- Email: [email protected]
Formation of chloroacetic acids from soil, humic acid and phenolic moieties
Chemosphere 52 (2003) 513–520 www.elsevier.com/locate/chemosphere
Formation of chloroacetic acids from soil, humic acid and phenolic moieties I.J. Fahimi *, F. Keppler, H.F. Sch€ oler Institute of Environmental Geochemistry, Heidelberg University, Im Neuenheimer Feld 236, D-69120 Heidelberg, Germany Received 15 April 2002; received in revised form 19 July 2002; accepted 23 July 2002
Abstract The mechanism of formation of chloroacetates, which are important toxic environmental substances, has been controversial. Whereas the anthropogenic production has been well established, a natural formation has also been suggested. In this study the natural formation of chloroacetic acids from soil, as well as from humic material which is present in soil and from phenolic model substances has been investigated. It is shown that chloroacetates are formed from humic material with a linear relationship between the amount of humic acid used and chloroacetates found. More dichloroacetate (DCA) than trichloroacetate (TCA) is produced. The addition of Fe2þ , Fe3þ and H2 O2 leads to an increased yield. NaCl was added as a source of chloride. We further examined the relationship between the structure and reactivity of phenolic substances, which can be considered as monomeric units of humic acids. Ethoxyphenol with built-in ethyl groups forms large amounts of DCA and TCA. The experiments with phenoxyacetic acid yielded large amounts of monochloroacetate (MCA). With other phenolic substances a ring cleavage was observed. Our investigations indicate that chloroacetates are formed abiotically from humic material and soils in addition to their known biotic mode of formation. Ó 2003 Published by Elsevier Science Ltd. Keywords: DCA; TCA; Humic acids; Phenolic model substances; Ring cleavage
1. Introduction In recent years haloacetic acids have attracted considerable scientific interest because of their possible environmental impact. All haloacetic acids are thought to be phytotoxic and some of them are presumably cancerogenic. There is evidence that forest trees at medium elevation in mountain ranges of Central Europe exposed to air masses, advectively transporting water as clouds or fog, exhibit pronounced phytotoxic symptoms
* Corresponding author. Tel.: +49-6221-544818; fax: +496221-545228. E-mail address: [email protected] (I.J. Fahimi).
(R€ ompp et al., 2001), suggesting haloacetic acids may be one of the sources of forest decline (Frank et al., 1995). Chloroacetates are formed via atmospheric breakdown of airborne C2 -chlorocarbons: tetrachloroethene, 1,1,1-trichloroethane and trichloroethene (McCulloch, 2002) which are used as solvents for degreasing and dry cleaning. Other anthropogenic sources include usage as a herbicide, incineration (Mowrer and Nordin, 1987), water chlorination and anti-fouling applications (Rook, 1977). TCA is widespread in precipitations in the Northern and Southern Hemisphere and, despite large emissions of possible precursors in the Northern Hemisphere, TCA concentrations in snow are not significantly different in arctic and subarctic regions than in Antarctica. A firn core from Antarctica with snow accumulated from the past 200 years exhibits haloacetic acids even in
0045-6535/03/$ - see front matter Ó 2003 Published by Elsevier Science Ltd. doi:10.1016/S0045-6535(03)00212-1
514
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
preindustrial layers (von Sydow et al., 2000). Moreover, TCA is also present in glacier ice of preindustrial origin from Northern Sweden (von Sydow et al., 1998) and in glacier ice samples from Monte Rosa, Switzerland (Haiber et al., 1996), dated around 1900, from the time before the mass production of chlorinated solvents, indicating that natural sources must also exist. Indeed, in recent years it has been argued that mass balances and experiments suggest a natural formation of haloacetates in soil (Hoekstra, 1999, Sch€ oler et al., 2003). Asplund et al. (1993) and Laturnus et al. (1995), succeeded in isolating a soil extract with halogenating potential. Walter and Ballschmiter (1992) reported chloroperoxidase (CPO)-mediated formation of chloroform from simple organic compounds, e.g. acetone, propionic acid and citric acid. Niedan et al. (2000) showed the formation of DCA and TCA from the halogenation of fulvic acid via CPO. Hoekstra et al. (1995) demonstrated the CPO-mediated formation of TCA and trichloromethane from humic acids. A correlation of chloroform and TCA, but unexpectedly not between TCA and other chlorinated solvents, in soil was found, indicating the natural formation of TCA in soils (Hoekstra et al., 1999a,b). The necessity of CPO for the formation of TCA was questioned by Haiber et al. (1996), who found TCA formation in experiments using only H2 O2 and humic acid. The TCA production in absence of CPO suggests that an abiotic formation might also occur. Indeed, Keppler et al. (2000) showed that an abiotic formation of volatile halocarbons occurs in soil. The aim of this study was to investigate the natural abiotic formation of chloroacetic acids in soil, using soil, humic acids and phenolic compounds which represent natural monomeric units of organic matter. A series of experiments were conducted to test the influence of H2 O2 , Fe2þ and Fe3þ on those abiotic formation processes.
Coniferous forest soil was collected in May 2001 from a rural area, Rotwasser, located in the Odenwald/ Germany (49°360 3900 N/8°530 1100 E); soil samples were taken from the A-horizon (Ah ). The soil was freeze-dried and stored in a freezer ()24 °C) until laboratory experiments were conducted. The soil was not sterilized. Diazomethane was prepared using the method of de Boer and Baker (1954).
