- Email: [email protected]
An experimental ecotoxicological study and its application to the behavioural study of organic mercury (CH3HgCl) in the environment: influence of temperature and pH
Chemosphere 49 (2002) 1399–1405 www.elsevier.com/locate/chemosphere
An experimental ecotoxicological study and its application to the behavioural study of organic mercury (CH3HgCl) in the environment: influence of temperature and pH S. Jahanbakht
a,b,* ,
F. Livardjani a, A. Jaeger
a
a
b
Laboratoire du Centre Antipoison, Universit e Louis Pasteur de Strasbourg, Strasbourg Cedex 67005, France AD Scientifique Centre d’ etudes, d’analyses et diagnostics en toxicologie de l’environnement, BP 214 R5 67005 Strasbourg, France Received 8 May 2001; received in revised form 8 August 2002; accepted 23 August 2002
Abstract Laboratory experiments were carried out to study the influence of temperature (24, 28 and 30 °C) and pH (1–10) on organic mercury (CH3 HgCl) transfer and accumulation in an experimental ecotoxicological model. We followed the evolution of CH3 HgCl in a basic model (water þ air) by varying temperature and pH. In a second step, we completed the model by adding sediment and fish. We added CH3 HgCl to water at the beginning of each experiment which was repeated at least three times. Results demonstrated that mercury was released from methylmercury into the air regardless of water pH and its concentration in the air increased with increasing pH. By contrast, in presence of sediment, almost all the mercury was fixed onto the sediment and no mercury was traced in air or in water. Interestingly, in the presence of sediment, the life span of fish under methylmercury exposure lasted longer despite their higher mercury body level content at their death. These results indicate that water is a bad exposure indicator for aquatic pollution. In case of chronic pollution, sediments, fish and aquatic plants are more appropriate indicators. Ó 2002 Elsevier Science Ltd. All rights reserved. Keywords: Methylmercury; Air; Water; Sediment; pH; Fish
1. Introduction Many mercury applications in industries entailed several cases of contamination in the past, as in Japan (Minamata) or Iraq, and today they are still a serious issue in countries as Brazil (Lacerda, 1997; Moreira and Pivetta, 1997). Mercury is one of the few metals to have caused human deaths related to food contamination,
* Corresponding author. Address: Laboratoire du Centre Antipoison, Universite Louis Pasteur de Strasbourg, Strasbourg Cedex 67005, France. Tel.: +33-388-24-33-17; fax: +33390-24-40-07. E-mail address: [email protected]. fr (S. Jahanbakht).
and over the past 40 years, it is estimated to have caused the death of 1400 people and affected 20 000 people worldwide (DÕItri, 1992). Mercury can be released from anthropogenic sources into the atmosphere in various chemical forms. Mercury most common forms are Hg0 , Hg2þ , particle Hg and HgO (Marins and Tonietto, 1995). The effects of mercury and its derivatives, especially of methylmercury, on the environment and aquatic media, have been widely discussed in the literature (Lacerda, 1997; Moreira and Pivetta, 1997). Yet, little has been reported on the mechanisms of mercury release into the atmosphere from water or soil. In this paper, we investigated the evolution of mercury in the air after mercury water pollution and we analysed the influence of several biotic and abiotic parameters. Implications of this study are discussed.
0045-6535/02/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 5 - 6 5 3 5 ( 0 2 ) 0 0 5 0 6 - 4
1400
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405
2.1.2. Air compartment The system contained a 100 l volume of room air. After cleaning, we analysed the air of the system so as to check that no mercury traces were left.
2. Material and methods 2.1. Structure of the experimental model The main characteristics of the compartments introduced into the experimental system (glass tank, 200 l capacity) are presented in Fig. 1. To minimize Hg adsorption, we chose materials most inert to mercury, such as Teflon tubes and glass walls. For air and water sampling, we used two pumps with a sampling site on each to create permanent flows, one for air and one for water. We then connected the mercury analyser to the air sampling site and air was automatically sampled every fifteen minutes throughout the duration of the experiment. We cleaned the model (sides of the glass tank) after each experiment with 5% diluted nitric acid and aired it for several days.
2.1.1. Water compartment A 100 l volume of tap water from the Strasbourg network was introduced into the system and was not renewed during the experiments. The water initial physico-chemical characteristics were checked by the Direction Departementale des Affaires Sanitaires et Sociales de la ville de Strasbourg. They were as follows: chloride (Cl ) 70 mg/l, calcium (Ca2þ ) 97.7 mg/l, potassium (Kþ ) 3.8 mg/l, sodium (Naþ ) 49.8 mg/l, NO 2 2 0.04 mg/l, NO 42 mg/l. At time zero, 3 9 mg/l, SO4 approximately 0.1 mg/l of CH3 HgCl (purity: 99%; purchased from Riedel-de-Ha€en) was added to the water.
2.1.3. Sediment compartment For this study, we used two types of sediments, natural sediment and sand without organic matter. For each experiment, we used new sand and sediment samples with exactly the same characteristics, the same weight 1 and from the same source. The sediment was taken from a local gravel pit (Wittisheim, France) and its geochemical and granulometric characteristics were determined by the Institut de Geologie de Strasbourg (ULP, Strasbourg University, France). It was mainly composed of quartz (SiO2 ) (40%), limestone (40%) and white mica (20%) and organic matters such as wood residues, plants, insects and shells. The particle size of the sediment was about 0.05 mm. The total carbon content, determined with the Carbon Sulfur Analyser CS 125 (LECO), was 14%. The natural level of total mercury was 52:7 19 ng Hg/g dry weight. The sand used was taken from a river of fluvial origin and the average size of the particles was 1 mm (2–0.5 mm), of which 95% were square while less than 5% were round. The sand was composed of 80% quartz, 10–15% feldspath and the remaining were ferro-magnesian minerals (pyroxene and amphiboles). No mercury traces were found in this sand. 2.1.4. Fish We selected Scardinius erythrophtalmus (of the cyprimide family) which lives in quiet rivers, ponds and lakes.
Fig. 1. Diagram of the model.
