Anal. Chem. 1996, 68, 1262-1266
Determination of Monomethylcadmium in the Environment by Differential Pulse Anodic Stripping Voltammetry Richard Pongratz and Klaus G. Heumann*
Institute of Inorganic Chemistry of the University of Regensburg, Universita¨ tsstrasse 31, D-93040 Regensburg, Germany
A differential pulse anodic stripping voltammetric (DPASV) method was used to differentiate between the cadmium species Cd2+ and MeCd+ (Me ) methyl) in aquatic systems. These two species show peaks in the DPASV voltammogram which differ by 112 mV. In model experiments, it was demonstrated that monomethylcadmium is not stable at pH 2, but under higher pH conditions, normally found in fresh and ocean water samples, the identity of MeCd+ was verified by different investigations, including cyclic voltammetry, selective extraction of a complex of diethyldithiocarbamate with MeCd+ into nhexane, and photochemical dissociation of MeCd+ by UV irradiation. It was also shown that humic acids do not influence the voltammetric determination of monomethylcadmium. For the first time, it was possible to analyze MeCd+ in environmental samples. During different expeditions with the German research vessel Polarstern, monomethylcadmium could be determined above the detection limit of 470 pg L-1 in nearly all surface water samples of the South Atlantic with spot concentrations of up to about 700 pg L-1, whereas in the North Atlantic only 15-30% of the total samples showed MeCd+ concentrations above this limit. The existence of MeCd+ in the remote area of the South Atlantic, as well as positive correlations with the local bioactivity in the ocean, indicates biomethylation as the most probable formation process for this methylated cadmium species. This assumption is supported by the simultaneous occurrence of other methylated heavy metal compounds, such as Me3Pb+. Up to 48% of the total cadmium was found to be monomethylcadmium in some Arctic meltwater ponds. Elemental speciation is an important key to a better understanding of the global geochemical cycle of the elements.1 Different elemental species can occur due to different oxidation states or due to the coordination of an element to various compounds, e.g., to organic substituents. Organometallic compounds with covalent metal-carbon bonds show significantly different behaviors compared with hydrated metal cations. The volatility and toxicity, for example, of inorganic mercury and lead ions significantly differ from the corresponding properties of dimethylmercury and tetraalkyllead, respectively.2 The process which transforms inorganic mercury into dimethylmercury in the (1) Frimmel, F. H. Fresenius J. Anal. Chem. 1994, 350, 7-13. (2) Craig, P. J. Organometallic Compounds in the Environment; Longman: Essex, UK, 1986.
1262 Analytical Chemistry, Vol. 68, No. 7, April 1, 1996
environment is biomethylation.3 The biogenic production of methylated mercury and lead compounds is well known due to model experiments4,5 and due to the determination of these compounds in unpolluted areas.6,7 On the other hand, only very little is known about biomethylation of cadmium. Although model experiments with inorganic cadmium and vitamin-B12 as methylation agent have been successfully carried out,8,9 no methylcadmium compound could be detected until now in the environment. A differential pulse anodic stripping voltammetric (DPASV) method has been developed in this work, which enables the detection of methylcadmium (MeCd+) in aquatic systems in the presence of “free” Cd2+ ions. This DPASV method was applied to ocean water and fresh water samples from remote areas. MeCd+ was determined for the first time in environmental samples. The analysis of samples in remote areas guaranteed that biomethylation was the most probable process responsible for the natural existence of monomethylcadmium. EXPERIMENTAL SECTION Instrumentation. A polarographic analyzer with digital plotter (EG&G, Princeton Applied Research), Model 384 B, was used for DPASV determinations. A rotating mercury film electrode, type Rotel-2, plated in situ on a glassy carbon substrate, was applied as working electrode and Ag/AgCl as reference electrode. Chemicals. Monomethylcadmium was prepared from dimethylcadmium by reaction with cadmium chloride. Dimethylcadmium was synthesized by a Grignard reaction in absolute ether with subsequent distillation of Me2Cd.10,11 Purification of dimethylcadmium was carried out in a second distillation step. Me2Cd was then converted into MeCd+ by adding a substoichiometric amount of cadmium chloride in ether under nitrogen inert gas conditions. Because monomethylcadmium is insoluble in ether, the MeCd+ precipitate could be isolated and dried under nitrogen atmosphere. This substance was used to produce standard solutions of MeCd+ which were calibrated by flame atomic (3) Wood, J. M.; Kennedy, F. S.; Rosen, C. G. Nature (London) 1968, 220, 173-174. (4) Campeau, G. C.; Bartha, R. Appl. Environ. Microbiol. 