2.2. Method The soil and the humic acid, with or without oxidizing agents (Fe2 (SO4 )3 , H2 O2 ) or FeCl2 4H2 O and chloride were added to 200 ml bidistilled water and shaken for 1 h. In the experiments with model substances 0.5 mmol model substance, 0.5 g NaCl, 100 mg Fe2 (SO4 )3 or FeCl2 4H2 O and 100 ll H2 O2 were combined, shaken for 1 h in 200 ml of aqua bidest. and analyzed. To extract the chloroacetic acids the mixture
2. Experimental section 2.1. Materials The humic acid, the TCA and the hydroquinone were obtained from Fluka, the resorcinol, MCA, DCA and the phloroglucinol were from Merck, the guaiacol, the phenoxyacetic acid and the ethoxyphenol were from Aldrich, the catechol was from Lancaster, the H2 SO4 and the FeCl2 4H2 O were from Riedel-de Haen, the Na2 SO4 was from Applichem, the Fe2 (SO4 )3 was from Fluka (heated at 400 °C for several hours); in the case of reactions of phenoxyacetic acid the Fe2 (SO4 )3 was from Alfa Aesar. The NaCl and the H2 O2 were from J.T. Baker. NaCl was heated overnight at 600 °C.
Fig. 1. (a) Chromatogram of DCA and TCA (1 lg) from a standard solution measured by GC-ECD. (b) Chromatogram of an experiment: catechol with Fe2 (SO4 )3 , NaCl and H2 O2: .
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
was acidified with concentrated H2 SO4 to pH 0.8 and shaken with 20 ml of t-butyl-methyl ether in a separatory funnel. After saturation with NaCl the aqueous solution was extracted two times and the solution was then centrifuged in the case of soil and humic material. The extracts were combined, dried over anhydrous Na2 SO4 and reduced to 1 ml in a graduated pear shaped flask. The obtained extracts were derivatized with diazomethane and measured by GC/ECD. The GC/ECD parameters are: GC: Fisons HRGC 8165, splitless, 1 ll, carrier gas: nitrogen, 2.0 ml/min, column: BP10, 30 m, 0.25 mm ID, 0.25 lm film thickness, temperature program: 35 °C (10 min), 8 °C min1 to 200 °C; ECD: Carlo-Erba ECD 400, 63 Ni, 270 °C, make-up gas: argon/ methane (95/5) 50 ml/min. The compounds were quantified via external standard. A chromatogram of a standard solution is shown in Fig. 1a; a chromatogram of an experiment with 55 mg catechol, 100 mg Fe2 (SO4 )3 , 0.5 g NaCl and 100 ll H2 O2 is shown in Fig. 1b. All experiments on model substances were carried out in triplicate, all other experiments were carried out in duplicate. The relative standard deviation ranged from 2.28% to 11.4% for DCA and from 2.6% to 13.1% for TCA; see captions of Figs. 7 and 8.
3. Results and discussion The formation of chloroacetic acids from soil, humic acid and model substances and the influence of oxidizing agents was examined.
515
3.1. Soil and Fluka humic acid suspended in water form chloroacetic acids The soil, which we used for our first set of experiments, formed chloroacetic acids, and this formation was increased when oxidizing agents were present (Fig. 2). Assuming that humic acid in soil is responsible for these reactions, we then used commercially available humic acid. The experiments with humic acid in water show that there was no need to add an oxidizing agent or halide source to produce DCA and TCA (Fig. 3). There is a positive correlation between haloacetic acid formation and humic acid content. DCA was always formed in higher concentrations than TCA, the ratio of DCA/TCA ranging between 1 and 3. Monochloroacetic acid (MCA) could not be measured, as it was below the detection limit (<1 lg). This linear relationship of humic acid and chlorinated acids formed could be explained by two effects: either by in vitro production or by contamination of the humic acid with DCA and TCA. DCA and TCA are omnipresent in the environment and have been shown to contaminate samples in significant amounts. If the commercially available humic acid is contaminated by DCA and TCA, either through exposure to air or from the production process, it is evident that their amounts would increase proportionately by increasing the humic acid concentration. On the other hand, there is also evidence for an in vitro production of DCA and TCA: an experiment with 1 g humic acid which was not shaken for an hour but directly acidified was performed, and indeed only about half of the amount of DCA and TCA,
Fig. 2. Formation of DCA and TCA from soil of the Rotwasser Reserve with 0.5 g NaCl.
516
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
.
.
.
Fig. 3. Formation of DCA and TCA as a function of amount of humic acid.
compared to the experiment in which the solution was shaken for an hour, was found. Some time-dependant measurements of DCA, using 1 g humic acid with varying time periods of shaking (5, 10, 40 min) showed an increase of DCA formation in the first 10 min. Another hint to the in vitro production is given by recent studies of Keppler et al. (2002). They demonstrated the in vitro formation of chloromethane, chloroethane and vinylchloride from the same humic acid which was used in this study. Thus, it seems that although some of the chloroacetic acids might be due to humic acid contamination, there is indeed evidence of genuine in-vitro formation as well. As far as the mode of formation of DCA and TCA is concerned, an explanation could be that the humic acid contained Fe (total iron: 0.9%) and chloride (1.5%) from the manufacturing process. Iron and chloride can react with the organic matter to form chloroacetic acids. The fact that the supplement of iron in the following experiments did not lead to a considerable increase in chloroacetic acid formation might be explained as follows: there are probably complexation sites in the humic acid, all of which were occupied by iron. When additional iron was supplemented the cations competed for these complexation sites, leading to a saturation in the following experiments. The iron reacts as an oxidant in a redox process with the humic acid and during this process organic compounds could be chlorinated, resulting in the production of short-chained halogenated acids. A possible reaction mechanism is discussed below.