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405
For our study, we used 4–5 month-old farm-raised fish, with an average weight of 13:312 0:041 g and a size of 12:4 0:5 cm. To avoid risks of food contamination, they had not been fed during the 30 day-long experiment and neither had the control group. For consistent results, we used a batch of fish. At time zero of the experiment, we introduced a group of nine fish into the model, and every 10th day a batch of three fish was sacrificed. For each batch, the average background level (Hg natural content) was determined. The aim was to find mercury transfer in the model. Since we were looking for total mercury in fish we did not care for mercury distribution in the organs; so after mixing the fish we analysed the mixture for total mercury traces. We repeated this type of experiments three times:
1401
ple of cold vapour atomic absorption (Hatch and Ott, 1968). Atmospheric mercury and inorganic forms of mercury in water were directly measured with the above described prototype (Livardjani et al., 1995). For organic forms, we carried out mineralisation according to the nature of the sample and optimised its parameters to reduce loss by volatilisation (see below). We used Merck reagents containing a minimum of mercury as recommended for mercury analysis (Heimburger et al., 1986; Jahanbakht et al., 1998). 2.3. Air analysis We directly sampled the air of the system three times every 15 min, then analysed total mercury in each sample and the average was taken (Figs. 2 and 3).
2.2. Mercury analyser 2.4. Sediment mineralisation In this study, we carried out the experiments with a mercury analyser developed in the laboratory and slightly modified; in this case, we used a 30 cm-long and 2.2 cm-diameter-measuring cell with a detection limit of 90 ng Hg/m3 in the air and 15 pg Hg in absolute weight with a 5% uncertainty determination (Jahanbakht et al., 1998). This mercury analyser is based on the princi-
The level of moisture in the sediment to be analysed was measured in each of the samples taken separately and assessed from triplicates. The sediment being a complex matrix, a higher temperature and a longer time were required to obtain complete mineralisation (Jahanbakht et al., 1998). The samples were placed into an
Fig. 2. Hg evolution in the air at different temperatures (pH ¼ 7:2; CH3 HgCl). Data points average of three independent experiments.
Fig. 3. Hg concentration according to pH in the air (T ¼ 24 °C, CH3 HgCl). Data points average of three independent experiments.
1402
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405
acid medium (HNO3 þ H2 SO4 , 6:3 v/v) in a Pyrex flask during 24 h at room temperature, and were then mineralised with a reflux cooling system at boiling temperature (about 110 °C) for 24 h. Oxidation was obtained by adding 5 ml KMnO4 (0.44 M) to the mineralised product. A solution of oxalic acid (0.5 M) was used to reduce permanganate until decoloration (Livardjani et al., 1996; Jahanbakht et al., 1998). Mercury was reduced to Hg0 by stannous chloride (5 ml of a 10% solution in double-distilled water). Mercury vapour (Hg0 ) was purged with an argon current, (Heimburger et al., 1986). The accuracy of this method was checked with the analysis of standard reference materials (e.g. BCR––Bureau Communautaire de Reference de la Commission des Communautes Europeennes) and was in good agreement with certified values (Livardjani et al., 1996). Results are expressed in concentration (ng Hg/g (dry weight)) and total mercury burden (mg Hg for the total amount of sediment). 2.5. Fish mineralisation The mineralisation parameters were optimised on aliquot fractions (500 mg) of fish mixtures (S. erythrophtalmus). The best results were obtained using a sulfonitric mixture (HNO3 þ H2 SO4 , 6:4 v/v). This operation, carried out in a reflux cooling system, lasted 1 h at 60 °C. Then, 3 ml of a KMnO4 solution (0.44 M) were added to oxidize mercury. The excess permanganate was removed by adding about 2 ml of oxalic acid (0.5 M). The method was checked against the standard addition method. 95% of the mercury was recovered (Jahanbakht et al., 1998). Results are expressed in concentration (ng Hg/g (fresh weight)). 2.6. Water mineralisation For mercury recuperation in environmental water, samples were treated using a mixture of nitric and sulfuric acid (3 ml HNO3 , 2 ml H2 SO4 ), at 20 °C for 10 min with a 200 ll sample. For mercury oxidation, 5 ml of a potassium permanganate solution (0.44 M) were added. The excess permanganate was eliminated by adding 5 ml of oxalic acid (0.5 M).
3. Results 3.1. Temperature influence Initial mercury concentration in the water was 0.10 mg/l. pH was constant at 7.2. Hg concentrations in the air and the water of the system were measured every 15 min during 48 h. Mercury concentration variations were
assessed at three different temperatures: 24, 28 and 30 °C. Results are shown in Fig. 2. A strong influence of temperature was observed. The delay of mercury release in the air of the system varied according to temperature. Indeed, when the water temperature was 24 °C, mercury was detected from the 4th h (240 min) whereas at 28 and 30 °C, it was detected from the 8th (480 min) and the 20th h (1200 min) respectively (Fig. 2). The mercury concentration in the air strictly increased with temperature and reached an equilibrium between 20 and 40 h. The value was related to the water temperature (0:155 0:06 mg/m3 at 24 °C, 11:52 0:05 mg/m3 at 28 °C, and 20:5 0:5 mg/m3 at 30 °C (Fig. 2)). 3.2. pH influence To study the pH influence, the temperature was kept constant. The initial water concentration was 0.09 mg/l. Each experiment was repeated at different pH values ranging from 1 to 12, using hydrochloric acid (HCl) and sodium hydroxide (NaOH), and was carried out over a 48-h period. The Hg concentrations in the air of the system found for each pH at the 48th h (after reaching equilibrium) are presented in Fig. 3. Methylmercury was released in the air regardless of the pH. Yet, this release was more intense in a basic medium. It was demonstrated (Driscol et al., 1995) that dissolved organic carbon (DOC) stabilizes methylmercury (DOC-HgCHþ 3 ) in solution. Besides, the presence of aluminium or other ions in the solution (competition principle) favours the transfer of dissolved methylmercury (HgCHþ 3 ) to the particle phase (Henry et al., 1996) according to the following reaction: 3þ DOC-HgCHþ ! DOC-Al3þ þ HgCHþ 3 þ Al 3
Dimethylmercury can be formed from HgCHþ 3 . This formation is favoured in a basic medium (Beijer and Jernelov, 1979; Baker et al., 1981). Due to its volatility, dimethylmerury is released in the air where it changes into Hg0 . 3.3. Sediment influence To assess the influence of sediment on the evolution of mercury in the air of our basic system (water þ air), we adjusted the initial water concentration to 0.050 mg/l, temperature was kept constant at 24 °C and pH at 7.2. Then we added 8.501 kg of natural sediment (dry weight) to the system with an initial mercury concentration of 52:7 19 ng Hg/g of sediment. During a 48 h monitoring, no mercury traces were detected in either compartment within the detection limit (90 ng Hg/m3 ). At the experiment completion, the sediment contained 606:99 83 ng Hg/g of sediment; the
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405
1403
Fig. 4. Mercury amount traced in the sediment according to pH (T ¼ 24 °C). Data points average of three independent experiments.