1985, 50, 498-502. (5) Walton, A. P.; Ebdon, L.; Millward, G. E. Appl. Organomet. Chem. 1988, 2, 87-90. (6) Harrison, R. M.; Allen, A. G. Appl. Organomet. Chem. 1989, 2, 49-58. (7) Mason, R. P.; Fitzgerald, W. F. Nature (London) 1990, 347, 457-459. (8) Huey, C. W.; Brinckman, F. E.; Iverson, W. P.; Grim, S. O. Proc. Int. Conf. Heavy Met. Environ. 1975, C214-C216. (9) Robinson, J. W.; Kiesel, E. L. J. Environ. Sci. Health 1981, A16, 341-352. (10) Nesmeyanov, A. N.; Kocheshkov, K. A. Methods of Elemento-Organic Chemistry; North-Holland: Amsterdam, the Netherlands, 1967; Vol. 3, Chapter 1. (11) Nu ¨ tzel, K. Organo-Cadmium-Verbindungen. In Methoden der organischen Chemie; Mu ¨ ller, E., Ed.; Thieme Verlag: Stuttgart, 1973; Band XIII/2a. 0003-2700/96/0368-1262$12.00/0
© 1996 American Chemical Society
could then be calculated with respect to the evaluation procedure described by Mart et al.13 Between analyses of different samples, the surface of the glassy carbon electrode was purified by polishing with aluminum oxide powder.13 The precision for MeCd+ determinations was 5% relative standard deviation for triplicate analyses by the DPASV method described. The detection limit of 470 pg L-1 was calculated as three times the background noise of the signal at the position of the MeCd+ peak.
Figure 1. Calibration curve for MeCd+ (solutions adjusted to pH 8 by an NH4Cl/NH3 buffer).
absorption spectrometry with Cd2+ standard solutions (Merck, Darmstadt). All standard solutions have been stored at 4 °C under exclusion of light and were prepared with ultrapure water produced in a twofold distillation apparatus of quartz. Hydrochloric acid was cleaned by subboiling distillation. The artificial seawater was produced with reference to the descriptions of Grasshoff et al.,12 including sodium, potassium, magnesium, calcium, and strontium as chlorides as well as potassium bromide, sodium fluoride, sodium sulfate, boric acid, and sodium hydrogen carbonate in concentrations adapted to the average ocean water values. The corresponding amounts of these salts have been dissolved in ultrapure water. Analytical Procedure. To 50 mL of each sample was added 250 µg of Hg(NO3)2. For samples in which the cadmium compound was dissolved in pure water instead of real or artificial ocean water, 50 µL of a 1000 ppm Mg(NO3)2 solution was added to increase the conductivity of the solution. Each sample was degassed by a pure nitrogen (99.999%) gas flow for 10 min. Afterward, the electrolysis took place for about 30 min at -1.0 V under stirring with the rotating electrode at 2000 rpm. After an interruption of 30 s, the voltage was scanned from -1.0 to -0.5 V in the differential pulse mode, running at a rate of 4 mV s-1 with a pulse amplitude of 50 mV and a pulse frequency of 2 s -1. The approximate concentration in the solutions, which was the basis for the amount of MeCd+ used in the following standard addition procedure, was estimated by a calibration curve (Figure 1) obtained for standard solutions prepared from the synthesized MeCd+ compound. As can be seen from Figure 1, the calibration curve for MeCd+ was found to be linear in the concentration range measured from 0.477 ng L-1 (detection limit) to 20 ng L-1. For a more exact quantitation, standard addition was then carried out with the original solution for determination using the described voltammetric procedure with shorter deposition times than before. Usually, two standard additions were applied per analysis. During the first standard addition step, the deposition time was halved, and the concentration was approximately duplicated by the MeCd+ addition. Samples of 5-10 µL of standard solutions with appropriate contents to double the concentration of MeCd+ were used. For the second standard addition step, about three times the concentration and one-third the deposition time were applied. This procedure yielded similar intensities in the peak signals of all solutions within one standard addition analysis. Concentrations (12) Grasshoff, K.; Ehrhardt, M.; Kremling, K. Methods of Seawater Analysis; Verlag Chemie: Weinheim, 1983.