3.2. Influence of H2 O2 , Fe3þ and Fe2þ on formation of chloroacetic acid Experiments were carried out to test the influence of Fe3þ , Fe2þ , H2 O2 and chloride (Figs. 2, 4 and 5). The addition of either H2 O2 or Fe3þ alone to the humic acid or soil had no significant effect on haloacetic acid formation (Figs. 2 and 4). However, when the oxidizing agents H2 O2 and Fe3þ were combined, a significant increase in haloacetic acid formation was observed with experiments using natural forest soil samples (Fig. 2). An addition of NaCl leads to a minor increase of DCA and a distinct increase of TCA (Fig. 4). Again, DCA was formed in higher concentrations than TCA, the ratio of DCA/TCA shifted to higher values (4.5) when using Fe3þ and H2 O2 . Those concentrations increased further with higher amounts of H2 O2 and Fe3þ , suggesting that Fe3þ was reduced to Fe2þ by redoxsensitive organic material and that probably a Fenton reaction (Fe2þ + H2 O2 ! Fe3þ + OH + OH ) was taking place, generating hydroxyl radicals and leading to a nonspecific chlorination of organic compounds when chloride is available in the system (Fig. 4). Hydroxyl radicals and chloride form an equilibrium system with the hypochlorous acid anion as well as with the chloride anion (Herrmann et al., 1999). The latter react with each other forming chlorine which induces the so-called ‘‘swimming pool chemistry’’. The C–Cl bond is probably created via a radical-related mechanism. To test this hypothesis, we performed experiments with Fe2þ and H2 O2 (the Fenton reagent) as well as with H2 O2 and UV-irradiation (Fig.
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
517
Fig. 4. The formation of DCA and TCA from humic acid increases dose-dependently with the concentration of H2 O2 and Fe2 (SO4 )3 .
Fig. 5. The formation of DCA and TCA from humic acid is increased with H2 O2 , FeCl2 4H2 O and UV-light.
5). A drastic increase of both DCA and TCA formation was found in both experiments, which indicates that OH radicals are probably involved in the reaction, either via a Fenton reaction, or the reaction of H2 O2 with UVirradiation.
An alternative explanation for the observed effect of the addition of Fe3þ could be the formation of a humic acid/Fe3þ -complex that acts similar to a heme group in the chloroperoxidase-mediated production of hypochlorous acid (Hoekstra, 2003).
518
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
Fig. 6. Model substances used for the experiments. (1) catechol, (2) resorcinol, (3) hydroquinone, (4) guaiacol, (5) ethoxyphenol, (6) phloroglucinol, (7) phenoxyacetic acid.
3.3. Model substances considered as natural organic matter form haloacetic acids The structure of humic substances in nature is only partly known. To reduce the complexity of natural systems, we carried out additional experiments with phenolic model substances, such as catechol, resorcinol, hydroquinone, guaiacol, ethoxyphenol, phloroglucinol and phenoxyacetic acid, which can be considered as monomeric units of the macromolecule humic acid (Fig. 6), with either Fe2þ and H2 O2 or Fe3þ and H2 O2 , to assess the effect of hydroxyl radical formation on the production of haloacetic acids. In the case of phenoxy-
acetic acid, when experiments with Fe3þ and H2 O2 were performed, a large amount of MCA was formed (11 lg in reactions with Fe3þ /H2 O2 ); phenoxyacetic acid and Fe2þ did not produce significant amounts of TCA compared to the blank (for blank measurements see the captions of Figs. 7 and 8). In experiments using only bidest. water, which was extracted and derivatized, a low background level was found (15 ng DCA and 2 ng TCA). In all experiments NaCl was supplemented as a source of chloride ions and DCA and TCA were formed, with higher values for DCA than TCA (Figs. 7 and 8). With ethoxyphenol there was a high formation of
Fig. 7. Amounts of DCA formed from model substances; the relative standard deviation (RSD) ranged from 2.28% to 11.4%. The blank value with Fe3þ (100 mg Fe2 (SO4 )3 , 0.5 g NaCl and 100 ll H2 O2 ) was 28 ng DCA and the blank value with Fe2þ (100 mg FeCl2 4H2 O, 0.5 g NaCl and 100 ll H2 O2: ) was 61.0 ng DCA.
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
519
Fig. 8. Amounts of TCA formed from model substances; the relative standard deviation (RSD) ranged from 2.06% to 13.1%. The blank value with Fe3þ (100 mg Fe2 (SO4 )3 , 0.5 g NaCl and 100 ll H2 O2 ) was 11 ng TCA and the blank value for Fe2þ (100 mg FeCl2 4H2 O, 0.5 g NaCl and 100 ll H2 O2: ) was 10 ng TCA.
Fig. 9. Reaction path of resorcinolic substances. This reaction scheme has been published by Hoekstra (1999).
haloacetic acids, as it has ethyl (C2 ) units which are suggested to be easily oxidized and halogenated, without the necessity of a cleavage of an aromatic ring system. But our experiments with other phenolic compounds showed a haloacetic acid formation also in the case of those compounds without an ethyl unit, suggesting that phenolic moieties react via a cleavage of the aromatic ring. A possible ring cleavage was also shown by Hoekstra (1999) for phenolic compounds with hydroxyl groups in meta-position (Fig. 9), which is a well-known reaction scheme elucidated during the chlorination of resorcinol in aqueous solutions.