total amount of sediment had thus adsorbed 5.16 mg of Hg, which means more than 95% of the mercury added to water (5.4 mg). Identical experiments carried out at 28 and 30 °C and various pH (1–10) showed that pH variation had no effect on the methylmercury adsorption by the sediment (Fig. 4). No significant difference was noted with the variation of pH, and the total amount of Hg added to the water was rapidly fixed onto the sediment. 3.4. Sand influence The same experiments were repeated (T ¼ 24 °C, pH ¼ 7:2), with sand instead of sediment. It was noted that nearly 17% of the mercury added to the water was fixed by sand and this amount remained stable during the month-long period while the system was monitored. In the presence of sand, the air was contaminated and mercury concentration varied as in a basic system (water þ air). 3.5. Fish influence The protocol was first simplified and fish were added to CH3 HgCl contaminated water at a 0.095 mg/l concentration. Then, sediment was added under the same conditions (T ¼ 24 1 °C, pH ¼ 7:2). In order to compare the two different aquatic matrices, we used sand instead of sediment. We repeated the following experiment three times. For each experiment, we used nine fish, three of which were to be sacrificed (five aliquots determined for each fish) every 10 days after the beginning of exposure. The control fish were maintained in the same conditions (temperature, pH) throughout the time of the experiment. No fish mortality was recorded in the control. 3.6. System: water þ fish We placed nine fish in the system and all died before the day foreseen for their sacrifice; they died within 9 h
following the addition of mercury to the water and had an average concentration of 0:696 0:103 mg Hg/kg of fish (0.474–0.824). No specific organ of the fish was analysed and the result shows the concentration of the whole fish. They had a restless behaviour before dying; hourly, they had serious behavioural problems for 5–10 min, with a loss of swimming ability, spinning round, head turned downwards and motionless, and their tail making circular movements. After a brutal stop, they seemed to swim normally again but actually went straight into the sides of the aquarium. The frequency of these crises increased until their death. 3.7. System: water þ sediment þ fish To carry out this experiment, temperature was set at 24 °C and pH at 7.2. After adjusting the water to obtain a 0.095 mg/l mercury concentration, we added 8.501 kg of natural sediment to the system and introduced 9 fish. Once the sediment had been added into the system, the fish remained alive about 35 h but had an average of 1:327 0:138 mg Hg/kg fish (1.218–1.677). The sediment contained 1:320 0:029 lg Hg/g sediment when the experiment was completed. 3.8. System: water þ sand þ fish The preceding experiment was repeated but instead of sediment, the same amount of sand was taken. After about 7 h, the fish showed the same symptoms as reported previously. All of them died between 8 and 20 h and had an average mercury concentration of 0:689 0:202 mg Hg/kg fish.
4. Discussion These findings show that the processes of mercury transformation and bioaccumulation in the environment are very complex. When pH and temperature are constant, an increase of the mercury amount in the air
1404
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405
related to time is noted. When pH is constant and temperature varies, the mercury amount in the air varies from 0.1% to 10% with increasing temperature. Schuhmacher et al. (1993) found that mercury concentrations in water and sediment were lower in summer than in winter. Airey (1982) noted that mercury concentrations in the air increased when temperature increased while Saouter (1990) reported a sharp fall of mercury concentration in the water related to temperature increase. Similar results were reported by Hutagalung (1987) and Prabhu and Hamdy (1977). However, no such correlation between temperature and total mercury amount in the water was observed by Frenet (1979). In these studies, the chemical form of mercury was not mentioned. When pH varied and temperature was kept constant, whatever the rate of water acidity, mercury was released into the air and its concentration increased according to pH. In our previous paper (Jahanbakht et al., 1998), we saw that when the source of water pollution was HgCl2 , Hg remained stable in the water in an acid medium (pH < 5) and the air was not contaminated. On the contrary, at pH > 5, mercury was released into the air of the system and reached an equilibrium value which was temperature-related. Thus, air contamination is related to the chemical form of the water pollutant (organic or inorganic). Schindler et al. (1980) reported an increase of mercury concentration in water and a decrease in sediment when pH decreased. Messaitfa (1997) noted a Hg increase in water when pH ranged between 7.4 and 7.8 and a decrease when it ranged between 7.8 and 8.4. In the presence of sediment, almost all the amount of methylmercury chloride added to the water is rapidly fixed onto the sediment and the air is not contaminated. This phenomenon was also observed with inorganic mercury (Jahanbakht et al., 1998). However, when sand is present, approximately 17% of the mercury added to the water is fixed onto the sand and the air is contaminated. Indeed, mercury concentrations are strongly affected by the sediment granulometry and this is observed in our case since the granulometry of sediments is higher in fine particles (sediment) than in sand. The influence of organic matters should not be minimized. MacNaughton and James (1974) found that Hg2þ adsorption on SiO2 increased sharply at pH 2 to 3, and in the absence of Cl . Barrow and Cox (1992) demonstrated that in the absence of Cl , the maximal adsorption was just above pH 4. In both studies, addition of Cl to the medium significantly reduced mercury adsorption. Yin et al. (1996) however showed that at low pH, Hg2þ was adsorbed by any types of sediments. The highest level of adsorbed Hg ranged from 86% to 98% and was observed at pH ranging from 3 to 5. Unlike Barrow and Cox (1992); Yin et al. (1996) found that Hg
adsorption did not decrease with the addition of Cl but that the impact of Cl on mercury adsorption by soil depended upon the amount of organic matter in the soil. Semu et al. (1987) studied pH incidence on mercury adsorption by tropical soil at pH ranging from 5 to 8, and noted an increase of adsorption at increasing pH. Our results indicate that, when temperature and pH varied, no significant difference in Hg concentration was noted in air or water; Hg was strongly fixed onto the sediment regardless of temperature and pH. Our results demonstrate that the sediment fixes very quickly (within 30 min) most of the mercury added to the water, whether it be CH3 HgCl or HgCl2 (Jahanbakht et al., 1998). Even though mercury contaminates sediments, it may not be detected in the water. Therefore, water is a bad exposure indicator for aquatic pollution. For chronic pollution assessment, indicators such as sediments, fish, aquatic plants are more appropriate.