RESULTS AND DISCUSSION Differentiation of Cd2+ and MeCd+ by DPASV. Model solutions containing only free cadmium ions or monomethylcadmium and a solution containing both Cd2+ and MeCd+ were analyzed by the DPASV method at pH values of 2 and 8 (pH control by pH electrode). The corresponding results are represented in Figures 2 and 3. At pH 2 (acidification with HCl), identical voltammograms were obtained, which means that only one cadmium species is present in the solution. The peaks of all samples fit the position of free cadmium ions, due to the instability of monomethylcadmium in acidic solutions. This result at pH 2 agrees well with descriptions by Weber and Witman of the instability of this compound in acidic solutions.14 On the other hand, at pH 8, two peaks at distinctly different potentials appeared that can be assigned to the cadmium species Cd2+ and MeCd+, respectively. Figure 3 shows the results of these species in a pure aqueous solution which was adjusted to pH 8 by a NH4Cl/NH3 buffer. The same compounds dissolved in artificial seawater resulted in voltammograms comparable to those of pure water samples at pH 8, with the same difference of 112 mV between free cadmium ions and monomethylcadmium. Depending on the actual electrolyte composition of the solution, the absolute positions of the two peaks of different cadmium species can slightly vary, but the potential difference between the two species remains nearly constant at 112 mV (compare Figures 2 and 3). The good reproducibility of results obtained from model solutions are transferable to real seawater samples, as demonstrated by the voltammogram of a sample from the Atlantic Ocean represented in Figure 4. The positions of the peaks for Cd2+ and MeCd+ ions were at exactly the same potentials as in the model solution (Figure 2c), and the difference between the two peaks of cadmium species was again 112 mV. Next to the MeCd+ peak, Pb2+ could simultaneously be analyzed in ocean water samples by DPASV at a potential of -0.630 V (Figure 4). The reason for the occurrence of two cadmium peaks during the voltammetric stripping process cannot exactly be explained at the moment. However, a similar result was found by others for solutions containing both free cadmium ions and cadmium complexes of humic substances (HUS). Helmers interpreted this phenomenon as resulting from HUS-associated and nonassociated Cd0, which arise during the voltammetric enrichment step,15 at the electrode/solution interface. It is assumed that during the subsequent stripping step, the nonassociated cadmium is oxidized first and then, at another potential, the HUS-associated cadmium. (13) Mart, L.; Nu ¨ rnberg, H. W.; Valenta, P. Fresenius Z. Anal. Chem. 1980, 300, 350-362. (14) Weber, J. H.; Witman, M. W. Unstable Organometallic Intermediates in a Protic Medium. In Organometalls and OrganometalloidssOccurence and Fate in the Environment; Brinckman, F. E., Bellama, J. M., Eds.; American Chemical Society: Washington, DC, 1978. (15) Helmers, E. Fresenius J. Anal. Chem. 1994, 350, 62-67.
Analytical Chemistry, Vol. 68, No. 7, April 1, 1996
1263
Figure 2. DPASV voltammograms of samples originally produced from (a) a Cd2+ solution, (b) a MeCd+ solution, and (c) a solution containing both cadmium species, adjusted to pH 2 by HCl (concentrations for all compounds are about 4 ng of Cd per liter).
Figure 3. DPASV voltammograms of samples adjusted to pH 8 by a NH4Cl/NH3 buffer containing (a) Cd2+ (1.6 ng L-1), (b) MeCd+ (0.5 ng L-1), and (c) both cadmium species (1.6 and 0.5 ng L-1, respectively).
Figure 4. DPASV voltammogram of a seawater sample from the Atlantic Ocean with a determined MeCd+ concentration of 492 pg L-1.
With reference to this model, the electrolytic reduction of MeCd+ can lead to MeCd radicals at the surface of the mercury film electrode during the enrichment step. MeCd is then oxidized after the oxidation of Cd0 has taken place. The possibility that radicals of organometallic compounds can be produced during electrolytic reductions was also postulated for trimethyllead and trialkyltin ions.16,17 1264 Analytical Chemistry, Vol. 68, No. 7, April 1, 1996
Verification of the Identity of MeCd+. No monomethylcadmium species has been identified until now in the environment. It was therefore necessary to verify the identity of the MeCd+ peak in the voltammograms with model solutions and real samples. The following investigations have been carried out. Cyclic Voltammogram. Figure 5 shows cyclic voltammograms of Cd2+ and MeCd+, which are identical for solutions of these ions in pure water adjusted to pH 8 by a NH4Cl/NH3 buffer and in artificial seawater. The concentrations of both compounds in these model experiments were set higher by a factor of about 1000 compared with the experiments represented by Figures 2 and 3 to give relevant peak intensities. The different positions of the peaks for Cd2+ and MeCd+ at potentials of -0.832 and -0.746 V, comparable to the voltammograms in Figures 3 and 4, can clearly be seen. The cyclic voltammogram of MeCd+ also significantly differs in the range between -0.1 and 0 V from that of Cd2+. Both results show two different cadmium species with different electrochemical behavior which can be transferred to real seawater samples. Selective Extraction. MeCd+ forms 1:1 complexes with sodium diethyldithiocarbamate (NaDDTC) that are extractable into nhexane. This is not possible for the corresponding Cd2+ complex. (16) Bond, A. M.; Bradbury, J. R.; Howell, G. N.; Hudson, H. A.; Hanna, P. J.; Shother, S. J. Electroanal. Chem. 1983, 154, 217-228. (17) Doretti, L.; Tagliavini, G. J. Organomet. Chem. 1968, 12, 203-208.