Recently, Pracht et al. (2001) investigated the abiotic mineralization of phenolic model substances with Fe3þ . The liberation of CO2 points to a cleavage of the aromatic ring. The keto groups of the quinone-like structure are oxidized and expelled as CO2 . A similar process seems to take place here with an additional halogenation in the presence of halide ions, forming DCA and TCA. Phloroglucinol forms more DCA and TCA than resorcinol, possibly because it has three hydroxyl groups in meta-position, whereas resorcinol only possesses two hydroxyl groups in metaposition.
520
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
4. Conclusions
References
Our experiments show that DCA and TCA are formed also abiotically from soil and humic acids. A multitude of phenolic substances, which are common structural components of natural organic matter, shows the abiotic formation of DCA and TCA. There is no need for either a microbial contribution or a photochemical reaction. Ethoxyphenol is degraded relatively rapidly, as it already has an ethyl unit; for the other phenolic substances a cleavage of the aromatic ring has to take place before C2 -units are available. The oxidizing mixtures Fe2þ /H2 O2 and Fe3þ /H2 O2 lead to an enormous increase in haloacetate formation. Those concentrations increased further with higher amounts of H2 O2 and Fe3þ , suggesting that Fe3þ was reduced to Fe2þ by redox-sensitive organic material and that probably a Fenton reaction (Fe2þ + H2 O2 ! Fe3þ + OH + OH ) was taking place. Thus hydroxyl radicals were generated which are able to oxidize chloride to elemental chlorine when chloride is available in the system and leading to a non-specific chlorination of organic compounds. The C–Cl bond is probably formed via a radical-related mechanism. The importance of this new abiotic way of haloacetic acid formation becomes most evident since humus, iron and chloride are very widespread in soil. Nevertheless, a transfer of our laboratory experiments to model complex natural systems with varying environmental conditions is difficult. Additional studies with laboratory experiments and field work on natural soils are necessary to make general assumptions about these processes.
Asplund, G., Christiansen, J.V., Grimvall, A., 1993. Soil Biol. Biochem. 25, 41–46. de Boer, T.J., Baker, H.J., 1954. Recl. Trav. Chim. Pay-Bas 73, 229–234. Frank, H., Renschen, D., Klein, A., Scholl, H., 1995. J. High Resol. Chromatogr. 18, 83–88. Haiber, G., Jacob, G., Niedan, V.W., Nkusi, G., Sch€ oler, H.F., 1996. Chemosphere 33, 839–849. Herrmann, H., Ervens, B., Nowacki, P., Wolke, R., Zellner, R., 1999. Chemosphere 38, 1223–1232. Hoekstra, E.J., 1999. On the Natural Formation of Chlorinated Organic Compounds in Soil, Dissertation, Febodruk BV, Enschede. Hoekstra, E.J., Lassen, P., van Leeuven, J.G.E., de Leer, E.W.B., Carlsen, 1995. Naturally-Produced Organohalogens, 149–158. Hoekstra, E.J., de Leer, E.W.B., Brinkman, U.A.T., 1999a. Chemosphere 38, 551–563. Hoekstra, E.J., de Leer, E.W.B., Brinkman, U.A., 1999b. Chemosphere 38, 2875–2883. Hoekstra, E.J., 2003. Chemosphere, PII: S0045-6535(03)002133. Keppler, F., Eiden, R., Niedan, V., Pracht, J., Sch€ oler, H.F., 2000. Nature 403, 298–301. Keppler, F., Borchers, R., Pracht, J., Rheinberger, S., Sch€ oler, H.F., 2002. Environ. Sci. Technol. 36, 2479–2483. Laturnus, F., Mehrtens, G., Gron, C., 1995. Chemosphere 31, 3709–3719. Mowrer, J., Nordin, J., 1987. Chemosphere 26, 1181–1192. McCulloch, A., 2002. Chemosphere 47, 667–686. € berg, G., 2000. Chemosphere 41, Niedan, V., Pavasars, I., O 779–785. Pracht, J., Boenigk, J., Isenbeck-Schr€ oter, M., Keppler, F., Sch€ oler, H.F., 2001. Chemosphere 44, 613–619. R€ ompp, A., Klemm, O., Fricke, W., Frank, H., 2001. Environ. Sci. Technol. 35, 1294–1298. Rook, J.J., 1977. Environ. Sci. Technol. 11, 478–482. Sch€ oler, H.F., Keppier, F., Fahimi, I.J., Niedan, V.W., 2003. Chemosphere, PII: S0045-6535(03)00214-5. von Sydow, L., Boren, H., Grimvall, A., 1998. Haloacetates in Precipitation, Dissertation, 183, Link€ oping University, pp. 1–12. von Sydow, L., Nielsen, A.T., Grimvall, A.B., Boren, H.B., 2000. Environ. Sci. Technol. 34, 239–245. Walter, B., Ballschmiter, K., 1992. Fresenius J. Anal. Chem. 342, 827–833.
Acknowledgements We express our gratitude to G. Kilian and F. Bernsdorff for their help; this research was funded by the German Research Society (DFG), more specifically the GRK 273. We would like to thank the reviewers for their valuable contributions.