References Airey, D., 1982. Contribution from coal and industrial materials to mercury in air, rain water and snow. The Science of the Total Environment 28, 19–40. Baker, M.D., Wong, P.T.S., Chau, Y.K., Mayfield, C.I., Innis, W.E., 1981. Methylation of lead, mercury, arsenic and selenium in the acidic aquatic environment. In: Heavy metals in the Environment International Conference, CEP Consultants Ltd., Amsterdam, pp. 645–647. Barrow, N.J., Cox, V.C., 1992. The effects of pH and chloride concentration on mercury sorption: II. By a soil. Journal of Soil Science 43, 305–312. Beijer, K., Jernelov, A., 1979. Methylation of mercury in aquatic environment. In: Nriagu, J.O. (Ed.), Biochemistry of Mercury in the Environment. Elsevier, New York, pp. 207–210. DÕItri, F.M., 1992. Mercury pollution and cycling in aquatic ecosystems. In: Proceeding 5th International Conference Environmental Contamination, CEP Consultants Ltd., Edinburgh, p. 64. Driscol, C.T., Blette, V., Yan, C., Schofield, C.L., Munson, R., Holsapple, J., 1995. The role of dissolved organic carbon in the chemistry and bio-availability of Hg in remote Adirondack. Water Air and Soil Pollution 80, 499– 508. Frenet, M., 1979. Phenomenes de fixation-desorption du mercure sur les argiles dans les eaux a salinite variable. Application a lÕestuaire de la Loire. These de Doctorat dÕetat, Universite de Nantes, Nantes. Hatch, W.R., Ott, W.L., 1968. Determination of sub-microgram quantities of mercury by atomic absorption spectrometry. Analytical Chemistry 40, 2085–2087. Heimburger, R., Livardjani, F., Leroy, M.J.F., 1986. Analyse du mercure par spectrometrie dÕAbsorption Atomique en vapeur froide. Analusis 14, 173–180. Henry, E.A., Dodge-Murphy, L.J., Driscol, C.T., Wang, W., 1996. Evaluation of total mercury loading to Onondaga lake
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405 via tributaries. In: Proceeding IV International Conference on Mercury as Global Pollutant. Hamburg, p. 295. Hutagalung, H.P., 1987. Mercury content in the water and marine organisms in Angke estuary, Jakarta Bay. Bulletin of Environmental Contamination and Toxicology 39, 406– 411. Jahanbakht, S., Livardjani, F., Ruhl, E., Jaeger, A., Lugnier, A.A., 1998. An experimental ecotoxicological model and its application to the behavioral study of inorganic mercury (HgCl2 ) in the environment. Analusis 26, 347– 350. Lacerda, L.D., 1997. Evolution of mercury contamination in Brazil. Water Air and Soil Pollution 97, 247–255. Livardjani, F., Heimburger, R., Leroy, M.J.F., Jaeger, A., Lugnier, A.A., 1995. Dosage du mercure dans lÕair. Mise au point dÕun appareil portatif. Analusis 23, 392–397. Livardjani, F., Heimburger, R., Leroy, M.J.F., Jaeger, A., Lugnier, A.A., 1996. Determination of total mercury in estuary, lake and river sediment. Central European Journal of Public Health 4 (Suppl.), 53. MacNaughton, M.G., James, R.O.J., 1974. Adsorption of aqueous mercury (II) complexes at the oxide/water interface. Journal of Colloid Interface Science 47, 431–440. Marins, R.V., Tonietto, G.B., 1995. An evaluation of sampling and measurements techniques of mercury in the air. In: Proceeding 5th Brazilian geochemical congress, Niteroi, Analytical Geochemistry Section, p. 1. Messaitfa, A., 1997. Transfert dÕeau, des sediments et de polluants associes sur le bassin de lÕIll: cas de mercure. These de lÕUniversite Louis Pasteur de Strasbourg, Strasbourg, France.
1405
Moreira, J.C., Pivetta, F.R., 1997. Human and Environmental contamination by mercury from industrial uses in Brazil. Water Air and Soil Pollution 97, 241–246. Prabhu, N.V., Hamdy, M.K., 1977. Behavior of mercury in biosystems. I. Uptake and concentration in food-chain. Bulletin of Environmental Contamination and Toxicology 18, 409–417. Saouter, E., 1990. Etude experimentale de la bioaccumulation des derives du mercure chez une larve dÕinsecte intrasedimentaire––Exagenia Rigida (Ephemeropteres)-incidence de differents facteurs ecotoxicologiques. These de lÕUniversite de Bordeaux I, Specialite: Toxicologie fondamentale et appliquee, Bordeaux, France. Schindler, D.W., Wagemann, R., Cook, R.B., Ruszczynski, T., Prokopowich, J., 1980. Experimental acidification of lake 223, experimental lakes area: background data and the first three years of acidification. Canadian Journal of Fisheries and Aquatic Sciences 37, 342–354. Schuhmacher, M., Domingo, J.L., Liobrt, J.M., Corbella, A.J., 1993. Evaluation of the effect of temperature, pH, and bioproduction of Hg concentration in sediments, water, mollusks and algae of the delta of the Ebro river. The Science of the Total Environment (Suppl.), 117–125. Semu, E., Singh, B.R., Selmer-Olsen, A.R., 1987. Adsorption of mercury compounds by tropical soils: II. Effect of soil: solution, ratio, ionic strength, pH, and organic matter. Water, Air and Soil Pollution 32, 1–10. Yin, Y., Allen, H.E., Li, Y., Huang, C.P., Sanders, P.F.J., 1996. Heavy metals in the environment: adsorption of mercury (II) by soil: effect of pH, chloride and organic matter. Journal of Environmental Quality 25, 837–844.
