- Email: [email protected]
Mechanisms for naphthalene removal during electrolytic aeration
Water Research 37 (2003) 891–901
Mechanisms for naphthalene removal during electrolytic aeration Ramesh K. Goela, Joseph R.V. Floraa,*, John Ferryb a
Department of Civil and Environmental Engineering, University of South Carolina, Columbia, SC 29208, USA b Department of Chemistry and Biochemistry, University of South Carolina, Columbia, SC 29208, USA
Abstract Batch tests were performed to investigate chemical and physical processes that may result during electrolytic aeration of a contaminated aquifer using naphthalene as a model contaminant. Naphthalene degradation of 58–66% took place electrolytically and occurred at the same rates at a pH of 4 and 7. 1,4-naphthoquinone was identified as a product of the electrolysis. Stripping due to gases produced at the electrodes did not result in any naphthalene loss. Hydrogen peroxide (which may be produced at the cathode) did not have any effect on naphthalene, but the addition of ferrous iron (which may be present in aquifers) resulted in 67–99% disappearance of naphthalene. Chlorine (which may be produced from the anodic oxidation of chloride) can effectively degrade naphthalene at pH of 4, but not at a pH of 7. Mono-, di- and poly chloronaphthalenes were identified as oxidation products. Ferric iron coagulation (due to the oxidation of ferrous iron) did not significantly contribute to naphthalene loss. Overall, electrolytic oxidation and chemical oxidation due to the electrolytic by-products formed are significant abiotic processes that could occur and should be accounted for if bioremediation of PAH-contaminated sites via electrolytic aeration is considered. Possible undesirable products such as chlorinated compounds may be formed when significant amounts of chlorides are present. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Naphthalene; Electrolysis; Aeration; Gas stripping; Hydrogen peroxide; Fenton’s reagents; Chlorine; Oxidation; Coagulation
1. Introduction Remediation of contaminated soil and groundwater is an important issue that is of concern among environmentalists and hydrogeologists. In 1977, the United States Environmental Protection Agency reported that a minimum of 17 million waste disposal facilities were discharging 6.5 billion m3 of liquid waste into the ground each year [1]. Gasoline and petroleum derived hydrocarbons from leaking underground storage tanks (UST) are another important contributor to groundwater contamination. There are more than 2 million UST systems located at about 700 thousand facilities existing in the United States [2]. Approximately 85% of all underground storage tanks are made of steel and about *Corresponding author. Tel.: +1-803-777-3614; fax: +1803-777-0670. E-mail address: fl[email protected] (J.R.V. Flora).
25% of the existing USTs fail tightness testing and may be leaking [3]. Remediation methods to address soil and groundwater contamination can be divided into two categories: in situ and ex situ methods. In situ remediation approaches include aeration to induce biodegradation, the use of capture zones, and oxidation through chemical injection [4]. Pump and treat and soil vapor extraction are two widely used ex situ remediation approaches. In situ remediation approaches are excellent alternatives for treating contaminated aquifers because they are more cost effective than ex situ remediation processes. The pump and treat approach has frequently been proven inefficient due to slow desorption, diffusion, and dissolution processes of contaminants [5]. Aeration is an in situ remediation approach that was developed in the late 1980s for treating aquifers contaminated with volatile organic compounds (VOCs) [6]. During aeration, the injection of air in the saturated
0043-1354/03/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 3 7 6 - 7
892
R.K. Goel et al. / Water Research 37 (2003) 891–901
zone of the aquifer physically removes dissolved and adsorbed VOCs and transfers oxygen into the groundwater. As a result, both physical removal and aerobic biodegradation of contaminants by indigenous organisms are enhanced [6,7]. Aeration can further be subdivided into two categories: in situ air sparging and in-well aeration (IWA) [7]. In IWA, air is injected into the well below the groundwater table. This aerated water flows to the rest of aquifer, either by the natural flow of groundwater or by pumping. Electrolysis of groundwater is one method for IWA. In electrolytic IWA, a small potential difference is applied across the cathode and anode of an electrolytic cell inserted in a well. The anodic oxidation of H2O then produces O2 in situ. Some of the applications of the electrolysis of water in environmental engineering include direct abiotic oxidation of contaminants [8–14] or disinfection of drinking water [15]. Furthermore, electrolytic reactions have been used to simulate biological activity such as denitrification [16–17], dechlorination [18], or to increase the nitrogen removal efficiency of biofilter processes [19]. Electrolytic oxygen generation for groundwater aeration and some of the side reactions that result from the process have been investigated and discussed by Franz et al. [20]. The side reactions included chlorine generation, hydrogen peroxide generation, and a drop in pH during laboratory and field-scale investigations. During a 77-day in-well aeration test using a similar electrode assembly at their field site, naphthalene levels decreased from a background concentration of 1.1–0.2 mg/l using a current density of 2.5 mA/cm2, chlorine was not detected, the pH dropped 2.4 units to a value of 3.8, ferrous iron ranged from 24 to 36 mg/l, and a possible upper limit of 1.5 mg/l hydrogen peroxide was reported in the injection well. Their results were used to determine the mechanisms that needed to be studied for contaminant removal during electrolytic groundwater remediation. The specific objectives of this research were to evaluate the mechanisms for contaminant removal that can occur during electrolytic aeration of contaminated groundwater. Of primary concern was the possibility that the anodic oxidation of chloride containing groundwater might lead to the formation of HOCl and cathodic reduction of oxygen might produce hydrogen peroxide, both of which could engage in the direct oxidation of contaminants. Contaminant loss can also occur due to direct electrolytic transformation, stripping due to gases formed at the electrodes, indirect reactions due to Fenton’s type chemistry between native ferrous iron and hydrogen peroxide, and coagulation from the ferric iron produced due to oxidation of native ferrous iron. This paper presents an evaluation of these side reactions using naphthalene as a model contaminant. Naphthalene is a polynuclear aromatic hydrocarbon that has
been found in numerous contaminated sites [21,22] and is of particular concern in the site reported in Franz et al. [20]. The experiments performed in this research were designed to complement the study of Franz et al. [20].
2. Materials and methods 2.1. Naphthalene test solution Naphthalene (Fisher Scientific, Fairlawn, NJ) was added to deionized water and stirred for 12–14 h to prepare a saturated solution. This stock solution was filtered and diluted to 9.871.5 mg/l with deionized H2O. 1.42 g/l of sodium sulfate (Fisher Scientific, Fair Lawn, NJ) was added to the diluted naphthalene solution to solution, simulating sulfate make a 0.01 M SO2 4 concentrations measured in the field site reported by Franz et al. [20]. Since it was expected that ionic strength (which was dominated by sulfate ions) would be critical during electrolysis, it was decided to consistently add sulfate in all the experiments. With the exception of gas stripping experiments, all the tests were performed at a pH of 470.2 and 770.2. The pH was adjusted using either concentrated sulfuric acid or 0.25 N NaOH. Neutral pH was used for the gas stripping experiments. 2.2. Electrolytic degradation experiments The electrode assembly in the present study was identical to the assembly used in Franz et al. [20]. The electrode assembly consisted of a stainless steel plate cathode (Type 304, #3 polish, 0.07 cm thickness (Metal Supermarket, Columbia, SC)) and a titanium anode with a mixed metal oxide coating (Type EC-600, 0.15 cm. thick, mesh anode (Eltec Systems Corp., Chardon, OH)) held together by nylon screws. The electrodes were 3.18 cm 6.35 cm separated by a distance of 1.4 cm. Titanium screws and nuts were used as the current collector on both the anode and the cathode to minimize the reactivity of the current collectors and to reduce the possibility of corrosion. The current collectors on the cathode and anode were connected to copper wires, which were connected directly to a power supply unit (HP Model E3612A, Agilent Technologies, Inc., Englewood, CO). The connection between the copper wires and titanium screws at the cathode and anode were sealed using a fast drying epoxy and coated with a silicon sealant. The reactor vessels used for the electrolytic experiments were 500 ml amber-colored bottles. Holes were drilled through the caps to run the wires connected to the electrode assembly. The gap between the copper wires running through the cap was sealed with an epoxy sealant. One 12 mm hole was drilled on the cap of all the bottles, including the experimental control, to keep the
R.K. Goel et al. / Water Research 37 (2003) 891–901