Formation of chloroacetic acids from soil, humic acid and phenolic moieties I.J. Fahimi *, F. Keppler, H.F. Sch€ oler Institute of Environmental Geochemistry, Heidelberg University, Im Neuenheimer Feld 236, D-69120 Heidelberg, Germany Received 15 April 2002; received in revised form 19 July 2002; accepted 23 July 2002
Abstract The mechanism of formation of chloroacetates, which are important toxic environmental substances, has been controversial. Whereas the anthropogenic production has been well established, a natural formation has also been suggested. In this study the natural formation of chloroacetic acids from soil, as well as from humic material which is present in soil and from phenolic model substances has been investigated. It is shown that chloroacetates are formed from humic material with a linear relationship between the amount of humic acid used and chloroacetates found. More dichloroacetate (DCA) than trichloroacetate (TCA) is produced. The addition of Fe2þ , Fe3þ and H2 O2 leads to an increased yield. NaCl was added as a source of chloride. We further examined the relationship between the structure and reactivity of phenolic substances, which can be considered as monomeric units of humic acids. Ethoxyphenol with built-in ethyl groups forms large amounts of DCA and TCA. The experiments with phenoxyacetic acid yielded large amounts of monochloroacetate (MCA). With other phenolic substances a ring cleavage was observed. Our investigations indicate that chloroacetates are formed abiotically from humic material and soils in addition to their known biotic mode of formation. Ó 2003 Published by Elsevier Science Ltd. Keywords: DCA; TCA; Humic acids; Phenolic model substances; Ring cleavage
1. Introduction In recent years haloacetic acids have attracted considerable scientific interest because of their possible environmental impact. All haloacetic acids are thought to be phytotoxic and some of them are presumably cancerogenic. There is evidence that forest trees at medium elevation in mountain ranges of Central Europe exposed to air masses, advectively transporting water as clouds or fog, exhibit pronounced phytotoxic symptoms
* Corresponding author. Tel.: +49-6221-544818; fax: +496221-545228. E-mail address: [email protected] (I.J. Fahimi).
(R€ ompp et al., 2001), suggesting haloacetic acids may be one of the sources of forest decline (Frank et al., 1995). Chloroacetates are formed via atmospheric breakdown of airborne C2 -chlorocarbons: tetrachloroethene, 1,1,1-trichloroethane and trichloroethene (McCulloch, 2002) which are used as solvents for degreasing and dry cleaning. Other anthropogenic sources include usage as a herbicide, incineration (Mowrer and Nordin, 1987), water chlorination and anti-fouling applications (Rook, 1977). TCA is widespread in precipitations in the Northern and Southern Hemisphere and, despite large emissions of possible precursors in the Northern Hemisphere, TCA concentrations in snow are not significantly different in arctic and subarctic regions than in Antarctica. A firn core from Antarctica with snow accumulated from the past 200 years exhibits haloacetic acids even in
0045-6535/03/$ - see front matter Ó 2003 Published by Elsevier Science Ltd. doi:10.1016/S0045-6535(03)00212-1
514
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
preindustrial layers (von Sydow et al., 2000). Moreover, TCA is also present in glacier ice of preindustrial origin from Northern Sweden (von Sydow et al., 1998) and in glacier ice samples from Monte Rosa, Switzerland (Haiber et al., 1996), dated around 1900, from the time before the mass production of chlorinated solvents, indicating that natural sources must also exist. Indeed, in recent years it has been argued that mass balances and experiments suggest a natural formation of haloacetates in soil (Hoekstra, 1999, Sch€ oler et al., 2003). Asplund et al. (1993) and Laturnus et al. (1995), succeeded in isolating a soil extract with halogenating potential. Walter and Ballschmiter (1992) reported chloroperoxidase (CPO)-mediated formation of chloroform from simple organic compounds, e.g. acetone, propionic acid and citric acid. Niedan et al. (2000) showed the formation of DCA and TCA from the halogenation of fulvic acid via CPO. Hoekstra et al. (1995) demonstrated the CPO-mediated formation of TCA and trichloromethane from humic acids. A correlation of chloroform and TCA, but unexpectedly not between TCA and other chlorinated solvents, in soil was found, indicating the natural formation of TCA in soils (Hoekstra et al., 1999a,b). The necessity of CPO for the formation of TCA was questioned by Haiber et al. (1996), who found TCA formation in experiments using only H2 O2 and humic acid. The TCA production in absence of CPO suggests that an abiotic formation might also occur. Indeed, Keppler et al. (2000) showed that an abiotic formation of volatile halocarbons occurs in soil. The aim of this study was to investigate the natural abiotic formation of chloroacetic acids in soil, using soil, humic acids and phenolic compounds which represent natural monomeric units of organic matter. A series of experiments were conducted to test the influence of H2 O2 , Fe2þ and Fe3þ on those abiotic formation processes.
Coniferous forest soil was collected in May 2001 from a rural area, Rotwasser, located in the Odenwald/ Germany (49°360 3900 N/8°530 1100 E); soil samples were taken from the A-horizon (Ah ). The soil was freeze-dried and stored in a freezer ()24 °C) until laboratory experiments were conducted. The soil was not sterilized. Diazomethane was prepared using the method of de Boer and Baker (1954).