An experimental ecotoxicological study and its application to the behavioural study of organic mercury (CH3HgCl) in the environment: influence of temperature and pH S. Jahanbakht
a,b,* ,
F. Livardjani a, A. Jaeger
a
a
b
Laboratoire du Centre Antipoison, Universit e Louis Pasteur de Strasbourg, Strasbourg Cedex 67005, France AD Scientifique Centre d’ etudes, d’analyses et diagnostics en toxicologie de l’environnement, BP 214 R5 67005 Strasbourg, France Received 8 May 2001; received in revised form 8 August 2002; accepted 23 August 2002
Abstract Laboratory experiments were carried out to study the influence of temperature (24, 28 and 30 °C) and pH (1–10) on organic mercury (CH3 HgCl) transfer and accumulation in an experimental ecotoxicological model. We followed the evolution of CH3 HgCl in a basic model (water þ air) by varying temperature and pH. In a second step, we completed the model by adding sediment and fish. We added CH3 HgCl to water at the beginning of each experiment which was repeated at least three times. Results demonstrated that mercury was released from methylmercury into the air regardless of water pH and its concentration in the air increased with increasing pH. By contrast, in presence of sediment, almost all the mercury was fixed onto the sediment and no mercury was traced in air or in water. Interestingly, in the presence of sediment, the life span of fish under methylmercury exposure lasted longer despite their higher mercury body level content at their death. These results indicate that water is a bad exposure indicator for aquatic pollution. In case of chronic pollution, sediments, fish and aquatic plants are more appropriate indicators. Ó 2002 Elsevier Science Ltd. All rights reserved. Keywords: Methylmercury; Air; Water; Sediment; pH; Fish
1. Introduction Many mercury applications in industries entailed several cases of contamination in the past, as in Japan (Minamata) or Iraq, and today they are still a serious issue in countries as Brazil (Lacerda, 1997; Moreira and Pivetta, 1997). Mercury is one of the few metals to have caused human deaths related to food contamination,
* Corresponding author. Address: Laboratoire du Centre Antipoison, Universite Louis Pasteur de Strasbourg, Strasbourg Cedex 67005, France. Tel.: +33-388-24-33-17; fax: +33390-24-40-07. E-mail address: [email protected]. fr (S. Jahanbakht).
and over the past 40 years, it is estimated to have caused the death of 1400 people and affected 20 000 people worldwide (DÕItri, 1992). Mercury can be released from anthropogenic sources into the atmosphere in various chemical forms. Mercury most common forms are Hg0 , Hg2þ , particle Hg and HgO (Marins and Tonietto, 1995). The effects of mercury and its derivatives, especially of methylmercury, on the environment and aquatic media, have been widely discussed in the literature (Lacerda, 1997; Moreira and Pivetta, 1997). Yet, little has been reported on the mechanisms of mercury release into the atmosphere from water or soil. In this paper, we investigated the evolution of mercury in the air after mercury water pollution and we analysed the influence of several biotic and abiotic parameters. Implications of this study are discussed.
0045-6535/02/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 5 - 6 5 3 5 ( 0 2 ) 0 0 5 0 6 - 4
1400
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405
2.1.2. Air compartment The system contained a 100 l volume of room air. After cleaning, we analysed the air of the system so as to check that no mercury traces were left.
2. Material and methods 2.1. Structure of the experimental model The main characteristics of the compartments introduced into the experimental system (glass tank, 200 l capacity) are presented in Fig. 1. To minimize Hg adsorption, we chose materials most inert to mercury, such as Teflon tubes and glass walls. For air and water sampling, we used two pumps with a sampling site on each to create permanent flows, one for air and one for water. We then connected the mercury analyser to the air sampling site and air was automatically sampled every fifteen minutes throughout the duration of the experiment. We cleaned the model (sides of the glass tank) after each experiment with 5% diluted nitric acid and aired it for several days.
2.1.1. Water compartment A 100 l volume of tap water from the Strasbourg network was introduced into the system and was not renewed during the experiments. The water initial physico-chemical characteristics were checked by the Direction Departementale des Affaires Sanitaires et Sociales de la ville de Strasbourg. They were as follows: chloride (Cl ) 70 mg/l, calcium (Ca2þ ) 97.7 mg/l, potassium (Kþ ) 3.8 mg/l, sodium (Naþ ) 49.8 mg/l, NO 2 2 0.04 mg/l, NO 42 mg/l. At time zero, 3 9 mg/l, SO4 approximately 0.1 mg/l of CH3 HgCl (purity: 99%; purchased from Riedel-de-Ha€en) was added to the water.
2.1.3. Sediment compartment For this study, we used two types of sediments, natural sediment and sand without organic matter. For each experiment, we used new sand and sediment samples with exactly the same characteristics, the same weight 1 and from the same source. The sediment was taken from a local gravel pit (Wittisheim, France) and its geochemical and granulometric characteristics were determined by the Institut de Geologie de Strasbourg (ULP, Strasbourg University, France). It was mainly composed of quartz (SiO2 ) (40%), limestone (40%) and white mica (20%) and organic matters such as wood residues, plants, insects and shells. The particle size of the sediment was about 0.05 mm. The total carbon content, determined with the Carbon Sulfur Analyser CS 125 (LECO), was 14%. The natural level of total mercury was 52:7 19 ng Hg/g dry weight. The sand used was taken from a river of fluvial origin and the average size of the particles was 1 mm (2–0.5 mm), of which 95% were square while less than 5% were round. The sand was composed of 80% quartz, 10–15% feldspath and the remaining were ferro-magnesian minerals (pyroxene and amphiboles). No mercury traces were found in this sand. 2.1.4. Fish We selected Scardinius erythrophtalmus (of the cyprimide family) which lives in quiet rivers, ponds and lakes.
Fig. 1. Diagram of the model.