Figure 6. North-south profile of MeCd+ in seawater samples at a depth of 10 m in the Atlantic Ocean from 51° N (British Channel) to 58° S (south of Cape Town at the pack-ice border of Antarctica).
After such an extraction of solutions containing both species, Cd2+ and MeCd+, the signal for monomethylcadmium disappeared in the voltammogram of the aqueous solution. On the other hand, the Cd2+ peak remained under these conditions, which is another argument that the peak at about -0.75 V can be attributed to MeCd+. Influence of Humic Acids. It is well known that humic substances are dissolved in natural water systems in more or less high concentrations. It must therefore also be proved that the signal at about -0.75 V, with a shift of 112 mV to the Cd2+ signal, is not due to complexes of cadmium with HUS. Using commercially available humic acids (Roth, Karlsruhe, Germany; article no. 6298) in aqueous solutions at concentrations up to 3 mg of HUS per liter, a difference between Cd2+ and CdHUS complexes of only 20 mV was detected (shift from -0.87 to -0.85 V). A similar shift of about 40 mV was also found by others using HUS concentrations up to 30 mg L-1.15,18 From these results, it follows that CdHUS complexes can probably not influence the DPASV voltammetric signal of MeCd+. Photochemical Dissociation. Monomethylcadmium is photochemically not very stable and can easily be dissociated by UV irradiation.19 Solutions containing both species, Cd2+ and MeCd+, were irradiated by UV (high-pressure mercury lamp, 750 W) for
about 30 min. After irradiation, the signal in the voltammogram at -0.75 V disappeared, and an increase of the Cd2+ peak could be detected, which demonstrates that monomethylcadmium was transformed into free cadmium ions. First Determinations of MeCd+ in the Environment. Although there had been some suggestions and indications in the literature that organocadmium compounds exist in the environment,15,20,21 methylated cadmium species could not be detected, until now, in environmental samples. Because of the extremely high reactivity of dimethylcadmium with water,10,11 only the presence of monomethylcadmium could be expected in the environment. Because the oceans are known to be one of the most productive environmental systems for biomethylation, seawater samples from the Atlantic Ocean have been analyzed. Sampling was carried out in October and November 1993 during expeditions ANT XI/1,2 as well as in July and August 1994 during expedition ARK X/1, of the German research vessel Polarstern. The samples were obtained from a snorkel system which continuously pumped seawater under clean conditions from the front of the ship’s bow at a depth of 10 m into a laboratory, where the samples were filled into precleaned bottles. During the first expedition, a north-south profile in seawater samples was determined from 51° N (British Channel) to 58° S (south of Cape Town at the pack-ice border). During ARK X/1, corresponding west-east profiles were measured in the Arctic Ocean at latitudes of 75° N and 79° N, respectively. The results of the north-south profile are represented in Figure 6, obtained by DPASV voltammograms as typically shown in Figure 4. During the expedition to the Arctic Ocean, only 30% of all samples could be analyzed for MeCd+ above the detection limit of 470 pg L-1. Also during the north-south cruise of RV Polarstern, only a few seawater samples (about 15% of all samples) could be analyzed for MeCd+ from 51° N to the equator. In contrast, most of the samples from the South Atlantic showed monomethylcadmium concentrations significantly higher than the detection limit. Peak concentrations up to 721 pg L-1 could be determined (Figure 6). This is the first time that monomethylcadmium was detected in environmental samples, and it is also the first set of results on the global distribution of this methylated cadmium species in the oceans.
(18) Labuda, J.; Saur, D.; Neeb, R. Fresenius J. Anal. Chem. 1994, 348, 312316. (19) Witman, M. W.; Weber, J. H. Inorg. Chem. 1976, 15, 2375-2378.