2.2. Method The soil and the humic acid, with or without oxidizing agents (Fe2 (SO4 )3 , H2 O2 ) or FeCl2 4H2 O and chloride were added to 200 ml bidistilled water and shaken for 1 h. In the experiments with model substances 0.5 mmol model substance, 0.5 g NaCl, 100 mg Fe2 (SO4 )3 or FeCl2 4H2 O and 100 ll H2 O2 were combined, shaken for 1 h in 200 ml of aqua bidest. and analyzed. To extract the chloroacetic acids the mixture
2. Experimental section 2.1. Materials The humic acid, the TCA and the hydroquinone were obtained from Fluka, the resorcinol, MCA, DCA and the phloroglucinol were from Merck, the guaiacol, the phenoxyacetic acid and the ethoxyphenol were from Aldrich, the catechol was from Lancaster, the H2 SO4 and the FeCl2 4H2 O were from Riedel-de Haen, the Na2 SO4 was from Applichem, the Fe2 (SO4 )3 was from Fluka (heated at 400 °C for several hours); in the case of reactions of phenoxyacetic acid the Fe2 (SO4 )3 was from Alfa Aesar. The NaCl and the H2 O2 were from J.T. Baker. NaCl was heated overnight at 600 °C.
Fig. 1. (a) Chromatogram of DCA and TCA (1 lg) from a standard solution measured by GC-ECD. (b) Chromatogram of an experiment: catechol with Fe2 (SO4 )3 , NaCl and H2 O2: .
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
was acidified with concentrated H2 SO4 to pH 0.8 and shaken with 20 ml of t-butyl-methyl ether in a separatory funnel. After saturation with NaCl the aqueous solution was extracted two times and the solution was then centrifuged in the case of soil and humic material. The extracts were combined, dried over anhydrous Na2 SO4 and reduced to 1 ml in a graduated pear shaped flask. The obtained extracts were derivatized with diazomethane and measured by GC/ECD. The GC/ECD parameters are: GC: Fisons HRGC 8165, splitless, 1 ll, carrier gas: nitrogen, 2.0 ml/min, column: BP10, 30 m, 0.25 mm ID, 0.25 lm film thickness, temperature program: 35 °C (10 min), 8 °C min1 to 200 °C; ECD: Carlo-Erba ECD 400, 63 Ni, 270 °C, make-up gas: argon/ methane (95/5) 50 ml/min. The compounds were quantified via external standard. A chromatogram of a standard solution is shown in Fig. 1a; a chromatogram of an experiment with 55 mg catechol, 100 mg Fe2 (SO4 )3 , 0.5 g NaCl and 100 ll H2 O2 is shown in Fig. 1b. All experiments on model substances were carried out in triplicate, all other experiments were carried out in duplicate. The relative standard deviation ranged from 2.28% to 11.4% for DCA and from 2.6% to 13.1% for TCA; see captions of Figs. 7 and 8.
3. Results and discussion The formation of chloroacetic acids from soil, humic acid and model substances and the influence of oxidizing agents was examined.
515
3.1. Soil and Fluka humic acid suspended in water form chloroacetic acids The soil, which we used for our first set of experiments, formed chloroacetic acids, and this formation was increased when oxidizing agents were present (Fig. 2). Assuming that humic acid in soil is responsible for these reactions, we then used commercially available humic acid. The experiments with humic acid in water show that there was no need to add an oxidizing agent or halide source to produce DCA and TCA (Fig. 3). There is a positive correlation between haloacetic acid formation and humic acid content. DCA was always formed in higher concentrations than TCA, the ratio of DCA/TCA ranging between 1 and 3. Monochloroacetic acid (MCA) could not be measured, as it was below the detection limit (<1 lg). This linear relationship of humic acid and chlorinated acids formed could be explained by two effects: either by in vitro production or by contamination of the humic acid with DCA and TCA. DCA and TCA are omnipresent in the environment and have been shown to contaminate samples in significant amounts. If the commercially available humic acid is contaminated by DCA and TCA, either through exposure to air or from the production process, it is evident that their amounts would increase proportionately by increasing the humic acid concentration. On the other hand, there is also evidence for an in vitro production of DCA and TCA: an experiment with 1 g humic acid which was not shaken for an hour but directly acidified was performed, and indeed only about half of the amount of DCA and TCA,
Fig. 2. Formation of DCA and TCA from soil of the Rotwasser Reserve with 0.5 g NaCl.
516
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
.
.
.
Fig. 3. Formation of DCA and TCA as a function of amount of humic acid.
compared to the experiment in which the solution was shaken for an hour, was found. Some time-dependant measurements of DCA, using 1 g humic acid with varying time periods of shaking (5, 10, 40 min) showed an increase of DCA formation in the first 10 min. Another hint to the in vitro production is given by recent studies of Keppler et al. (2002). They demonstrated the in vitro formation of chloromethane, chloroethane and vinylchloride from the same humic acid which was used in this study. Thus, it seems that although some of the chloroacetic acids might be due to humic acid contamination, there is indeed evidence of genuine in-vitro formation as well. As far as the mode of formation of DCA and TCA is concerned, an explanation could be that the humic acid contained Fe (total iron: 0.9%) and chloride (1.5%) from the manufacturing process. Iron and chloride can react with the organic matter to form chloroacetic acids. The fact that the supplement of iron in the following experiments did not lead to a considerable increase in chloroacetic acid formation might be explained as follows: there are probably complexation sites in the humic acid, all of which were occupied by iron. When additional iron was supplemented the cations competed for these complexation sites, leading to a saturation in the following experiments. The iron reacts as an oxidant in a redox process with the humic acid and during this process organic compounds could be chlorinated, resulting in the production of short-chained halogenated acids. A possible reaction mechanism is discussed below.