system under atmospheric pressure. Three liters of the naphthalene test solution was prepared and divided into two batches, with the pH of a batch adjusted to either 4 or 7. Five hundred ml of each solution was transferred to separate 500 ml amber-colored bottles. Duplicate tests were performed with passed currents of 25, 50, 100, and 200 mA, corresponding to current densities of 1.2, 2.5, 5.0, and 9.9 mA/cm2. An experimental control (i.e., no current) was performed using a reactor vessel containing an electrode assembly. The contents of the bottles were stirred continuously using a magnetic stirrer bar and plate. Samples were withdrawn directly from the bottles by opening the cap. The total organic carbon (TOC) and naphthalene concentrations of the samples were quantified. Samples for GC/MS analysis were withdrawn at the end of each experiment. 2.3. Gas stripping experiments Experiments were performed to investigate the contribution of gas stripping to the disappearance of naphthalene observed during the electrolytic experiments. Using Faraday’s law and ideal gas law, the volume of gas produced from oxygen generation at the anode and hydrogen generation at the cathode would be 1.6 l/day for a passed current of 100 mA. Stripping experiments corresponding to the gas flowrates for the four current intensities used in electrolytic experiments were performed. A 12.7 cm long stainless steel needle for gas delivery was run through a hole drilled through the cap of a 500 ml amber-colored bottle, with 6.4 cm of the needle penetrating through the cap of the bottle into the naphthalene solution. The gap between the needle and the cap was sealed with an epoxy sealant. One 12 mm hole was drilled on the cap of all the bottles, including the experimental control, to keep the system under atmospheric pressure. A 1-l naphthalene test solution was prepared without pH adjustment. 500 ml were transferred to two 500 ml amber-colored bottles and were stirred using a magnetic stirrer bar and plate. Scientific grade nitrogen gas was bubbled in the naphthalene solution in one bottle with the gas flow controlled using a regulator attached to a gas cylinder and an intermediate gas valve. The gas flow was quantified using a Riteflow flow meter with plain ends (Model 150 mmSZ1, Bel-art Products, Pequannock, NJ) calibrated using a wet-tip gas flow meter. Samples for naphthalene and TOC analysis were withdrawn in a similar manner as the electrolytic experiments. 2.4. Experiments with chlorine The chlorination experiments were performed under identical conditions to the electrolysis experiments. 1.2 ml of 5.25% by weight sodium hypochlorite solution
893
was added to each liter of the two test solutions (at pH 4 and 7) to give measured total chlorine concentrations of 35 mg/l in each solution. The pH of the solution was checked after chlorine addition and adjusted immediately to 4 or 7. Periodic samples were withdrawn from the bottles for naphthalene, TOC, and chlorine analysis. Samples for GC/MS analysis were withdrawn at the end of each experiment. 2.5. Experiments with fenton’s reagents and hydrogen peroxide Fenton’s experiments were carried with hydrogen peroxide concentrations of 2–3 and 10–12 mg/l and a ferrous iron concentration of 30 mg/l. Experiments with only hydrogen peroxide were also performed at a concentration of 10–12 mg/l. The experimental conditions were the same as in the electrolysis experiments. To get the desired hydrogen peroxide concentrations, 30% by weight hydrogen peroxide solution (Aldrich Chemical Company Inc, Milwaukee, WI) was used. Ferrous iron was added using a stock solution prepared by adding 1.49 g FeSO4 7H2O in 100 ml DI water at a pH of 3. Periodic samples were withdrawn from the bottles for naphthalene and TOC analysis. Samples for GC/MS analysis were withdrawn at the end of each experiment. 2.6. Coagulation with ferric iron Standard jar tests were performed using a Phipps and Bird jar test apparatus. Six liters of a naphthalene test solution was prepared and adjusted to either a pH of 4 or 7. Nine hundred ml of the solution were transferred to six 1-l glass beakers, which were open to the atmosphere. Varying amounts (0–25 ml) of a FeCl3 stock solution were added to produce Fe3+ doses ranging from 0–50 mg/l. FeCl3 stock solution was prepared by adding 1.74 g FeCl3 6H2O in 200 ml deionized water. Pre-determined amounts of NaOH were added simultaneously with the coagulant to maintain the pH at appropriate levels. After chemical addition, the solutions in the beakers were mixed at 150 RPM for 1 min, after which the mixing speed was reduced to 25 RPM for 30 min. Mixing was then stopped and quiescent settling was allowed to occur for 30 min prior to sample withdrawal for naphthalene and TOC analysis. 2.7. Analytical methods Naphthalene was quantified using a Varian Model 3380 gas chromatograph equipped with a flame ionization detector, Chrompack capillary column (Select 624 CB Df 1.8 mm, FS 30 0.32 mm ID), and an auto-sampler with a 100 mm PDMS coated SPME fiber assembly (Supelco, Bellefonte, PA). The sample
894
R.K. Goel et al. / Water Research 37 (2003) 891–901
adsorption time with the SPME fiber was 10 min in agitate mode and the desorption time was 2 min followed by a 1 min waiting period. The analysis was performed in splitless mode with an injection temperature of 2501C, isothermal oven temperature of 1801C, and detector temperature of 2751C. TOC was quantified using a Shimadzu TOC 5000A analyzer equipped with an autosampler. The total chlorine and ferrous iron concentrations were measured using Hach Methods 8167 and 8146, respectively, with a Hach DR2000 spectrophotometer [23]. Hydrogen peroxide was measured using CHEMets Kit K-5510 (Chemetrics, Inc., Calverton, VA). The pH was measured using an Ag/AgCl combination pH electrode attached to a pH meter (Ion Analyzer 220, Corning, Inc., Corning, NY). GC/MS analysis was performed on the samples at the end of experiments where naphthalene degradation was noted. The samples were extracted with hexane and approximately 25 ml of the extracted organic solution was concentrated in Rotavapor R 110 (BUCHI, Slawil, Switzerland) up to a capacity of 1–2 ml. This concentrated solution was analyzed using GC-IT (ion-trap) MS (Varian 3800 model GC equipped with a Varian 8200 autoinjector and a Varian Saturn 2000 ITMS). The analytical column was a 30 m DB-5 with a 25 mm-film thickness. The injector port was set for splitless operation at 2201C and was ramped to 2601C for 1 min after sample injection (autoinjector volume=1 ml).
3. Results and discussion 3.1. Electrolytic naphthalene degradation and gas stripping Naphthalene solutions were electrolyzed over a range of current densities. Under our conditions, naphthalene degradation rates were insensitive to current density and independent of system pH (see Figs. 1a and b). The zerocurrent control showed some naphthalene loss, probably due to volatilization or adsorption on the inactive electrode surface. Corresponding TOC values for all current intensities, including the control, are plotted in Figs. 2a and b, which shows that TOC values are constant, except for the decrease with 200 mA current after 2 h. However, these differences appear to be within the operational sensitivity of the instrument. Constant TOC values indicate little or no naphthalene mineralization during electrolysis. The gas stripping experiments showed some losses of naphthalene in the control experiment over a 9 h period (see Fig. 3). The addition of nitrogen gas supports the conclusion that the formation of gases on the electrodes does not contribute to significant losses of naphthalene due to volatilization under the conditions of the study.
Gas stripping may be a more significant mechanism for more volatile compounds. Electro-coagulation, electro-flotation and electro-oxidation are three processes that can contribute to contaminant loss in electrochemical systems [24]. In the present experiments, electro-coagulation and electrofloatation did not occur as turbidity or settled solids were not present. Anodic oxidation of aromatics have been reported by various researchers [25,26] where contaminants are either directly transformed on the anode into CO2 and water or into some other intermediate product, or contaminants are indirectly oxidized through electrocatalytic process [27]. Although electrocatalysts were not added in this study, water can undergo direct anodic oxidation to yield hydroxyl radicals [28], the presence of which has been reported by various authors during the electrolytic degradation of contaminants (e.g., [11,13,14,29]). For example, Panizza et al. [13] performed experiments on the anodic oxidation of 2-naphthol where they attributed the 2naphthol degradation to hydroxyl radicals formed in their system. Furthermore, hydroxyl radicals may be involved in the formation of ozone, another strong oxidant [28]. Under our experimental conditions, constant TOC values rule out the possibility of naphthalene transformation into CO2 and water due to direct anodic oxidation. Hence, naphthalene disappearance could be due to its chemical conversion into other intermediate products because of hydroxyl radicals formed during electrolysis. GC/MS analysis on samples withdrawn at the end of the electrolytic experiments confirmed the presence of at least one reaction product, 1,4-naphthoquinone. The presence of this compound is consistent with those found by other researchers who also observed aromatic quinones during the electrolytic oxidation of aromatic hydrocarbons (e.g., [11,13,26]). Hydroxyl radicals can directly attack aromatic rings by hydroxylation followed by oxidation [30]. Although hydroxyl radicals are known to produce quinones during the oxidation of aromatics, other oxidants may also mediate this transformation. Hence, product analysis cannot rule out other compounding pathways, including the direct anodic oxidation of naphthalene. Naphthalene degradation in the present electrolytic experiments could be due to direct anodic oxidation, oxidation by an intermediate (e.g., hydroxyl radicals), or a combination of both. 3.2. Naphthalene degradation by aqueous chlorine Batch experiments to evaluate the reactivity of chlorine with naphthalene were conducted with 35 mg/l total chlorine at a pH of 4 and 7. These experiments were performed because electrolysis of chloride–containing water produces chlorine at the anode, which subsequently hydrolyzes to hypochlorous acid
R.K. Goel et al. / Water Research 37 (2003) 891–901
895
Normalized Naphthalene Concentration, C/Co
1.2
pH=4 1.0
0.8
0.6
0.4 Blank 25 mA 50 mA 100 mA 200 mA
0.2
0.0 0
2
4
6
8
10
Time, hours
(a)
Normalized Naphthalene Concentration, C/Co
1.2
pH=7 1.0
0.8
0.6
0.4 Blank 25 mA 50 mA 100 mA 200 mA
0.2
0.0 0
2
4
(b)
6
8
10
Time, hours
Fig. 1. Naphthalene concentration normalized to the initial concentration during electrolysis at (a) a pH of 4 and (b) a pH of 7.