(20) Linnek, P. M.; Ishra, I. V. Microchem. J. 1994, 50, 184-190. (21) Capodaglio, G.; Scarponi, G.; Toscano, G.; Cescon, P. Anal. Chim. (Rome) 1991, 81, 279-296.
Figure 5. Cyclic voltammograms of (a) Cd2+ (1.3 µg L-1) and (b) MeCd+ (2.1 µg L-1) in artificial seawater.
Analytical Chemistry, Vol. 68, No. 7, April 1, 1996
1265
Table 1. MeCd+ in Meltwater Ponds on Arctic Ice Floes
Figure 7. Depth profile in the South Atlantic at 46°30′ S, 15°20′ W for MeCd+ and Me3Pb+.
Biogenic Formation of MeCd+. The described DPASV method and the verification of the identity of MeCd+ in seawater demonstrated in this work allow more extensive investigations on this substance in the future, e.g., on the biogenic formation of this species and its influence on the global geochemical cycle of the heavy metal cadmium. Although more detailed studies must be carried out, our north-south and west-east profiles in the Atlantic Ocean indicated positive correlations with the bioactivity of the ocean at the point of sampling. For example, during the expedition ANT XI/1, relatively high concentrations in the range of 150-400 ng L-1 of adenosine triphosphate (ATP) as an indicator for bioactivity were usually found in the South Atlantic,22 where in most of the samples significant contents of MeCd+ could also be determined (Figure 6). Contrary to that in the South Atlantic, the ATP concentration in the North Atlantic was normally in a lower range of 50-150 ng L-1, which correlated with the lower average content for MeCd+ in this region. It is also significant in this connection that an ATP spot concentration of about 558 ng L-1 was found in the same sample at 14°02′ S, 06°40′ E exactly where the highest MeCd+ content of 721 pg L-1 was determined for the north-south profile (Figure 6). Another indication for the biogenic origin of monomethylcadmium is the fact that an essential increase in the MeCd+ concentration was found at the pack-ice border (at about 58° S, see Figure 6), where the growth of algae in the Antarctic springtime had just started. The relatively high concentrations of MeCd+ in this remote area of the South Atlantic clearly show the nonanthropogenic origin of this compound. It was also possible to analyze a depth profile in the South
geographic position
MeCd+ (pg L-1)
MeCd+/total Cd (%)
75°34′ N, 12°08′ W 74°56′ N, 12°24′ W 75°01′ N, 12°55′ W 75°03′ N, 12°27′ W 74°29′ N, 11°38′ W 78°59′ N, 04°40′ W 79°01′ N, 06°02′ W 78°59′ N, 04°46′ W
676 706 824 676 1265 1029 527 502
26 19 10 9 48 30 15 15
Atlantic from 10 to 200 m, where in the same samples the Me3Pb+ species was also determined by a DPASV method after coprecipitation of free Pb2+ ions with BaSO423 (Figure 7). The biogenic formation of trimethyllead has already been verified in former model experiments5 and must therefore be assumed to be the major production process for this organolead species in remote areas. The positive correlation between both species suggests that similar biomethylation processes must be responsible for the production of these methylated compounds. During the expedition ARK X/1, different meltwater ponds on Arctic ice floes, which showed distinct bioactivity by their green color, were analyzed for their MeCd+ content and also for their total cadmium concentration (by DPASV after UV irradiation). The results are summarized in Table 1. On the one hand, relatively high monomethylcadmium concentrations up to nearly 1300 pg L-1 were determined, which essentially exceed the corresponding contents measured in ocean water. On the other hand, the methylated cadmium species was found to account for up to 48% of the total cadmium in these Arctic water ponds. This result demonstrates that MeCd+ can play an important role in the environment and must therefore be taken into consideration in many natural processes of cadmium, especially in the global biogeochemical cycle of this trace element. ACKNOWLEDGMENT We thank the “Deutsche Forschungsgemeinschaft” for financial support within the “Schwerpunktprogramm Antarktisforschung” and the crew of RV Polarstern for assistance during the expeditions. Received for review August 8, 1995. Accepted January 2, 1996.X AC9507911
(22) Schrems, O., personal communication. (23) Mikac, N.; Brancia, M. Anal. Chim. Acta 1988, 212, 349-353.
1266 Analytical Chemistry, Vol. 68, No. 7, April 1, 1996
X
Abstract published in Advance ACS Abstracts, February 1, 1996.