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405
For our study, we used 4–5 month-old farm-raised fish, with an average weight of 13:312 0:041 g and a size of 12:4 0:5 cm. To avoid risks of food contamination, they had not been fed during the 30 day-long experiment and neither had the control group. For consistent results, we used a batch of fish. At time zero of the experiment, we introduced a group of nine fish into the model, and every 10th day a batch of three fish was sacrificed. For each batch, the average background level (Hg natural content) was determined. The aim was to find mercury transfer in the model. Since we were looking for total mercury in fish we did not care for mercury distribution in the organs; so after mixing the fish we analysed the mixture for total mercury traces. We repeated this type of experiments three times:
1401
ple of cold vapour atomic absorption (Hatch and Ott, 1968). Atmospheric mercury and inorganic forms of mercury in water were directly measured with the above described prototype (Livardjani et al., 1995). For organic forms, we carried out mineralisation according to the nature of the sample and optimised its parameters to reduce loss by volatilisation (see below). We used Merck reagents containing a minimum of mercury as recommended for mercury analysis (Heimburger et al., 1986; Jahanbakht et al., 1998). 2.3. Air analysis We directly sampled the air of the system three times every 15 min, then analysed total mercury in each sample and the average was taken (Figs. 2 and 3).
2.2. Mercury analyser 2.4. Sediment mineralisation In this study, we carried out the experiments with a mercury analyser developed in the laboratory and slightly modified; in this case, we used a 30 cm-long and 2.2 cm-diameter-measuring cell with a detection limit of 90 ng Hg/m3 in the air and 15 pg Hg in absolute weight with a 5% uncertainty determination (Jahanbakht et al., 1998). This mercury analyser is based on the princi-
The level of moisture in the sediment to be analysed was measured in each of the samples taken separately and assessed from triplicates. The sediment being a complex matrix, a higher temperature and a longer time were required to obtain complete mineralisation (Jahanbakht et al., 1998). The samples were placed into an
Fig. 2. Hg evolution in the air at different temperatures (pH ¼ 7:2; CH3 HgCl). Data points average of three independent experiments.
Fig. 3. Hg concentration according to pH in the air (T ¼ 24 °C, CH3 HgCl). Data points average of three independent experiments.
1402
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405
acid medium (HNO3 þ H2 SO4 , 6:3 v/v) in a Pyrex flask during 24 h at room temperature, and were then mineralised with a reflux cooling system at boiling temperature (about 110 °C) for 24 h. Oxidation was obtained by adding 5 ml KMnO4 (0.44 M) to the mineralised product. A solution of oxalic acid (0.5 M) was used to reduce permanganate until decoloration (Livardjani et al., 1996; Jahanbakht et al., 1998). Mercury was reduced to Hg0 by stannous chloride (5 ml of a 10% solution in double-distilled water). Mercury vapour (Hg0 ) was purged with an argon current, (Heimburger et al., 1986). The accuracy of this method was checked with the analysis of standard reference materials (e.g. BCR––Bureau Communautaire de Reference de la Commission des Communautes Europeennes) and was in good agreement with certified values (Livardjani et al., 1996). Results are expressed in concentration (ng Hg/g (dry weight)) and total mercury burden (mg Hg for the total amount of sediment). 2.5. Fish mineralisation The mineralisation parameters were optimised on aliquot fractions (500 mg) of fish mixtures (S. erythrophtalmus). The best results were obtained using a sulfonitric mixture (HNO3 þ H2 SO4 , 6:4 v/v). This operation, carried out in a reflux cooling system, lasted 1 h at 60 °C. Then, 3 ml of a KMnO4 solution (0.44 M) were added to oxidize mercury. The excess permanganate was removed by adding about 2 ml of oxalic acid (0.5 M). The method was checked against the standard addition method. 95% of the mercury was recovered (Jahanbakht et al., 1998). Results are expressed in concentration (ng Hg/g (fresh weight)). 2.6. Water mineralisation For mercury recuperation in environmental water, samples were treated using a mixture of nitric and sulfuric acid (3 ml HNO3 , 2 ml H2 SO4 ), at 20 °C for 10 min with a 200 ll sample. For mercury oxidation, 5 ml of a potassium permanganate solution (0.44 M) were added. The excess permanganate was eliminated by adding 5 ml of oxalic acid (0.5 M).
3. Results 3.1. Temperature influence Initial mercury concentration in the water was 0.10 mg/l. pH was constant at 7.2. Hg concentrations in the air and the water of the system were measured every 15 min during 48 h. Mercury concentration variations were
assessed at three different temperatures: 24, 28 and 30 °C. Results are shown in Fig. 2. A strong influence of temperature was observed. The delay of mercury release in the air of the system varied according to temperature. Indeed, when the water temperature was 24 °C, mercury was detected from the 4th h (240 min) whereas at 28 and 30 °C, it was detected from the 8th (480 min) and the 20th h (1200 min) respectively (Fig. 2). The mercury concentration in the air strictly increased with temperature and reached an equilibrium between 20 and 40 h. The value was related to the water temperature (0:155 0:06 mg/m3 at 24 °C, 11:52 0:05 mg/m3 at 28 °C, and 20:5 0:5 mg/m3 at 30 °C (Fig. 2)). 3.2. pH influence To study the pH influence, the temperature was kept constant. The initial water concentration was 0.09 mg/l. Each experiment was repeated at different pH values ranging from 1 to 12, using hydrochloric acid (HCl) and sodium hydroxide (NaOH), and was carried out over a 48-h period. The Hg concentrations in the air of the system found for each pH at the 48th h (after reaching equilibrium) are presented in Fig. 3. Methylmercury was released in the air regardless of the pH. Yet, this release was more intense in a basic medium. It was demonstrated (Driscol et al., 1995) that dissolved organic carbon (DOC) stabilizes methylmercury (DOC-HgCHþ 3 ) in solution. Besides, the presence of aluminium or other ions in the solution (competition principle) favours the transfer of dissolved methylmercury (HgCHþ 3 ) to the particle phase (Henry et al., 1996) according to the following reaction: 3þ DOC-HgCHþ ! DOC-Al3þ þ HgCHþ 3 þ Al 3
Dimethylmercury can be formed from HgCHþ 3 . This formation is favoured in a basic medium (Beijer and Jernelov, 1979; Baker et al., 1981). Due to its volatility, dimethylmerury is released in the air where it changes into Hg0 . 3.3. Sediment influence To assess the influence of sediment on the evolution of mercury in the air of our basic system (water þ air), we adjusted the initial water concentration to 0.050 mg/l, temperature was kept constant at 24 °C and pH at 7.2. Then we added 8.501 kg of natural sediment (dry weight) to the system with an initial mercury concentration of 52:7 19 ng Hg/g of sediment. During a 48 h monitoring, no mercury traces were detected in either compartment within the detection limit (90 ng Hg/m3 ). At the experiment completion, the sediment contained 606:99 83 ng Hg/g of sediment; the
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405
1403
Fig. 4. Mercury amount traced in the sediment according to pH (T ¼ 24 °C). Data points average of three independent experiments.