[8,13,24,31,32]. Franz et al. [20] found total chlorine to be as high as 43 mg/l in laboratory column studies evaluating potential side reactions during electrolytic aeration of groundwater. Furthermore, chloride concentrations as high as 110 mg/l were also observed in other wells at the field site where electrolytic aeration of contaminated groundwater is being tested. Fig. 4a shows a 98% decrease in naphthalene concentration in the presence of chlorine at a pH of 4 and a negligible change in naphthalene at a pH of 7. Corresponding measured total chlorine concentrations fell by 33% at a pH of 4, but the chlorine concentration
remained almost constant at a pH of 7 (see Fig. 4b). Based on the following equilibria [33]: Cl2ðaqÞ þ H2 O"HOCl þ Hþ þ Cl HOCl"Hþ þ OCl
ðlog K ¼ 3:3Þ;ð1Þ
ðlog K ¼ 7:3Þ:
ð2Þ
The primary chlorine species at a pH of 7 are HOCl and OCl–, and Fig. 4a indicates that these species do not react with naphthalene. Although the HOCl and OCl– species would still predominate at a pH of 4.0 at the low levels of chloride in the experiments, the Cl2(aq) concentrations at a pH of 4 is 1500 times greater than
R.K. Goel et al. / Water Research 37 (2003) 891–901
896
Normalized TOC Concentration, C/Co
1.2
1.0
0.8
0.6
0.4 Blank 25 mA 50 mA 100 mA 200 mA
0.2
pH=4
0.0 0
2
4
(a)
6
8
10
Time, hours
Normalized TOC Concentration, C/Co
1.2
1.0
0.8
0.6
0.4 Blank 25 mA 50 mA 100 mA 200 mA
0.2
pH=7
0.0 0
2
4
(b)
6
8
10
Time, hours
Fig. 2. Total organic concentration normalized to the initial concentration during electrolysis at (a) a pH of 4 and (b) a pH of 7.
the Cl2(aq) levels at a pH of 7 (with the actual concentration dependent on the chloride levels). Thus, it is possible that Cl2(aq) is the species that reacts with naphthalene, which is consistent with the observed faster kinetics at low pH values. The TOC was essentially constant throughout the experiments in both cases, indicating no mineralization of naphthalene. GC/MS analysis of samples at the end of the chlorination experiments at a pH of 4 showed monochlorinated (1-chloro and 2-chloronaphthalene) and various congeners of di- and tri-chloronaphthalenes. Chlorinated naphthoquinones were also identified.
These results are consistent with studies reported in the literature [34–36]. However, in the above-mentioned literature, addition of chlorine atoms at one or more positions on the naphthalene molecule took place either in the presence of catalysts like FeCl3 [35], in the gas phase [34], or by radiolysis in a chlorinated solvent [36]. Reports on the direct chlorination of naphthalene by aqueous chlorine alone are limited. Reinhard et al. [37] added a total free chlorine concentration of 100 mg/l to the aromatic fraction of diesel fuel (with naphthalene as one of the constituents). After 30 h of chlorination, all naphthalene had disappeared, but no chlorinated
R.K. Goel et al. / Water Research 37 (2003) 891–901
897
Normalized Naphthalene Concentration, C/Co
1.2
1.0
0.8
0.6
0.4 Blank 0.4 l/d (=25 mA equivalent) 0.8 l/d (=50 mA equivalent) 1.6 l/d (=100 mA equivalent) 3.2 l/d (=200 mA equivalent)
0.2
0.0 0
2
4
6
8
10
Time, hours Fig. 3. Naphthalene concentration normalized to the initial concentration during gas stripping with nitrogen.
naphthalene products were detected. They attributed this lack of chlorinated products to the direct oxidation of naphthalene, rather than chlorination of naphthalene. 3.3. Naphthalene degradation by hydrogen peroxide and Fenton’s reagents During the electrolytic aeration of groundwater, hydrogen peroxide may be produced at the cathode, O2 þ 2Hþ þ 2e -H2 O2 :
ð3Þ
Franz et al. [20] reported the production of hydrogen peroxide during the laboratory column experiments with an identical electrode assembly and that hydrogen peroxide may have been produced in their field site. Hydrogen peroxide is a strong oxidant and can oxidize many organic compounds, alone or in combination with other chemicals. Hydrogen peroxide also produces hydroxyl radicals during the oxidation of ferrous iron, Fenton’s reaction d þ Feþ 2 þ H2 O2 ¼ Fe3 þ OH þ HO :
ð4Þ
If ferrous iron is also present in groundwater, Fenton’s reaction can also lead to degradation of contaminants during electrolytic aeration of groundwater. Batch tests were performed to investigate the degradation of naphthalene by hydrogen peroxide alone and by Fenton’s reagent to simulate cases when ferrous iron is present in the groundwater. Franz et al. [20] reported ferrous iron concentrations ranging from 24 to 36 mg/l in their field site. Thus, 30 mg/l of ferrous iron was used to represent the native ferrous iron levels during these experiments. A low (2–3 mg/l) and a high (10–12 mg/l)
hydrogen peroxide concentration range were selected to evaluate the effect of hydrogen peroxide concentration on the rate of naphthalene disappearance with Fenton’s reagents. Fig. 5 shows that naphthalene levels remained constant in the presence of hydrogen peroxide, which is consistent with the study of Tuhkanen and Beltran [38]. Thus, hydrogen peroxide that may have been generated during the electrolytic experiments is not responsible for the decrease in naphthalene levels. Hydrogen peroxide in the presence of ferrous iron resulted in a 99% and 96% decrease of naphthalene concentrations at a pH of 4 and 7, respectively, at a high hydrogen peroxide dose, with a corresponding decrease of 84% and 68% at a pH of 4 and 7, respectively, at a low hydrogen peroxide dose. Fenton’s reaction caused a faster rate of naphthalene decrease at a lower pH, which is consistent with Lin et al. [24] who reported that the degradation rate of phenol during electrochemical oxidation with Fenton’s reaction decreased when pH of the solution increased from 3 to 9. The generation of hydroxyl radicals is known to be pH dependent [30,39] and ferrous iron and hydrogen peroxide is less stable at pH values above 5 as hydrogen peroxide may reduce to water and ferrous iron more readily oxidizes to ferric iron [40]. The TOC values remained fairly constant, indicating minimal mineralization of naphthalene. It was expected that GC/MS analysis on samples at the end of the experiments with Fenton’s reagents would show hydroquinone or naphthoquinone as intermediate compounds due to the intermediately generated hydroxyl radical. However, these products were not detected. Tuhkanen and Beltran [38] reported the formation of
R.K. Goel et al. / Water Research 37 (2003) 891–901
898
Normalized Naphthalene Concentration, C/Co
1.2
1.0
0.8
0.6 Blank, pH=4 35 mg/l Cl2, pH=4
0.4
Blank, pH=7 35 mg/l Cl2, pH=7 0.2
0.0 0
2
4
6
8
10
Time, hours
(a)
Normalized Total Chlorine Concentration, C/Co
1.2
1.0
0.8
0.6
0.4
0.2
pH=4 pH=7
0.0 0
2
4
(b)
6
8
10
Time, hours
Fig. 4. (a) Naphthalene concentration normalized to the initial concentration and (b) total chlorine concentration normalized to the initial concentration during chlorination with an initial chlorine dose of 35 mg/l.
carboxylic acids and aldehydes during the degradation of naphthalene using ultraviolet light and hydrogen peroxide. It is possible that because of the reactivity of Fenton’s reagents, naphthalene oxidation went beyond hydroquinone and naphthoquinone and the products were not detected with the current instrument settings. 3.4. Coagulation with ferric iron Electrolytic aeration of groundwater would oxidize native ferrous iron to ferric iron, a common coagulant.
Since Franz et al. [20] reported ferrous iron levels ranging from 24 to 36 mg/l in the field site, removal of naphthalene through adsorption onto ferric hydroxide flocs in the presence of 0–50 mg/l of ferric iron was evaluated. Considering that the coagulation experiments were about 1 h and 11–31% losses of naphthalene were observed, volatilization can significantly contribute to naphthalene loss in an open system. Since the addition of ferric chloride did not enhance the disappearance of naphthalene relative to the no ferric chloride control, the removal of naphthalene through coagulation is negligible (see Fig. 6).
R.K. Goel et al. / Water Research 37 (2003) 891–901
899
Normalized Naphthalene Concentration, C/Co
1.2
pH=4 1.0
0.8 Blank 1 Blank 2 10-12 mg/l H 2O 2
0.6
10-12 mg/l H 2O 2 + 30 mg/l Fe 0.4
2-3 mg/l H 2 O2 + 30 mg/l Fe
2+
2+
0.2
0.0 0
2
4
(a)
6
8
10
Time, hours
Normalized Naphthalene Concentration, C/Co
1.2
pH=7 1.0
0.8 Blank 1 Blank 2 10-12 mg/l H2O 2
0.6
10-12 mg/l H2O 2 + 30 mg/l Fe 2+ 2-3 mg/l H 2O 2 + 30 mg/l Fe
0.4
2+
0.2
0.0 0
2
4
(b)
6
8
10
Time, hours
Fig. 5. Naphthalene concentration normalized to the initial concentration during hydrogen peroxide oxidation and oxidation with Fenton’s reagents at (a) a pH of 4 and (b) a pH of 7.
4. Conclusions This research demonstrates that the main abiotic mechanisms that could cause naphthalene disappearance during the electrolytic aeration of anoxic groundwater include direct anodic oxidation, oxidation by Fenton’s type reaction when hydrogen peroxide is produced at the cathode and native ferrous iron is present in solution, and oxidation by aqueous chlorine produced due to the oxidation of chloride at the anode. Oxidation using Fenton’s type reaction and aqueous chlorine was faster at a pH level of 4, while the rate of
direct anodic oxidation of naphthalene was the same at a pH level of 4 and 7. The rates of naphthalene disappearance during direct anodic oxidation did not vary with the passed current density. Naphthalene stripping due to gases produced at the electrodes and naphthalene removal via coagulation when Fe2+ is oxidized to Fe3+ were minimal. Although the mechanisms involved in the naphthalene disappearance through oxidation by aqueous chlorine is clear, the products of oxidation using Fenton’s reagents and the mechanisms involved in the electrolytic degradation of naphthalene still require further investigation.