3.2. Influence of H2 O2 , Fe3þ and Fe2þ on formation of chloroacetic acid Experiments were carried out to test the influence of Fe3þ , Fe2þ , H2 O2 and chloride (Figs. 2, 4 and 5). The addition of either H2 O2 or Fe3þ alone to the humic acid or soil had no significant effect on haloacetic acid formation (Figs. 2 and 4). However, when the oxidizing agents H2 O2 and Fe3þ were combined, a significant increase in haloacetic acid formation was observed with experiments using natural forest soil samples (Fig. 2). An addition of NaCl leads to a minor increase of DCA and a distinct increase of TCA (Fig. 4). Again, DCA was formed in higher concentrations than TCA, the ratio of DCA/TCA shifted to higher values (4.5) when using Fe3þ and H2 O2 . Those concentrations increased further with higher amounts of H2 O2 and Fe3þ , suggesting that Fe3þ was reduced to Fe2þ by redoxsensitive organic material and that probably a Fenton reaction (Fe2þ + H2 O2 ! Fe3þ + OH + OH ) was taking place, generating hydroxyl radicals and leading to a nonspecific chlorination of organic compounds when chloride is available in the system (Fig. 4). Hydroxyl radicals and chloride form an equilibrium system with the hypochlorous acid anion as well as with the chloride anion (Herrmann et al., 1999). The latter react with each other forming chlorine which induces the so-called ‘‘swimming pool chemistry’’. The C–Cl bond is probably created via a radical-related mechanism. To test this hypothesis, we performed experiments with Fe2þ and H2 O2 (the Fenton reagent) as well as with H2 O2 and UV-irradiation (Fig.
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
517
Fig. 4. The formation of DCA and TCA from humic acid increases dose-dependently with the concentration of H2 O2 and Fe2 (SO4 )3 .
Fig. 5. The formation of DCA and TCA from humic acid is increased with H2 O2 , FeCl2 4H2 O and UV-light.
5). A drastic increase of both DCA and TCA formation was found in both experiments, which indicates that OH radicals are probably involved in the reaction, either via a Fenton reaction, or the reaction of H2 O2 with UVirradiation.
An alternative explanation for the observed effect of the addition of Fe3þ could be the formation of a humic acid/Fe3þ -complex that acts similar to a heme group in the chloroperoxidase-mediated production of hypochlorous acid (Hoekstra, 2003).
518
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
Fig. 6. Model substances used for the experiments. (1) catechol, (2) resorcinol, (3) hydroquinone, (4) guaiacol, (5) ethoxyphenol, (6) phloroglucinol, (7) phenoxyacetic acid.
3.3. Model substances considered as natural organic matter form haloacetic acids The structure of humic substances in nature is only partly known. To reduce the complexity of natural systems, we carried out additional experiments with phenolic model substances, such as catechol, resorcinol, hydroquinone, guaiacol, ethoxyphenol, phloroglucinol and phenoxyacetic acid, which can be considered as monomeric units of the macromolecule humic acid (Fig. 6), with either Fe2þ and H2 O2 or Fe3þ and H2 O2 , to assess the effect of hydroxyl radical formation on the production of haloacetic acids. In the case of phenoxy-
acetic acid, when experiments with Fe3þ and H2 O2 were performed, a large amount of MCA was formed (11 lg in reactions with Fe3þ /H2 O2 ); phenoxyacetic acid and Fe2þ did not produce significant amounts of TCA compared to the blank (for blank measurements see the captions of Figs. 7 and 8). In experiments using only bidest. water, which was extracted and derivatized, a low background level was found (15 ng DCA and 2 ng TCA). In all experiments NaCl was supplemented as a source of chloride ions and DCA and TCA were formed, with higher values for DCA than TCA (Figs. 7 and 8). With ethoxyphenol there was a high formation of
Fig. 7. Amounts of DCA formed from model substances; the relative standard deviation (RSD) ranged from 2.28% to 11.4%. The blank value with Fe3þ (100 mg Fe2 (SO4 )3 , 0.5 g NaCl and 100 ll H2 O2 ) was 28 ng DCA and the blank value with Fe2þ (100 mg FeCl2 4H2 O, 0.5 g NaCl and 100 ll H2 O2: ) was 61.0 ng DCA.
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
519
Fig. 8. Amounts of TCA formed from model substances; the relative standard deviation (RSD) ranged from 2.06% to 13.1%. The blank value with Fe3þ (100 mg Fe2 (SO4 )3 , 0.5 g NaCl and 100 ll H2 O2 ) was 11 ng TCA and the blank value for Fe2þ (100 mg FeCl2 4H2 O, 0.5 g NaCl and 100 ll H2 O2: ) was 10 ng TCA.
Fig. 9. Reaction path of resorcinolic substances. This reaction scheme has been published by Hoekstra (1999).
haloacetic acids, as it has ethyl (C2 ) units which are suggested to be easily oxidized and halogenated, without the necessity of a cleavage of an aromatic ring system. But our experiments with other phenolic compounds showed a haloacetic acid formation also in the case of those compounds without an ethyl unit, suggesting that phenolic moieties react via a cleavage of the aromatic ring. A possible ring cleavage was also shown by Hoekstra (1999) for phenolic compounds with hydroxyl groups in meta-position (Fig. 9), which is a well-known reaction scheme elucidated during the chlorination of resorcinol in aqueous solutions.