total amount of sediment had thus adsorbed 5.16 mg of Hg, which means more than 95% of the mercury added to water (5.4 mg). Identical experiments carried out at 28 and 30 °C and various pH (1–10) showed that pH variation had no effect on the methylmercury adsorption by the sediment (Fig. 4). No significant difference was noted with the variation of pH, and the total amount of Hg added to the water was rapidly fixed onto the sediment. 3.4. Sand influence The same experiments were repeated (T ¼ 24 °C, pH ¼ 7:2), with sand instead of sediment. It was noted that nearly 17% of the mercury added to the water was fixed by sand and this amount remained stable during the month-long period while the system was monitored. In the presence of sand, the air was contaminated and mercury concentration varied as in a basic system (water þ air). 3.5. Fish influence The protocol was first simplified and fish were added to CH3 HgCl contaminated water at a 0.095 mg/l concentration. Then, sediment was added under the same conditions (T ¼ 24 1 °C, pH ¼ 7:2). In order to compare the two different aquatic matrices, we used sand instead of sediment. We repeated the following experiment three times. For each experiment, we used nine fish, three of which were to be sacrificed (five aliquots determined for each fish) every 10 days after the beginning of exposure. The control fish were maintained in the same conditions (temperature, pH) throughout the time of the experiment. No fish mortality was recorded in the control. 3.6. System: water þ fish We placed nine fish in the system and all died before the day foreseen for their sacrifice; they died within 9 h
following the addition of mercury to the water and had an average concentration of 0:696 0:103 mg Hg/kg of fish (0.474–0.824). No specific organ of the fish was analysed and the result shows the concentration of the whole fish. They had a restless behaviour before dying; hourly, they had serious behavioural problems for 5–10 min, with a loss of swimming ability, spinning round, head turned downwards and motionless, and their tail making circular movements. After a brutal stop, they seemed to swim normally again but actually went straight into the sides of the aquarium. The frequency of these crises increased until their death. 3.7. System: water þ sediment þ fish To carry out this experiment, temperature was set at 24 °C and pH at 7.2. After adjusting the water to obtain a 0.095 mg/l mercury concentration, we added 8.501 kg of natural sediment to the system and introduced 9 fish. Once the sediment had been added into the system, the fish remained alive about 35 h but had an average of 1:327 0:138 mg Hg/kg fish (1.218–1.677). The sediment contained 1:320 0:029 lg Hg/g sediment when the experiment was completed. 3.8. System: water þ sand þ fish The preceding experiment was repeated but instead of sediment, the same amount of sand was taken. After about 7 h, the fish showed the same symptoms as reported previously. All of them died between 8 and 20 h and had an average mercury concentration of 0:689 0:202 mg Hg/kg fish.
4. Discussion These findings show that the processes of mercury transformation and bioaccumulation in the environment are very complex. When pH and temperature are constant, an increase of the mercury amount in the air
1404
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405
related to time is noted. When pH is constant and temperature varies, the mercury amount in the air varies from 0.1% to 10% with increasing temperature. Schuhmacher et al. (1993) found that mercury concentrations in water and sediment were lower in summer than in winter. Airey (1982) noted that mercury concentrations in the air increased when temperature increased while Saouter (1990) reported a sharp fall of mercury concentration in the water related to temperature increase. Similar results were reported by Hutagalung (1987) and Prabhu and Hamdy (1977). However, no such correlation between temperature and total mercury amount in the water was observed by Frenet (1979). In these studies, the chemical form of mercury was not mentioned. When pH varied and temperature was kept constant, whatever the rate of water acidity, mercury was released into the air and its concentration increased according to pH. In our previous paper (Jahanbakht et al., 1998), we saw that when the source of water pollution was HgCl2 , Hg remained stable in the water in an acid medium (pH < 5) and the air was not contaminated. On the contrary, at pH > 5, mercury was released into the air of the system and reached an equilibrium value which was temperature-related. Thus, air contamination is related to the chemical form of the water pollutant (organic or inorganic). Schindler et al. (1980) reported an increase of mercury concentration in water and a decrease in sediment when pH decreased. Messaitfa (1997) noted a Hg increase in water when pH ranged between 7.4 and 7.8 and a decrease when it ranged between 7.8 and 8.4. In the presence of sediment, almost all the amount of methylmercury chloride added to the water is rapidly fixed onto the sediment and the air is not contaminated. This phenomenon was also observed with inorganic mercury (Jahanbakht et al., 1998). However, when sand is present, approximately 17% of the mercury added to the water is fixed onto the sand and the air is contaminated. Indeed, mercury concentrations are strongly affected by the sediment granulometry and this is observed in our case since the granulometry of sediments is higher in fine particles (sediment) than in sand. The influence of organic matters should not be minimized. MacNaughton and James (1974) found that Hg2þ adsorption on SiO2 increased sharply at pH 2 to 3, and in the absence of Cl . Barrow and Cox (1992) demonstrated that in the absence of Cl , the maximal adsorption was just above pH 4. In both studies, addition of Cl to the medium significantly reduced mercury adsorption. Yin et al. (1996) however showed that at low pH, Hg2þ was adsorbed by any types of sediments. The highest level of adsorbed Hg ranged from 86% to 98% and was observed at pH ranging from 3 to 5. Unlike Barrow and Cox (1992); Yin et al. (1996) found that Hg
adsorption did not decrease with the addition of Cl but that the impact of Cl on mercury adsorption by soil depended upon the amount of organic matter in the soil. Semu et al. (1987) studied pH incidence on mercury adsorption by tropical soil at pH ranging from 5 to 8, and noted an increase of adsorption at increasing pH. Our results indicate that, when temperature and pH varied, no significant difference in Hg concentration was noted in air or water; Hg was strongly fixed onto the sediment regardless of temperature and pH. Our results demonstrate that the sediment fixes very quickly (within 30 min) most of the mercury added to the water, whether it be CH3 HgCl or HgCl2 (Jahanbakht et al., 1998). Even though mercury contaminates sediments, it may not be detected in the water. Therefore, water is a bad exposure indicator for aquatic pollution. For chronic pollution assessment, indicators such as sediments, fish, aquatic plants are more appropriate.