R.K. Goel et al. / Water Research 37 (2003) 891–901
900
Normalized Naphthalene Concentration, C/Co
1.2
1.0
0.8
0.6
0.4 pH = 4 pH = 7
0.2
0.0 0
10
20
30 3+
Fe
40
50
60
Concentration, mg/l
Fig. 6. Naphthalene concentration normalized to the initial concentration before coagulant addition during jar tests with ferric chloride.
In addition, the biodegradability or toxicity of the products must be evaluated to ascertain whether the electrolytic process might be detrimental to the goal of aerating the groundwater and promoting aerobic biodegradation. For example, as shown in the current study, chlorine that may be produced during electrolysis would form chlorinated naphthalenes, which are more difficult to biodegrade and would be worse from an environmental standpoint [41]. Thus, electrolytic in well aeration may not be a desirable technology for high chloride-bearing groundwater. Overall, electrolytic in well aeration to enhance in situ bioremediation at a site is a promising technology. Like any remediation technology, the applicability of electrolytic in well aeration for a specific site must be ascertained first prior to implementation.
Acknowledgements We are thankful to Spencer Walse for help with GC/ MS analysis of samples. This material is based upon work supported in part by the National Science Foundation under Grant No. BES-9733377 and the South Carolina Electric and Gas Company, Inc. Any opinions, findings, and conclusions or recommendations expressed in this manuscript are those of the authors and do not necessarily reflect the views of the funding agencies. Mention of any specific trade name does not constitute endorsement of the product by the authors or the sponsors.
References [1] Freeze AL, Cherry F. Groundwater. Englewood Cliffs, NJ: Prentice-Hall, 1979. [2] Chen CT, Tafuri AN, Rahman M, Foerst MB. Chemical oxidation treatment of petroleum contaminated soil using Fenton’s reagent. J Environ Sci Health 1998;A33(6):987–1008. [3] Warith M, Fernandes L, Gaudet N. Design of in-situ microbial filter for the remediation of naphthalene. Waste Manage 1999;19(1):9–25. [4] Zhu K, Chen H, Li GH, Liu GZ. In situ remediation of petroleum compounds in groundwater aquifer with chlorine dioxide. Water Res 1998;32(5):1471–80. [5] Langwaldt JH, Puhakka JA. On-site biological remediation of contaminated groundwater: a review. Environ Pollut 2000;107(2):187–97. [6] Bass DH, Hastings NA, Brown RA. Performance of air sparging systems: a review of case studies. J Hazard Mater 2000;72(2–3):101–19. [7] Hall BL, Lachmar TE, Dupont RR. Field monitoring and performance evaluation of a field-scale in-well aeration system at a gasoline-contaminated site. J Hazard Mater 2001;82(2):197–212. [8] Ogutveren UB, Toru E, Koparal S. Removal of cyanide by anodic oxidation for wastewater treatment. Water Res 1999;33(8):1851–6. [9] Vlyssides AG, Karlis PK, Zorpas AA. Electrochemical oxidation of noncyanide strippers wastes. Environ Int 1999;25(5):663–70. [10] Vlyssides AG, Loizidou M, Karlis PK, Zorpas AA, Papaioannou D. Electrochemical oxidation of a textile dye wastewater using a Pt/Ti electrode. J Hazard Mater 1999;70(1–2):41–52.
R.K. Goel et al. / Water Research 37 (2003) 891–901 [11] Brillas E, Calpe JC, Casado J. Mineralization of 2,4-D by advanced electrochemical oxidation processes. Water Res 2000;34(8):2253–62. [12] Motheo AJ, Pinhedo L. Electrochemical degradation of humic acid. Sci Total Environ 2000;256(1):67–76. [13] Panizza M, Bocca C, Cerisola G. Electrochemical treatment of wastewater containing polyaromatic organic pollutants. Water Res 2000;34(9):2601–5. [14] Vlyssides AG, Papaioannou D, Loizidoy M, Karlis PK, Zorpas AA. Testing an electrochemical method for treatment of textile dye wastewater. Waste Manage 2000; 20(7):569–74. [15] Okochi M, Lim TK, Nakamura N, Matsunaga T. Electrochemical disinfection of drinking water using an activated-carbon-fiber reactor capable of monitoring its microbial fouling. Appl Microbiol Biotechnol 1997; 47(1):18–22. [16] Hayes AM, Flora JRV, Khan J. Electrolytic stimulation of denitrification in sand columns. Water Res 1998; 32(9):2830–4. [17] Szekeres S, Kiss I, Bejerano TT. Soares MIM Hydrogendependent denitrification in a two-rector bio-electrochemical system. Water Res 2001;35(3):715–9. [18] Skadberg B, Geoly Horn SL, Sangamalli VK, Flora JRV. Influence of pH, current, and copper on the biological dechlorination of 2,6-dichlorophenol in an electrochemical cell. Water Res 1999;33(9):1997–2010. [19] Tanaka T, Kuroda M. Improvement of submerged biofilter process by bioelectrochemical method. ASCE J Environ Eng 2000;126(6):541–8. [20] Franz JA, Williams RA, Flora JRV, Meadows ME, Irwin WE. Electrolytic oxygen generation for subsurface delivery: effects of precipitation at the cathode and an assessment of side reactions. Water Res, 2002;36(9):2243–54. [21] Baker JF, Patrick GC, Major D. Natural attenuation of aromatic hydrocarbons in a shallow sand aquifer. Ground Water Monit Rev 1987;7:64–71. [22] Durant ND, Jonkers CAA, Bouwer EJ. Spatial variability in the naphthalene mineralization response to oxygen, nitrate, and orthophosphate amendments in MGP aquifer sediments. Biodegradation 1997;8:77–86. [23] Hach Company. Hach water analysis handbook. Loveland, CO: Hach Company, 1997. [24] Lin SH, Shyu CT, Sun MC. Saline wastewater treatment by electrochemical method. Water Res 1998;32(4):1059–66. [25] De Sucre VS, Watkinson AP. Anodic oxidation of phenol for waste water treatment. Canad J Chem Eng 1981;59:52–9. [26] Saracco G, Solarino L, Aigotti R, Specchia V, Maja M. Electrochemical oxidation of organic pollutants at low electrolyte concentrations. Electrochim Acta 2000; 46(2–3):373–80.
901
[27] Grimm J, Bessarabov D, Sanderson R. Review of electroassisted methods for water purification. Desalination 1998;115(3):285–94. [28] Amadelli R, De Battisti A, Girenko DV, Kovalyov SV, Velichenko AB. Electrochemical Oxidation of trans-3,4dihydroxycinnamic acid at PbO2 electrodes: direct electrolysis and ozone mediated reactions compared. Electrochim Acta 2000;46(2–3):341–7. [29] Brillas E, Bastida RM, Llosa EL, Casado J. Electrochemical destruction of aniline and 4-chloroaniline for wastewater treatment using a carbon-PTFE O2–fed cathode. J Electrochem Soc 1995;142(6):1733–41. [30] Walling C, Johnson RA. Fenton’s reagents. V. Hydroxylation and side chain cleavage of aromatics. J. Am Chem Soc 1975;97(2):363–7. [31] Mendia L. Electrochemical processes for wastewater treatment. Water Sci Technol 1982;14:331–44. [32] Chiang LC, Chang JE, Wen TC. Indirect oxidation effect in electrochemical oxidation treatment of landfill leachate. Water Res 1995;29(2):671–8. [33] Stumm W, Morgan JJ. Aquatic chemistry, 3rd ed.. New York: Wiley, 1996. [34] Perez G, Caponecchi G, Keheyan Y, Lilla E. Gasphase naphthalene chlorination. Chemosphere 1993; 26(12):2139–46. [35] Singh AP, Kumar SB. Selective monochlorination of naphthalene over zeolite catalysis. Catal Lett 1994;27 (1–2):171–6. [36] De Stefanis A, Perez G, Caponecchi G. Radiation induced competitive chlorination. J Radioanal Nucl Chem Lett 1994;188(2):117–25. [37] Reinhard M, Drevenkar V, Giger W. Effect of aqueous chlorination on the aromatic fraction of diesel fuel: analysis by computer assisted gas chromatography-mass spectrometry. J Chromatogr 1976;116:43–51. [38] Tuhkanen TA, Beltran FJ. Intermediates of the oxidation of naphthalene in water with the combination of hydrogen peroxide and UV radiation. Chemosphere 1995; 30(8):1463–75. [39] Zhu WP, Yang ZH, Wang L. Application of ferroushydrogen peroxide for the treatment of H-acid manufacturing process wastewater. Water Res 1996;30(12): 2949–54. [40] Rivas FJ, Beltran FJ, Frades J, Buxeda P. Oxidation of phydroxybenzoic acid by Fenton’s reagents. Water Res 2001;35(2):387–96. [41] Kannan K, Imawaga T, Blankenship AL, Giesy JP. Isomer-specific analysis and toxic evaluation of polychlorinated naphthalenes in soil, sediment, and biota collected near the site of a former chlor-alkali plant. Environ Sci Technol 1998;32(17):2507–14.
