITRC, 2005; Huling and Pivetz, 2006). At the conventional molar ratio [percarbonate]:[Fe2+] 1:1 (exp. #7 and 11), degradation was low, ca 50% of initial BTEX were still present after 48 h (Fig. 2). Gradual Fe(II) addition (exp. #11) had no effect compared to direct addition (exp. #7). Increasing 5-fold both oxidant and catalyst concentrations (exp. #8) did not lead to better results: ca 20% of initial BTEX were still detected after 20 h. Raising activation 10-fold while maintaining [percarbonate]:[BETX] ratio 10:1 (exp. #10) had no effect, ca 18% of initial BTEX were present after 40 h. The last conditions, consisting of setting [percarbonate]:[Fe2+]:[BETX] at 25:100:1 led to a complete BTEX degradation in less than 1 h, the major fraction being degraded in 10 min. These experiments enabled us to determine oxidant and catalyst concentrations to degrade BTEX, resulting in high percarbonate concentration (4-fold the SMR given in Eq. 3) and very high Fe(II) concentration. Percarbonate and Fe(II) addition affected GW pH (6.9 initially). Percarbonate addition increased pH because of carbonate addition while Fe(II) addition made it decrease because of iron carbonate precipitation. These results were consistent with the predictions obtained with the speciation modeling program CHESS. In our experiments, pH did not change at the molar ratio [percarbon-
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187
1185
Fig. 1. Evolution of the residual BTEX ratio during batch oxidation in groundwater with potassium permanganate and potassium activated persulfate. [oxidant]:[Fe2+]:[BTEX] were 100:0:1 (s), 10:0:1 (j) for permanganate, 100:100:1 (), 10:10:1 (4) for persulfate and 0:0:1 (x) for control experiment. The dotted lines are construction lines and the curves represent first-order rates for the highest oxidant concentrations.
Fig. 2. Evolution of the residual BTEX ratio during batch oxidation in groundwater with activated sodium percarbonate. [oxidant]:[Fe2+]:[BTEX] were 100:0:1 (}), 50:50:1 (x), 10:10:1 (j), 10:10:1 with gradual supply of Fe2+ (s), 10:100:1 (), 25:100:1 (4) and 0:0:1 () for control experiment. Dotted lines are construction lines.
ate]:[Fe2+]:[BTEX] 10:10:1 (exp. #7 and 11). It increased to 7.6 at 50:50:1 (exp. #8) and, at high Fe(II) concentrations, it drastically decreased to 5.3 at 10:100:1 (exp. #10) and 3.9 at 25:100:1 (exp. #11). Again precipitation occurred as with persulfate, solutions turned from colorless to dark orange. Dramatic pH decrease and precipitation could be a major drawback for ISCO feasibility at the field scale.
3.1.3. Conclusions on BTEX oxidation in GW Considering all the experiments run with GW, the three ones leading to the best degradation results were (i) exp. #1 with permanganate at [oxidant]:[Fe2+]:[BTEX] set at 100:0:1, (ii) exp. #2 with persulfate at 100:100:1 and (iii) exp. #9 with percarbonate at 25:100:1. In the former case, manganese dioxide precipitation occurred and fine particles were formed. In the other cases, the
1186
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187
addition of Fe(II) led to the formation of flocs and precipitation of iron carbonate that was confirmed by speciation modelling. Other experiments (not presented here) run with citric acid as an iron chelatant (Tsitonaki et al., 2010) made precipitation decrease but did not improve significantly BTEX degradation. Then as a partial conclusion, in the objective of upscaling, we could chose potassium permanganate because of its persistence in solution as well as the formation of finer particles than iron carbonate precipitates. 3.2. BTEX degradation in chalk samples Oxidation experiments were performed with chalk slurries to approach field reality. They were run with a ratio (solid mass): (solution volume) of 2 gdry chalk mL1. Initial pH was 8.9. In the field, the accessible porosity of chalk was ca 5%. One must keep in mind that laboratory conditions were more favorable to BTEX oxidation than field conditions because of lower (solid mass):(solution volume) ratio, stirring and higher temperature (20 °C instead of ca 10 °C in the aquifer). Besides we used rather low oxidant concentrations in order to ensure feasibility at the field scale. 3.2.1. BTEX oxidation with the 3 oxidants Oxidation experiments with chalk were performed at molar ratios [oxidant]:[Fe2+]:[BTEX] 10:0:1 with permanganate (exp. #13), 10:10:1 with persulfate (exp. #14) and 100:100:1 (exp. #15) with percarbonate. Results are presented in Fig. 3. With permanganate and persulfate, we observed that BTEX residual ratio increased at first. This increase was attributed to BTEX desorption from the solid matrix occurring at a higher rate than oxidation. Exp. #15 with percarbonate was run afterward and we mixed the slurry for 2 h before oxidant addition; for this reason the initial BTEX concentration was higher than that of other. After 144 h (6 d), BTEX concentration was still high with the three oxidants. For permanganate and persulfate, a second addition was done at a higher concentration. The experiment with percarbonate was stopped since it had shown to be ineffective: after 2 d, percarbonate was completely depleted while BTEX were not
degraded. Permanganate led to rapid and almost complete BTEX degradation (exp. #13) and persulfate to complete degradation (exp. #14) after 2 d. Again manganese dioxide or iron carbonate precipitation occurred. Further experiments run with persulfate without activation showed that it was not required. Indeed, complete BTEX degradation was reached after 2 d with the molar ratio [K2S2O8]:[BTEX] 140:1. These results are not plotted for clarity reasons since they were obtained with the chalk sample CCh2 that was less contaminated than CCh1. However, they clearly showed the possibility of decontaminating with persulfate without activation. Our results showed that radical reactions involving hydroxyl radicals released by percarbonate according to Eq. 3 were effective in GW but not in chalk slurries. This result was attributed to the high carbonate concentration in chalk slurries responsible for radical scavenging. Carbonate was initially present in solution because of chalk-water equilibrium (initial pH: 8.9) and was added through percarbonate addition as expressed in Eq. 3. Moreover, Fe(II) failed to activate percarbonate since it was consumed by precipitation reaction. On the contrary, permanganate and persulfate with and without activation by Fe(II) enabled us to attain high BTEX degradation rates in both media (we recall that benzene was absent). Oxidation by permanganate and activated persulfate were accompanied by precipitation that may be a drawback especially with iron carbonate occurring with activated persulfate. This drawback was avoided by using persulfate without activation.