References Airey, D., 1982. Contribution from coal and industrial materials to mercury in air, rain water and snow. The Science of the Total Environment 28, 19–40. Baker, M.D., Wong, P.T.S., Chau, Y.K., Mayfield, C.I., Innis, W.E., 1981. Methylation of lead, mercury, arsenic and selenium in the acidic aquatic environment. In: Heavy metals in the Environment International Conference, CEP Consultants Ltd., Amsterdam, pp. 645–647. Barrow, N.J., Cox, V.C., 1992. The effects of pH and chloride concentration on mercury sorption: II. By a soil. Journal of Soil Science 43, 305–312. Beijer, K., Jernelov, A., 1979. Methylation of mercury in aquatic environment. In: Nriagu, J.O. (Ed.), Biochemistry of Mercury in the Environment. Elsevier, New York, pp. 207–210. DÕItri, F.M., 1992. Mercury pollution and cycling in aquatic ecosystems. In: Proceeding 5th International Conference Environmental Contamination, CEP Consultants Ltd., Edinburgh, p. 64. Driscol, C.T., Blette, V., Yan, C., Schofield, C.L., Munson, R., Holsapple, J., 1995. The role of dissolved organic carbon in the chemistry and bio-availability of Hg in remote Adirondack. Water Air and Soil Pollution 80, 499– 508. Frenet, M., 1979. Phenomenes de fixation-desorption du mercure sur les argiles dans les eaux a salinite variable. Application a lÕestuaire de la Loire. These de Doctorat dÕetat, Universite de Nantes, Nantes. Hatch, W.R., Ott, W.L., 1968. Determination of sub-microgram quantities of mercury by atomic absorption spectrometry. Analytical Chemistry 40, 2085–2087. Heimburger, R., Livardjani, F., Leroy, M.J.F., 1986. Analyse du mercure par spectrometrie dÕAbsorption Atomique en vapeur froide. Analusis 14, 173–180. Henry, E.A., Dodge-Murphy, L.J., Driscol, C.T., Wang, W., 1996. Evaluation of total mercury loading to Onondaga lake
S. Jahanbakht et al. / Chemosphere 49 (2002) 1399–1405 via tributaries. In: Proceeding IV International Conference on Mercury as Global Pollutant. Hamburg, p. 295. Hutagalung, H.P., 1987. Mercury content in the water and marine organisms in Angke estuary, Jakarta Bay. Bulletin of Environmental Contamination and Toxicology 39, 406– 411. Jahanbakht, S., Livardjani, F., Ruhl, E., Jaeger, A., Lugnier, A.A., 1998. An experimental ecotoxicological model and its application to the behavioral study of inorganic mercury (HgCl2 ) in the environment. Analusis 26, 347– 350. Lacerda, L.D., 1997. Evolution of mercury contamination in Brazil. Water Air and Soil Pollution 97, 247–255. Livardjani, F., Heimburger, R., Leroy, M.J.F., Jaeger, A., Lugnier, A.A., 1995. Dosage du mercure dans lÕair. Mise au point dÕun appareil portatif. Analusis 23, 392–397. Livardjani, F., Heimburger, R., Leroy, M.J.F., Jaeger, A., Lugnier, A.A., 1996. Determination of total mercury in estuary, lake and river sediment. Central European Journal of Public Health 4 (Suppl.), 53. MacNaughton, M.G., James, R.O.J., 1974. Adsorption of aqueous mercury (II) complexes at the oxide/water interface. Journal of Colloid Interface Science 47, 431–440. Marins, R.V., Tonietto, G.B., 1995. An evaluation of sampling and measurements techniques of mercury in the air. In: Proceeding 5th Brazilian geochemical congress, Niteroi, Analytical Geochemistry Section, p. 1. Messaitfa, A., 1997. Transfert dÕeau, des sediments et de polluants associes sur le bassin de lÕIll: cas de mercure. These de lÕUniversite Louis Pasteur de Strasbourg, Strasbourg, France.
1405
Moreira, J.C., Pivetta, F.R., 1997. Human and Environmental contamination by mercury from industrial uses in Brazil. Water Air and Soil Pollution 97, 241–246. Prabhu, N.V., Hamdy, M.K., 1977. Behavior of mercury in biosystems. I. Uptake and concentration in food-chain. Bulletin of Environmental Contamination and Toxicology 18, 409–417. Saouter, E., 1990. Etude experimentale de la bioaccumulation des derives du mercure chez une larve dÕinsecte intrasedimentaire––Exagenia Rigida (Ephemeropteres)-incidence de differents facteurs ecotoxicologiques. These de lÕUniversite de Bordeaux I, Specialite: Toxicologie fondamentale et appliquee, Bordeaux, France. Schindler, D.W., Wagemann, R., Cook, R.B., Ruszczynski, T., Prokopowich, J., 1980. Experimental acidification of lake 223, experimental lakes area: background data and the first three years of acidification. Canadian Journal of Fisheries and Aquatic Sciences 37, 342–354. Schuhmacher, M., Domingo, J.L., Liobrt, J.M., Corbella, A.J., 1993. Evaluation of the effect of temperature, pH, and bioproduction of Hg concentration in sediments, water, mollusks and algae of the delta of the Ebro river. The Science of the Total Environment (Suppl.), 117–125. Semu, E., Singh, B.R., Selmer-Olsen, A.R., 1987. Adsorption of mercury compounds by tropical soils: II. Effect of soil: solution, ratio, ionic strength, pH, and organic matter. Water, Air and Soil Pollution 32, 1–10. Yin, Y., Allen, H.E., Li, Y., Huang, C.P., Sanders, P.F.J., 1996. Heavy metals in the environment: adsorption of mercury (II) by soil: effect of pH, chloride and organic matter. Journal of Environmental Quality 25, 837–844.