Mechanisms for naphthalene removal during electrolytic aeration Ramesh K. Goela, Joseph R.V. Floraa,*, John Ferryb a
Department of Civil and Environmental Engineering, University of South Carolina, Columbia, SC 29208, USA b Department of Chemistry and Biochemistry, University of South Carolina, Columbia, SC 29208, USA
Abstract Batch tests were performed to investigate chemical and physical processes that may result during electrolytic aeration of a contaminated aquifer using naphthalene as a model contaminant. Naphthalene degradation of 58–66% took place electrolytically and occurred at the same rates at a pH of 4 and 7. 1,4-naphthoquinone was identified as a product of the electrolysis. Stripping due to gases produced at the electrodes did not result in any naphthalene loss. Hydrogen peroxide (which may be produced at the cathode) did not have any effect on naphthalene, but the addition of ferrous iron (which may be present in aquifers) resulted in 67–99% disappearance of naphthalene. Chlorine (which may be produced from the anodic oxidation of chloride) can effectively degrade naphthalene at pH of 4, but not at a pH of 7. Mono-, di- and poly chloronaphthalenes were identified as oxidation products. Ferric iron coagulation (due to the oxidation of ferrous iron) did not significantly contribute to naphthalene loss. Overall, electrolytic oxidation and chemical oxidation due to the electrolytic by-products formed are significant abiotic processes that could occur and should be accounted for if bioremediation of PAH-contaminated sites via electrolytic aeration is considered. Possible undesirable products such as chlorinated compounds may be formed when significant amounts of chlorides are present. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Naphthalene; Electrolysis; Aeration; Gas stripping; Hydrogen peroxide; Fenton’s reagents; Chlorine; Oxidation; Coagulation
1. Introduction Remediation of contaminated soil and groundwater is an important issue that is of concern among environmentalists and hydrogeologists. In 1977, the United States Environmental Protection Agency reported that a minimum of 17 million waste disposal facilities were discharging 6.5 billion m3 of liquid waste into the ground each year [1]. Gasoline and petroleum derived hydrocarbons from leaking underground storage tanks (UST) are another important contributor to groundwater contamination. There are more than 2 million UST systems located at about 700 thousand facilities existing in the United States [2]. Approximately 85% of all underground storage tanks are made of steel and about *Corresponding author. Tel.: +1-803-777-3614; fax: +1803-777-0670. E-mail address: fl[email protected] (J.R.V. Flora).
25% of the existing USTs fail tightness testing and may be leaking [3]. Remediation methods to address soil and groundwater contamination can be divided into two categories: in situ and ex situ methods. In situ remediation approaches include aeration to induce biodegradation, the use of capture zones, and oxidation through chemical injection [4]. Pump and treat and soil vapor extraction are two widely used ex situ remediation approaches. In situ remediation approaches are excellent alternatives for treating contaminated aquifers because they are more cost effective than ex situ remediation processes. The pump and treat approach has frequently been proven inefficient due to slow desorption, diffusion, and dissolution processes of contaminants [5]. Aeration is an in situ remediation approach that was developed in the late 1980s for treating aquifers contaminated with volatile organic compounds (VOCs) [6]. During aeration, the injection of air in the saturated
0043-1354/03/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 3 7 6 - 7
892
R.K. Goel et al. / Water Research 37 (2003) 891–901
zone of the aquifer physically removes dissolved and adsorbed VOCs and transfers oxygen into the groundwater. As a result, both physical removal and aerobic biodegradation of contaminants by indigenous organisms are enhanced [6,7]. Aeration can further be subdivided into two categories: in situ air sparging and in-well aeration (IWA) [7]. In IWA, air is injected into the well below the groundwater table. This aerated water flows to the rest of aquifer, either by the natural flow of groundwater or by pumping. Electrolysis of groundwater is one method for IWA. In electrolytic IWA, a small potential difference is applied across the cathode and anode of an electrolytic cell inserted in a well. The anodic oxidation of H2O then produces O2 in situ. Some of the applications of the electrolysis of water in environmental engineering include direct abiotic oxidation of contaminants [8–14] or disinfection of drinking water [15]. Furthermore, electrolytic reactions have been used to simulate biological activity such as denitrification [16–17], dechlorination [18], or to increase the nitrogen removal efficiency of biofilter processes [19]. Electrolytic oxygen generation for groundwater aeration and some of the side reactions that result from the process have been investigated and discussed by Franz et al. [20]. The side reactions included chlorine generation, hydrogen peroxide generation, and a drop in pH during laboratory and field-scale investigations. During a 77-day in-well aeration test using a similar electrode assembly at their field site, naphthalene levels decreased from a background concentration of 1.1–0.2 mg/l using a current density of 2.5 mA/cm2, chlorine was not detected, the pH dropped 2.4 units to a value of 3.8, ferrous iron ranged from 24 to 36 mg/l, and a possible upper limit of 1.5 mg/l hydrogen peroxide was reported in the injection well. Their results were used to determine the mechanisms that needed to be studied for contaminant removal during electrolytic groundwater remediation. The specific objectives of this research were to evaluate the mechanisms for contaminant removal that can occur during electrolytic aeration of contaminated groundwater. Of primary concern was the possibility that the anodic oxidation of chloride containing groundwater might lead to the formation of HOCl and cathodic reduction of oxygen might produce hydrogen peroxide, both of which could engage in the direct oxidation of contaminants. Contaminant loss can also occur due to direct electrolytic transformation, stripping due to gases formed at the electrodes, indirect reactions due to Fenton’s type chemistry between native ferrous iron and hydrogen peroxide, and coagulation from the ferric iron produced due to oxidation of native ferrous iron. This paper presents an evaluation of these side reactions using naphthalene as a model contaminant. Naphthalene is a polynuclear aromatic hydrocarbon that has
been found in numerous contaminated sites [21,22] and is of particular concern in the site reported in Franz et al. [20]. The experiments performed in this research were designed to complement the study of Franz et al. [20].
2. Materials and methods 2.1. Naphthalene test solution Naphthalene (Fisher Scientific, Fairlawn, NJ) was added to deionized water and stirred for 12–14 h to prepare a saturated solution. This stock solution was filtered and diluted to 9.871.5 mg/l with deionized H2O. 1.42 g/l of sodium sulfate (Fisher Scientific, Fair Lawn, NJ) was added to the diluted naphthalene solution to solution, simulating sulfate make a 0.01 M SO2 4 concentrations measured in the field site reported by Franz et al. [20]. Since it was expected that ionic strength (which was dominated by sulfate ions) would be critical during electrolysis, it was decided to consistently add sulfate in all the experiments. With the exception of gas stripping experiments, all the tests were performed at a pH of 470.2 and 770.2. The pH was adjusted using either concentrated sulfuric acid or 0.25 N NaOH. Neutral pH was used for the gas stripping experiments. 2.2. Electrolytic degradation experiments The electrode assembly in the present study was identical to the assembly used in Franz et al. [20]. The electrode assembly consisted of a stainless steel plate cathode (Type 304, #3 polish, 0.07 cm thickness (Metal Supermarket, Columbia, SC)) and a titanium anode with a mixed metal oxide coating (Type EC-600, 0.15 cm. thick, mesh anode (Eltec Systems Corp., Chardon, OH)) held together by nylon screws. The electrodes were 3.18 cm 6.35 cm separated by a distance of 1.4 cm. Titanium screws and nuts were used as the current collector on both the anode and the cathode to minimize the reactivity of the current collectors and to reduce the possibility of corrosion. The current collectors on the cathode and anode were connected to copper wires, which were connected directly to a power supply unit (HP Model E3612A, Agilent Technologies, Inc., Englewood, CO). The connection between the copper wires and titanium screws at the cathode and anode were sealed using a fast drying epoxy and coated with a silicon sealant. The reactor vessels used for the electrolytic experiments were 500 ml amber-colored bottles. Holes were drilled through the caps to run the wires connected to the electrode assembly. The gap between the copper wires running through the cap was sealed with an epoxy sealant. One 12 mm hole was drilled on the cap of all the bottles, including the experimental control, to keep the
R.K. Goel et al. / Water Research 37 (2003) 891–901