3.2.2. Oxidant persistence Residual permanganate and persulfate concentrations were measured after 48 h and after 6 wk in control experiments (Table 3, exp. #11–13, 24–26) to estimate oxidant loss. It was negligible compared to the amounts that were used (<3% after 6 wk). Permanganate and persulfate were very persistent in our experimental conditions. In field conditions, very low decomposition rates of permanganate and persulfate are expected, which is in agreement with reported data (Brown, 2003; Huling and Pivetz, 2006).
Fig. 3. Evolution of the residual BTEX concentration in chalk-groundwater slurries with permanganate (N), activated persulfate () and activated percarbonate (}). Dotted lines are construction lines.
J. Lemaire et al. / Chemosphere 84 (2011) 1181–1187
3.2.3. NOD NOD was calculated according to the following equation:
NOD ¼
½Oxcontrol V control ½Oxsup V sup m
ð4Þ
where NOD is the Natural Oxidant Demand (g kg1), [Ox]sup the oxidant concentration in the supernatant (g L1), Vsup the supernatant volume (L), [Ox]control the oxidant concentration in the control (g L1), Vcontrol the control solution volume (L), m the mass of dry chalk (kg). NOD was measured after 48 h with three samples of non contaminated chalk with permanganate and with persulfate at three initial concentrations (Table 3). In each case, NOD was lower than 2% of the initial oxidant concentration (not statistically different from 0 according to error bars). The same was observed at longer times, after 13 and 21 d (NOD <5% of the initial oxidant content). The totality of the oxidant was consumed for BTEX degradation. This result was very important in the perspective of upscaling.
4. Conclusions This work showed that permanganate, activated persulfate and activated percarbonate (activated with Fe(II)) enabled us to attain BTEX destruction in the groundwater of a chalky aquifer. Activated percarbonate failed to degrade BTEX in a slurry composed of groundwater and chalk particles. This result was attributed to (i) radical scavenging in carbonated medium and (ii) inefficiency of unactivated percarbonate because of siderite precipitation involving Fe(II) and carbonate. Permanganate and activated persulfate were both effective in this slurry. Moreover, they were persistent for at least 6 wk and NOD was very low. However, both of them induced significant precipitation. Further experiments showed that persulfate could be used without activation, no precipitation occurred. In conclusion, this study enabled us to demonstrate ISCO feasibility to remediate a chalky aquifer contaminated by BTEX (that were mainly xylenes without benzene). Further experiments could be performed but we proved the efficacy of two possible oxidants at the bench scale. Permanganate and unactivated persulfate were selected for pilot tests that were conducted before implementing ISCO at the field scale. References ATSDR, 2004. Interaction Profile for Benzene, Toluene, Ethylbenzene, and Xylenes (BTEX). US Department of Health and Human Services, Public Health Service, Atlanta. Bennett, G.F., 2002. Principles and practices of in situ chemical oxidation using permanganate. Book Review of: Siegrist, R.L., Urynowicz, M.A., West, O.R., Crimi, M.L., Lowe, K.S., Battelle Press, Columbus, OH, USA, 001. J. Hazard. Mater. 90, 323–324. Birnstingl, J., Kelley, B., Koenigsberg, S., 2006. Field results with an alkaline in situ chemical oxidation process. Contam. Reclam. 14, 539. Brown, R.A., 2003. In Situ Chemical Oxidation: Performance, Practice, and Pitfalls. AFCEE Technology Transfer Workshop, San Antonio, Texas. Brumblay, R.U., 1971. Quantitative Analysis – College Outline, second ed. HarperCollins, New York. Crimi, M., Ko, S., 2009. Control of manganese dioxide particles resulting from in situ chemical oxidation using permanganate. Chemosphere 74, 847–853.
1187
Crimi, M.L., Taylor, J., 2007. Experimental evaluation of catalyzed hydrogen peroxide and sodium persulfate for destruction of BTEX contaminants. Soil Sediment Contam. 16, 29–45. Crimi, M., Quickel, M., Ko, S., 2009. Enhanced permanganate in situ chemical oxidation through MnO2 particle stabilization: evaluation in 1-D transport systems. J. Contam. Hydrol. 105, 69–79. De Souza e Silva, P.T., Da Silva, V.L., De Barros Neto, B., Simonnot, M.-O., 2009. Phenanthrene and pyrene oxidation in contaminated soils using Fenton’s reagent. J. Hazard. Mater. 161, 967–973. Frasco, K.A., Osborne, M.W., Vella, P.A., 2005. The permanganate soil oxidant demand: a critical parameter for ISCO remediation. In: The 21st Annual Conference on Soils, Sediments and Water. University of Massachusetts at Amherst, USA, pp. 32–33.