Recently, Pracht et al. (2001) investigated the abiotic mineralization of phenolic model substances with Fe3þ . The liberation of CO2 points to a cleavage of the aromatic ring. The keto groups of the quinone-like structure are oxidized and expelled as CO2 . A similar process seems to take place here with an additional halogenation in the presence of halide ions, forming DCA and TCA. Phloroglucinol forms more DCA and TCA than resorcinol, possibly because it has three hydroxyl groups in meta-position, whereas resorcinol only possesses two hydroxyl groups in metaposition.
520
I.J. Fahimi et al. / Chemosphere 52 (2003) 513–520
4. Conclusions
References
Our experiments show that DCA and TCA are formed also abiotically from soil and humic acids. A multitude of phenolic substances, which are common structural components of natural organic matter, shows the abiotic formation of DCA and TCA. There is no need for either a microbial contribution or a photochemical reaction. Ethoxyphenol is degraded relatively rapidly, as it already has an ethyl unit; for the other phenolic substances a cleavage of the aromatic ring has to take place before C2 -units are available. The oxidizing mixtures Fe2þ /H2 O2 and Fe3þ /H2 O2 lead to an enormous increase in haloacetate formation. Those concentrations increased further with higher amounts of H2 O2 and Fe3þ , suggesting that Fe3þ was reduced to Fe2þ by redox-sensitive organic material and that probably a Fenton reaction (Fe2þ + H2 O2 ! Fe3þ + OH + OH ) was taking place. Thus hydroxyl radicals were generated which are able to oxidize chloride to elemental chlorine when chloride is available in the system and leading to a non-specific chlorination of organic compounds. The C–Cl bond is probably formed via a radical-related mechanism. The importance of this new abiotic way of haloacetic acid formation becomes most evident since humus, iron and chloride are very widespread in soil. Nevertheless, a transfer of our laboratory experiments to model complex natural systems with varying environmental conditions is difficult. Additional studies with laboratory experiments and field work on natural soils are necessary to make general assumptions about these processes.
Asplund, G., Christiansen, J.V., Grimvall, A., 1993. Soil Biol. Biochem. 25, 41–46. de Boer, T.J., Baker, H.J., 1954. Recl. Trav. Chim. Pay-Bas 73, 229–234. Frank, H., Renschen, D., Klein, A., Scholl, H., 1995. J. High Resol. Chromatogr. 18, 83–88. Haiber, G., Jacob, G., Niedan, V.W., Nkusi, G., Sch€ oler, H.F., 1996. Chemosphere 33, 839–849. Herrmann, H., Ervens, B., Nowacki, P., Wolke, R., Zellner, R., 1999. Chemosphere 38, 1223–1232. Hoekstra, E.J., 1999. On the Natural Formation of Chlorinated Organic Compounds in Soil, Dissertation, Febodruk BV, Enschede. Hoekstra, E.J., Lassen, P., van Leeuven, J.G.E., de Leer, E.W.B., Carlsen, 1995. Naturally-Produced Organohalogens, 149–158. Hoekstra, E.J., de Leer, E.W.B., Brinkman, U.A.T., 1999a. Chemosphere 38, 551–563. Hoekstra, E.J., de Leer, E.W.B., Brinkman, U.A., 1999b. Chemosphere 38, 2875–2883. Hoekstra, E.J., 2003. Chemosphere, PII: S0045-6535(03)002133. Keppler, F., Eiden, R., Niedan, V., Pracht, J., Sch€ oler, H.F., 2000. Nature 403, 298–301. Keppler, F., Borchers, R., Pracht, J., Rheinberger, S., Sch€ oler, H.F., 2002. Environ. Sci. Technol. 36, 2479–2483. Laturnus, F., Mehrtens, G., Gron, C., 1995. Chemosphere 31, 3709–3719. Mowrer, J., Nordin, J., 1987. Chemosphere 26, 1181–1192. McCulloch, A., 2002. Chemosphere 47, 667–686. € berg, G., 2000. Chemosphere 41, Niedan, V., Pavasars, I., O 779–785. Pracht, J., Boenigk, J., Isenbeck-Schr€ oter, M., Keppler, F., Sch€ oler, H.F., 2001. Chemosphere 44, 613–619. R€ ompp, A., Klemm, O., Fricke, W., Frank, H., 2001. Environ. Sci. Technol. 35, 1294–1298. Rook, J.J., 1977. Environ. Sci. Technol. 11, 478–482. Sch€ oler, H.F., Keppier, F., Fahimi, I.J., Niedan, V.W., 2003. Chemosphere, PII: S0045-6535(03)00214-5. von Sydow, L., Boren, H., Grimvall, A., 1998. Haloacetates in Precipitation, Dissertation, 183, Link€ oping University, pp. 1–12. von Sydow, L., Nielsen, A.T., Grimvall, A.B., Boren, H.B., 2000. Environ. Sci. Technol. 34, 239–245. Walter, B., Ballschmiter, K., 1992. Fresenius J. Anal. Chem. 342, 827–833.
Acknowledgements We express our gratitude to G. Kilian and F. Bernsdorff for their help; this research was funded by the German Research Society (DFG), more specifically the GRK 273. We would like to thank the reviewers for their valuable contributions.