system under atmospheric pressure. Three liters of the naphthalene test solution was prepared and divided into two batches, with the pH of a batch adjusted to either 4 or 7. Five hundred ml of each solution was transferred to separate 500 ml amber-colored bottles. Duplicate tests were performed with passed currents of 25, 50, 100, and 200 mA, corresponding to current densities of 1.2, 2.5, 5.0, and 9.9 mA/cm2. An experimental control (i.e., no current) was performed using a reactor vessel containing an electrode assembly. The contents of the bottles were stirred continuously using a magnetic stirrer bar and plate. Samples were withdrawn directly from the bottles by opening the cap. The total organic carbon (TOC) and naphthalene concentrations of the samples were quantified. Samples for GC/MS analysis were withdrawn at the end of each experiment. 2.3. Gas stripping experiments Experiments were performed to investigate the contribution of gas stripping to the disappearance of naphthalene observed during the electrolytic experiments. Using Faraday’s law and ideal gas law, the volume of gas produced from oxygen generation at the anode and hydrogen generation at the cathode would be 1.6 l/day for a passed current of 100 mA. Stripping experiments corresponding to the gas flowrates for the four current intensities used in electrolytic experiments were performed. A 12.7 cm long stainless steel needle for gas delivery was run through a hole drilled through the cap of a 500 ml amber-colored bottle, with 6.4 cm of the needle penetrating through the cap of the bottle into the naphthalene solution. The gap between the needle and the cap was sealed with an epoxy sealant. One 12 mm hole was drilled on the cap of all the bottles, including the experimental control, to keep the system under atmospheric pressure. A 1-l naphthalene test solution was prepared without pH adjustment. 500 ml were transferred to two 500 ml amber-colored bottles and were stirred using a magnetic stirrer bar and plate. Scientific grade nitrogen gas was bubbled in the naphthalene solution in one bottle with the gas flow controlled using a regulator attached to a gas cylinder and an intermediate gas valve. The gas flow was quantified using a Riteflow flow meter with plain ends (Model 150 mmSZ1, Bel-art Products, Pequannock, NJ) calibrated using a wet-tip gas flow meter. Samples for naphthalene and TOC analysis were withdrawn in a similar manner as the electrolytic experiments. 2.4. Experiments with chlorine The chlorination experiments were performed under identical conditions to the electrolysis experiments. 1.2 ml of 5.25% by weight sodium hypochlorite solution
893
was added to each liter of the two test solutions (at pH 4 and 7) to give measured total chlorine concentrations of 35 mg/l in each solution. The pH of the solution was checked after chlorine addition and adjusted immediately to 4 or 7. Periodic samples were withdrawn from the bottles for naphthalene, TOC, and chlorine analysis. Samples for GC/MS analysis were withdrawn at the end of each experiment. 2.5. Experiments with fenton’s reagents and hydrogen peroxide Fenton’s experiments were carried with hydrogen peroxide concentrations of 2–3 and 10–12 mg/l and a ferrous iron concentration of 30 mg/l. Experiments with only hydrogen peroxide were also performed at a concentration of 10–12 mg/l. The experimental conditions were the same as in the electrolysis experiments. To get the desired hydrogen peroxide concentrations, 30% by weight hydrogen peroxide solution (Aldrich Chemical Company Inc, Milwaukee, WI) was used. Ferrous iron was added using a stock solution prepared by adding 1.49 g FeSO4 7H2O in 100 ml DI water at a pH of 3. Periodic samples were withdrawn from the bottles for naphthalene and TOC analysis. Samples for GC/MS analysis were withdrawn at the end of each experiment. 2.6. Coagulation with ferric iron Standard jar tests were performed using a Phipps and Bird jar test apparatus. Six liters of a naphthalene test solution was prepared and adjusted to either a pH of 4 or 7. Nine hundred ml of the solution were transferred to six 1-l glass beakers, which were open to the atmosphere. Varying amounts (0–25 ml) of a FeCl3 stock solution were added to produce Fe3+ doses ranging from 0–50 mg/l. FeCl3 stock solution was prepared by adding 1.74 g FeCl3 6H2O in 200 ml deionized water. Pre-determined amounts of NaOH were added simultaneously with the coagulant to maintain the pH at appropriate levels. After chemical addition, the solutions in the beakers were mixed at 150 RPM for 1 min, after which the mixing speed was reduced to 25 RPM for 30 min. Mixing was then stopped and quiescent settling was allowed to occur for 30 min prior to sample withdrawal for naphthalene and TOC analysis. 2.7. Analytical methods Naphthalene was quantified using a Varian Model 3380 gas chromatograph equipped with a flame ionization detector, Chrompack capillary column (Select 624 CB Df 1.8 mm, FS 30 0.32 mm ID), and an auto-sampler with a 100 mm PDMS coated SPME fiber assembly (Supelco, Bellefonte, PA). The sample
894
R.K. Goel et al. / Water Research 37 (2003) 891–901
adsorption time with the SPME fiber was 10 min in agitate mode and the desorption time was 2 min followed by a 1 min waiting period. The analysis was performed in splitless mode with an injection temperature of 2501C, isothermal oven temperature of 1801C, and detector temperature of 2751C. TOC was quantified using a Shimadzu TOC 5000A analyzer equipped with an autosampler. The total chlorine and ferrous iron concentrations were measured using Hach Methods 8167 and 8146, respectively, with a Hach DR2000 spectrophotometer [23]. Hydrogen peroxide was measured using CHEMets Kit K-5510 (Chemetrics, Inc., Calverton, VA). The pH was measured using an Ag/AgCl combination pH electrode attached to a pH meter (Ion Analyzer 220, Corning, Inc., Corning, NY). GC/MS analysis was performed on the samples at the end of experiments where naphthalene degradation was noted. The samples were extracted with hexane and approximately 25 ml of the extracted organic solution was concentrated in Rotavapor R 110 (BUCHI, Slawil, Switzerland) up to a capacity of 1–2 ml. This concentrated solution was analyzed using GC-IT (ion-trap) MS (Varian 3800 model GC equipped with a Varian 8200 autoinjector and a Varian Saturn 2000 ITMS). The analytical column was a 30 m DB-5 with a 25 mm-film thickness. The injector port was set for splitless operation at 2201C and was ramped to 2601C for 1 min after sample injection (autoinjector volume=1 ml).
3. Results and discussion 3.1. Electrolytic naphthalene degradation and gas stripping Naphthalene solutions were electrolyzed over a range of current densities. Under our conditions, naphthalene degradation rates were insensitive to current density and independent of system pH (see Figs. 1a and b). The zerocurrent control showed some naphthalene loss, probably due to volatilization or adsorption on the inactive electrode surface. Corresponding TOC values for all current intensities, including the control, are plotted in Figs. 2a and b, which shows that TOC values are constant, except for the decrease with 200 mA current after 2 h. However, these differences appear to be within the operational sensitivity of the instrument. Constant TOC values indicate little or no naphthalene mineralization during electrolysis. The gas stripping experiments showed some losses of naphthalene in the control experiment over a 9 h period (see Fig. 3). The addition of nitrogen gas supports the conclusion that the formation of gases on the electrodes does not contribute to significant losses of naphthalene due to volatilization under the conditions of the study.
Gas stripping may be a more significant mechanism for more volatile compounds. Electro-coagulation, electro-flotation and electro-oxidation are three processes that can contribute to contaminant loss in electrochemical systems [24]. In the present experiments, electro-coagulation and electrofloatation did not occur as turbidity or settled solids were not present. Anodic oxidation of aromatics have been reported by various researchers [25,26] where contaminants are either directly transformed on the anode into CO2 and water or into some other intermediate product, or contaminants are indirectly oxidized through electrocatalytic process [27]. Although electrocatalysts were not added in this study, water can undergo direct anodic oxidation to yield hydroxyl radicals [28], the presence of which has been reported by various authors during the electrolytic degradation of contaminants (e.g., [11,13,14,29]). For example, Panizza et al. [13] performed experiments on the anodic oxidation of 2-naphthol where they attributed the 2naphthol degradation to hydroxyl radicals formed in their system. Furthermore, hydroxyl radicals may be involved in the formation of ozone, another strong oxidant [28]. Under our experimental conditions, constant TOC values rule out the possibility of naphthalene transformation into CO2 and water due to direct anodic oxidation. Hence, naphthalene disappearance could be due to its chemical conversion into other intermediate products because of hydroxyl radicals formed during electrolysis. GC/MS analysis on samples withdrawn at the end of the electrolytic experiments confirmed the presence of at least one reaction product, 1,4-naphthoquinone. The presence of this compound is consistent with those found by other researchers who also observed aromatic quinones during the electrolytic oxidation of aromatic hydrocarbons (e.g., [11,13,26]). Hydroxyl radicals can directly attack aromatic rings by hydroxylation followed by oxidation [30]. Although hydroxyl radicals are known to produce quinones during the oxidation of aromatics, other oxidants may also mediate this transformation. Hence, product analysis cannot rule out other compounding pathways, including the direct anodic oxidation of naphthalene. Naphthalene degradation in the present electrolytic experiments could be due to direct anodic oxidation, oxidation by an intermediate (e.g., hydroxyl radicals), or a combination of both. 3.2. Naphthalene degradation by aqueous chlorine Batch experiments to evaluate the reactivity of chlorine with naphthalene were conducted with 35 mg/l total chlorine at a pH of 4 and 7. These experiments were performed because electrolysis of chloride–containing water produces chlorine at the anode, which subsequently hydrolyzes to hypochlorous acid
R.K. Goel et al. / Water Research 37 (2003) 891–901
895
Normalized Naphthalene Concentration, C/Co
1.2
pH=4 1.0
0.8
0.6
0.4 Blank 25 mA 50 mA 100 mA 200 mA
0.2
0.0 0
2
4
6
8
10
Time, hours
(a)
Normalized Naphthalene Concentration, C/Co
1.2
pH=7 1.0
0.8
0.6
0.4 Blank 25 mA 50 mA 100 mA 200 mA
0.2
0.0 0
2
4
(b)
6
8
10
Time, hours
Fig. 1. Naphthalene concentration normalized to the initial concentration during electrolysis at (a) a pH of 4 and (b) a pH of 7.
[8,13,24,31,32]. Franz et al. [20] found total chlorine to be as high as 43 mg/l in laboratory column studies evaluating potential side reactions during electrolytic aeration of groundwater. Furthermore, chloride concentrations as high as 110 mg/l were also observed in other wells at the field site where electrolytic aeration of contaminated groundwater is being tested. Fig. 4a shows a 98% decrease in naphthalene concentration in the presence of chlorine at a pH of 4 and a negligible change in naphthalene at a pH of 7. Corresponding measured total chlorine concentrations fell by 33% at a pH of 4, but the chlorine concentration
remained almost constant at a pH of 7 (see Fig. 4b). Based on the following equilibria [33]: Cl2ðaqÞ þ H2 O"HOCl þ Hþ þ Cl HOCl"Hþ þ OCl
ðlog K ¼ 3:3Þ;ð1Þ
ðlog K ¼ 7:3Þ:
ð2Þ
The primary chlorine species at a pH of 7 are HOCl and OCl–, and Fig. 4a indicates that these species do not react with naphthalene. Although the HOCl and OCl– species would still predominate at a pH of 4.0 at the low levels of chloride in the experiments, the Cl2(aq) concentrations at a pH of 4 is 1500 times greater than
R.K. Goel et al. / Water Research 37 (2003) 891–901
896
Normalized TOC Concentration, C/Co
1.2
1.0
0.8
0.6
0.4 Blank 25 mA 50 mA 100 mA 200 mA
0.2
pH=4
0.0 0
2
4
(a)
6
8
10
Time, hours
Normalized TOC Concentration, C/Co
1.2
1.0
0.8
0.6
0.4 Blank 25 mA 50 mA 100 mA 200 mA
0.2
pH=7
0.0 0
2
4
(b)
6
8
10
Time, hours
Fig. 2. Total organic concentration normalized to the initial concentration during electrolysis at (a) a pH of 4 and (b) a pH of 7.
the Cl2(aq) levels at a pH of 7 (with the actual concentration dependent on the chloride levels). Thus, it is possible that Cl2(aq) is the species that reacts with naphthalene, which is consistent with the observed faster kinetics at low pH values. The TOC was essentially constant throughout the experiments in both cases, indicating no mineralization of naphthalene. GC/MS analysis of samples at the end of the chlorination experiments at a pH of 4 showed monochlorinated (1-chloro and 2-chloronaphthalene) and various congeners of di- and tri-chloronaphthalenes. Chlorinated naphthoquinones were also identified.
These results are consistent with studies reported in the literature [34–36]. However, in the above-mentioned literature, addition of chlorine atoms at one or more positions on the naphthalene molecule took place either in the presence of catalysts like FeCl3 [35], in the gas phase [34], or by radiolysis in a chlorinated solvent [36]. Reports on the direct chlorination of naphthalene by aqueous chlorine alone are limited. Reinhard et al. [37] added a total free chlorine concentration of 100 mg/l to the aromatic fraction of diesel fuel (with naphthalene as one of the constituents). After 30 h of chlorination, all naphthalene had disappeared, but no chlorinated
R.K. Goel et al. / Water Research 37 (2003) 891–901
897
Normalized Naphthalene Concentration, C/Co
1.2
1.0
0.8
0.6
0.4 Blank 0.4 l/d (=25 mA equivalent) 0.8 l/d (=50 mA equivalent) 1.6 l/d (=100 mA equivalent) 3.2 l/d (=200 mA equivalent)
0.2
0.0 0
2
4
6
8
10
Time, hours Fig. 3. Naphthalene concentration normalized to the initial concentration during gas stripping with nitrogen.
naphthalene products were detected. They attributed this lack of chlorinated products to the direct oxidation of naphthalene, rather than chlorination of naphthalene. 3.3. Naphthalene degradation by hydrogen peroxide and Fenton’s reagents During the electrolytic aeration of groundwater, hydrogen peroxide may be produced at the cathode, O2 þ 2Hþ þ 2e -H2 O2 :
ð3Þ
Franz et al. [20] reported the production of hydrogen peroxide during the laboratory column experiments with an identical electrode assembly and that hydrogen peroxide may have been produced in their field site. Hydrogen peroxide is a strong oxidant and can oxidize many organic compounds, alone or in combination with other chemicals. Hydrogen peroxide also produces hydroxyl radicals during the oxidation of ferrous iron, Fenton’s reaction d þ Feþ 2 þ H2 O2 ¼ Fe3 þ OH þ HO :
ð4Þ
If ferrous iron is also present in groundwater, Fenton’s reaction can also lead to degradation of contaminants during electrolytic aeration of groundwater. Batch tests were performed to investigate the degradation of naphthalene by hydrogen peroxide alone and by Fenton’s reagent to simulate cases when ferrous iron is present in the groundwater. Franz et al. [20] reported ferrous iron concentrations ranging from 24 to 36 mg/l in their field site. Thus, 30 mg/l of ferrous iron was used to represent the native ferrous iron levels during these experiments. A low (2–3 mg/l) and a high (10–12 mg/l)
hydrogen peroxide concentration range were selected to evaluate the effect of hydrogen peroxide concentration on the rate of naphthalene disappearance with Fenton’s reagents. Fig. 5 shows that naphthalene levels remained constant in the presence of hydrogen peroxide, which is consistent with the study of Tuhkanen and Beltran [38]. Thus, hydrogen peroxide that may have been generated during the electrolytic experiments is not responsible for the decrease in naphthalene levels. Hydrogen peroxide in the presence of ferrous iron resulted in a 99% and 96% decrease of naphthalene concentrations at a pH of 4 and 7, respectively, at a high hydrogen peroxide dose, with a corresponding decrease of 84% and 68% at a pH of 4 and 7, respectively, at a low hydrogen peroxide dose. Fenton’s reaction caused a faster rate of naphthalene decrease at a lower pH, which is consistent with Lin et al. [24] who reported that the degradation rate of phenol during electrochemical oxidation with Fenton’s reaction decreased when pH of the solution increased from 3 to 9. The generation of hydroxyl radicals is known to be pH dependent [30,39] and ferrous iron and hydrogen peroxide is less stable at pH values above 5 as hydrogen peroxide may reduce to water and ferrous iron more readily oxidizes to ferric iron [40]. The TOC values remained fairly constant, indicating minimal mineralization of naphthalene. It was expected that GC/MS analysis on samples at the end of the experiments with Fenton’s reagents would show hydroquinone or naphthoquinone as intermediate compounds due to the intermediately generated hydroxyl radical. However, these products were not detected. Tuhkanen and Beltran [38] reported the formation of
R.K. Goel et al. / Water Research 37 (2003) 891–901
898
Normalized Naphthalene Concentration, C/Co
1.2
1.0
0.8
0.6 Blank, pH=4 35 mg/l Cl2, pH=4
0.4
Blank, pH=7 35 mg/l Cl2, pH=7 0.2
0.0 0
2
4
6
8
10
Time, hours
(a)
Normalized Total Chlorine Concentration, C/Co
1.2
1.0
0.8
0.6
0.4
0.2
pH=4 pH=7
0.0 0
2
4
(b)
6
8
10
Time, hours
Fig. 4. (a) Naphthalene concentration normalized to the initial concentration and (b) total chlorine concentration normalized to the initial concentration during chlorination with an initial chlorine dose of 35 mg/l.
carboxylic acids and aldehydes during the degradation of naphthalene using ultraviolet light and hydrogen peroxide. It is possible that because of the reactivity of Fenton’s reagents, naphthalene oxidation went beyond hydroquinone and naphthoquinone and the products were not detected with the current instrument settings. 3.4. Coagulation with ferric iron Electrolytic aeration of groundwater would oxidize native ferrous iron to ferric iron, a common coagulant.
Since Franz et al. [20] reported ferrous iron levels ranging from 24 to 36 mg/l in the field site, removal of naphthalene through adsorption onto ferric hydroxide flocs in the presence of 0–50 mg/l of ferric iron was evaluated. Considering that the coagulation experiments were about 1 h and 11–31% losses of naphthalene were observed, volatilization can significantly contribute to naphthalene loss in an open system. Since the addition of ferric chloride did not enhance the disappearance of naphthalene relative to the no ferric chloride control, the removal of naphthalene through coagulation is negligible (see Fig. 6).
R.K. Goel et al. / Water Research 37 (2003) 891–901
899
Normalized Naphthalene Concentration, C/Co
1.2
pH=4 1.0
0.8 Blank 1 Blank 2 10-12 mg/l H 2O 2
0.6
10-12 mg/l H 2O 2 + 30 mg/l Fe 0.4
2-3 mg/l H 2 O2 + 30 mg/l Fe
2+
2+
0.2
0.0 0
2
4
(a)
6
8
10
Time, hours
Normalized Naphthalene Concentration, C/Co
1.2
pH=7 1.0
0.8 Blank 1 Blank 2 10-12 mg/l H2O 2
0.6
10-12 mg/l H2O 2 + 30 mg/l Fe 2+ 2-3 mg/l H 2O 2 + 30 mg/l Fe
0.4
2+
0.2
0.0 0
2
4
(b)
6
8
10
Time, hours
Fig. 5. Naphthalene concentration normalized to the initial concentration during hydrogen peroxide oxidation and oxidation with Fenton’s reagents at (a) a pH of 4 and (b) a pH of 7.
4. Conclusions This research demonstrates that the main abiotic mechanisms that could cause naphthalene disappearance during the electrolytic aeration of anoxic groundwater include direct anodic oxidation, oxidation by Fenton’s type reaction when hydrogen peroxide is produced at the cathode and native ferrous iron is present in solution, and oxidation by aqueous chlorine produced due to the oxidation of chloride at the anode. Oxidation using Fenton’s type reaction and aqueous chlorine was faster at a pH level of 4, while the rate of
direct anodic oxidation of naphthalene was the same at a pH level of 4 and 7. The rates of naphthalene disappearance during direct anodic oxidation did not vary with the passed current density. Naphthalene stripping due to gases produced at the electrodes and naphthalene removal via coagulation when Fe2+ is oxidized to Fe3+ were minimal. Although the mechanisms involved in the naphthalene disappearance through oxidation by aqueous chlorine is clear, the products of oxidation using Fenton’s reagents and the mechanisms involved in the electrolytic degradation of naphthalene still require further investigation.
R.K. Goel et al. / Water Research 37 (2003) 891–901
900
Normalized Naphthalene Concentration, C/Co
1.2
1.0
0.8
0.6
0.4 pH = 4 pH = 7
0.2
0.0 0
10
20
30 3+
Fe
40
50
60
Concentration, mg/l
Fig. 6. Naphthalene concentration normalized to the initial concentration before coagulant addition during jar tests with ferric chloride.
In addition, the biodegradability or toxicity of the products must be evaluated to ascertain whether the electrolytic process might be detrimental to the goal of aerating the groundwater and promoting aerobic biodegradation. For example, as shown in the current study, chlorine that may be produced during electrolysis would form chlorinated naphthalenes, which are more difficult to biodegrade and would be worse from an environmental standpoint [41]. Thus, electrolytic in well aeration may not be a desirable technology for high chloride-bearing groundwater. Overall, electrolytic in well aeration to enhance in situ bioremediation at a site is a promising technology. Like any remediation technology, the applicability of electrolytic in well aeration for a specific site must be ascertained first prior to implementation.
Acknowledgements We are thankful to Spencer Walse for help with GC/ MS analysis of samples. This material is based upon work supported in part by the National Science Foundation under Grant No. BES-9733377 and the South Carolina Electric and Gas Company, Inc. Any opinions, findings, and conclusions or recommendations expressed in this manuscript are those of the authors and do not necessarily reflect the views of the funding agencies. Mention of any specific trade name does not constitute endorsement of the product by the authors or the sponsors.
References [1] Freeze AL, Cherry F. Groundwater. Englewood Cliffs, NJ: Prentice-Hall, 1979. [2] Chen CT, Tafuri AN, Rahman M, Foerst MB. Chemical oxidation treatment of petroleum contaminated soil using Fenton’s reagent. J Environ Sci Health 1998;A33(6):987–1008. [3] Warith M, Fernandes L, Gaudet N. Design of in-situ microbial filter for the remediation of naphthalene. Waste Manage 1999;19(1):9–25. [4] Zhu K, Chen H, Li GH, Liu GZ. In situ remediation of petroleum compounds in groundwater aquifer with chlorine dioxide. Water Res 1998;32(5):1471–80. [5] Langwaldt JH, Puhakka JA. On-site biological remediation of contaminated groundwater: a review. Environ Pollut 2000;107(2):187–97. [6] Bass DH, Hastings NA, Brown RA. Performance of air sparging systems: a review of case studies. J Hazard Mater 2000;72(2–3):101–19. [7] Hall BL, Lachmar TE, Dupont RR. Field monitoring and performance evaluation of a field-scale in-well aeration system at a gasoline-contaminated site. J Hazard Mater 2001;82(2):197–212. [8] Ogutveren UB, Toru E, Koparal S. Removal of cyanide by anodic oxidation for wastewater treatment. Water Res 1999;33(8):1851–6. [9] Vlyssides AG, Karlis PK, Zorpas AA. Electrochemical oxidation of noncyanide strippers wastes. Environ Int 1999;25(5):663–70. [10] Vlyssides AG, Loizidou M, Karlis PK, Zorpas AA, Papaioannou D. Electrochemical oxidation of a textile dye wastewater using a Pt/Ti electrode. J Hazard Mater 1999;70(1–2):41–52.
R.K. Goel et al. / Water Research 37 (2003) 891–901 [11] Brillas E, Calpe JC, Casado J. Mineralization of 2,4-D by advanced electrochemical oxidation processes. Water Res 2000;34(8):2253–62. [12] Motheo AJ, Pinhedo L. Electrochemical degradation of humic acid. Sci Total Environ 2000;256(1):67–76. [13] Panizza M, Bocca C, Cerisola G. Electrochemical treatment of wastewater containing polyaromatic organic pollutants. Water Res 2000;34(9):2601–5. [14] Vlyssides AG, Papaioannou D, Loizidoy M, Karlis PK, Zorpas AA. Testing an electrochemical method for treatment of textile dye wastewater. Waste Manage 2000; 20(7):569–74. [15] Okochi M, Lim TK, Nakamura N, Matsunaga T. Electrochemical disinfection of drinking water using an activated-carbon-fiber reactor capable of monitoring its microbial fouling. Appl Microbiol Biotechnol 1997; 47(1):18–22. [16] Hayes AM, Flora JRV, Khan J. Electrolytic stimulation of denitrification in sand columns. Water Res 1998; 32(9):2830–4. [17] Szekeres S, Kiss I, Bejerano TT. Soares MIM Hydrogendependent denitrification in a two-rector bio-electrochemical system. Water Res 2001;35(3):715–9. [18] Skadberg B, Geoly Horn SL, Sangamalli VK, Flora JRV. Influence of pH, current, and copper on the biological dechlorination of 2,6-dichlorophenol in an electrochemical cell. Water Res 1999;33(9):1997–2010. [19] Tanaka T, Kuroda M. Improvement of submerged biofilter process by bioelectrochemical method. ASCE J Environ Eng 2000;126(6):541–8. [20] Franz JA, Williams RA, Flora JRV, Meadows ME, Irwin WE. Electrolytic oxygen generation for subsurface delivery: effects of precipitation at the cathode and an assessment of side reactions. Water Res, 2002;36(9):2243–54. [21] Baker JF, Patrick GC, Major D. Natural attenuation of aromatic hydrocarbons in a shallow sand aquifer. Ground Water Monit Rev 1987;7:64–71. [22] Durant ND, Jonkers CAA, Bouwer EJ. Spatial variability in the naphthalene mineralization response to oxygen, nitrate, and orthophosphate amendments in MGP aquifer sediments. Biodegradation 1997;8:77–86. [23] Hach Company. Hach water analysis handbook. Loveland, CO: Hach Company, 1997. [24] Lin SH, Shyu CT, Sun MC. Saline wastewater treatment by electrochemical method. Water Res 1998;32(4):1059–66. [25] De Sucre VS, Watkinson AP. Anodic oxidation of phenol for waste water treatment. Canad J Chem Eng 1981;59:52–9. [26] Saracco G, Solarino L, Aigotti R, Specchia V, Maja M. Electrochemical oxidation of organic pollutants at low electrolyte concentrations. Electrochim Acta 2000; 46(2–3):373–80.
901
[27] Grimm J, Bessarabov D, Sanderson R. Review of electroassisted methods for water purification. Desalination 1998;115(3):285–94. [28] Amadelli R, De Battisti A, Girenko DV, Kovalyov SV, Velichenko AB. Electrochemical Oxidation of trans-3,4dihydroxycinnamic acid at PbO2 electrodes: direct electrolysis and ozone mediated reactions compared. Electrochim Acta 2000;46(2–3):341–7. [29] Brillas E, Bastida RM, Llosa EL, Casado J. Electrochemical destruction of aniline and 4-chloroaniline for wastewater treatment using a carbon-PTFE O2–fed cathode. J Electrochem Soc 1995;142(6):1733–41. [30] Walling C, Johnson RA. Fenton’s reagents. V. Hydroxylation and side chain cleavage of aromatics. J. Am Chem Soc 1975;97(2):363–7. [31] Mendia L. Electrochemical processes for wastewater treatment. Water Sci Technol 1982;14:331–44. [32] Chiang LC, Chang JE, Wen TC. Indirect oxidation effect in electrochemical oxidation treatment of landfill leachate. Water Res 1995;29(2):671–8. [33] Stumm W, Morgan JJ. Aquatic chemistry, 3rd ed.. New York: Wiley, 1996. [34] Perez G, Caponecchi G, Keheyan Y, Lilla E. Gasphase naphthalene chlorination. Chemosphere 1993; 26(12):2139–46. [35] Singh AP, Kumar SB. Selective monochlorination of naphthalene over zeolite catalysis. Catal Lett 1994;27 (1–2):171–6. [36] De Stefanis A, Perez G, Caponecchi G. Radiation induced competitive chlorination. J Radioanal Nucl Chem Lett 1994;188(2):117–25. [37] Reinhard M, Drevenkar V, Giger W. Effect of aqueous chlorination on the aromatic fraction of diesel fuel: analysis by computer assisted gas chromatography-mass spectrometry. J Chromatogr 1976;116:43–51. [38] Tuhkanen TA, Beltran FJ. Intermediates of the oxidation of naphthalene in water with the combination of hydrogen peroxide and UV radiation. Chemosphere 1995; 30(8):1463–75. [39] Zhu WP, Yang ZH, Wang L. Application of ferroushydrogen peroxide for the treatment of H-acid manufacturing process wastewater. Water Res 1996;30(12): 2949–54. [40] Rivas FJ, Beltran FJ, Frades J, Buxeda P. Oxidation of phydroxybenzoic acid by Fenton’s reagents. Water Res 2001;35(2):387–96. [41] Kannan K, Imawaga T, Blankenship AL, Giesy JP. Isomer-specific analysis and toxic evaluation of polychlorinated naphthalenes in soil, sediment, and biota collected near the site of a former chlor-alkali plant. Environ Sci Technol 1998;32(17):2507–14.