Anal. Chem. 1997, 69, 1782-1784
Mechanism for the Electrochemical Stripping Reduction of the Nickel and Cobalt Dimethylglyoxime Complexes Feng Ma, Daniel Jagner,* and Lars Renman
Department of Analytical and Marine Chemistry, University of Go¨ teborg, S-412 96 Go¨ teborg, Sweden
Reductive coulometric stripping potentiometry, a technique not hitherto described, has been used to establish that the reduction of Ni(II) and Co(II) dimethylglyoximates, adsorbed on a mercury film electrode, is a 10electron process. Exhaustive adsorption of Ni(II) or Co(II) complexes, in the 0-4 µg L-1 concentration range, was achieved by vibrationally promoted electrolysis for 3 min of ∼25 µL volume samples, hanging under the working electrode in a nitrogen atmosphere. The adsorbed complexes were reduced by means of a constant current of 50 µA. The technique was successfully used for the calibration-free determination of Ni(II) in certified seawater and river water reference samples. Adsorptive stripping voltammetry is frequently used for the determination of trace elements,1 the determination of nickel(II) and cobalt(II) as their dimethylglyoxime (DMG) complexes, no doubt, being the most common application. Since the method was first characterized as an adsorptive stripping process2 it has been the subject of many studies.3-7 One reason for the large interest in the DMG chelates is their anomalous behavior manifested by an unexpectedly high sensitivity. For example, assuming that reduction of Ni(DMG)2 or Co(DMG)2 saturatedly adsorbed on a mercury drop electrode, is a 2-electron process, whereby the central ion is reduced to its atomic state, each molecule would be allowed to occupy only 0.15 nm2 of the electrode surface.3 If the planar and highly symmetrical complexes were adsorbed as planar monolayers, a most reasonable assumption, each complex would, however, need a surface area of 0.8-0.9 nm2 according to crystal structure data.8 Furthermore, when operating below electrode surface saturation, i.e., under normal analytical conditions, the sensitivity implies that the diffusion coefficient for the dimethylglyoxime complexes is several times higher than that estimated by the Cottrell equation.3 In a recent communication Vukomanovic et al.9 proposed, from results obtained using a combination of voltammetry and atomic (1) Wang, J. Stripping Analysis: Principles, Instrumentation and Applications; VCH: New York, 1985. (2) Pihlar, B.; Valenta, P.; Nu ¨ rnberg, H. W. Fresenius Z. Anal. Chem. 1981, 307, 337-346. (3) Pihlar, B.; Valenta, P.; Nu ¨ rnberg, H. W. J. Electroanal. Chem. 1986, 214, 157-177. (4) Jin, W.; Liu, K. J. Electroanal. Chem. 1987, 216, 181-201. (5) Flora, C. J.; Nieboer, E. Anal. Chem. 1980, 52, 1013-1020. (6) Weinzierl, J,; Umland, F. Fresenius Z. Anal. Chem. 1982, 312, 608-610. (7) Mairanovskii, S. G.; Prokhorova, G. V.; Osipova, E. A. J. Electroanal. Chem. 1989, 266, 205-214. (8) Godycki, L. E.; Rundle, R. E. Acta Crystallogr. 1953, 6, 487-491. (9) Vukomanovic, D. V.; Page, J. A.; vanLoon, G. W. Anal. Chem. 1996, 68, 829-833.
1782 Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
spectroscopy, that the explanation for the high sensitivity is that the reduction of the M(DMG)2 complexes is a 16- or, possibly, an 18-electron process. Here we show, using reductive coulometric stripping potentiometry, that it is, in fact, a 10-electron process. We also show how this information can be used for the calibration-free determination of nickel(II) in natural waters. THEORY Reductive Coulometric Stripping Potentiometry. Oxidative coulometric stripping potentiometry is a recently described technique for the calibration-free determination of some trace elements. The theory for oxidative coulometric stripping potentiometry, applied to the determination of mercury- or gold-soluble elements, has been described elsewhere.10 Reductive coulometric stripping potentiometry, using adsorptive accumulation of metal complexes, MmLp, is based on the quantitative adsorption of the complexes onto a mercury drop or film electrode from, typically 10-30 µL, samples, i.e.,
MmLp(sample) f MmLp(adsorbed)
(1)
After adsorption the metal complex is reduced by means of an applied constant current, istrip (A), schematically represented by
MmLp(adsorbed) + ne- f MmLpn-
(2)
during which process the metal ion(s), the ligand(s), or both are reduced. From known values for the sample volume, v (L), the number of electrons involved in the reduction, n, and the time needed for quantitative reduction of the adsorbed complex, tstrip (s), the sample concentration of complex, C (M), can be calculated from Faraday’s law as
C ) tstripistrip/nFv
(3)
provided, however, that there are no reduction processes competing with the reduction of the complex. Reducible species present in the sample will diffuse to the electrode during reduction in quiescent conditions, thus consuming a fraction of the applied current.10 Denoting this “chemical current”, ichem (A), in which the electrons needed for the formation of the double layer are also included, eq 3 has to be modified to
C ) tstrip(istrip - ichem)/nFv
(4)
(10) Jagner, D.; Wang, Y. Electroanalysis 1995, 7, 614-618. S0003-2700(96)01023-2 CCC: $14.00
© 1997 American Chemical Society
The chemical current, ichem, which is normally at least 1 order of magnitude smaller than istrip, can be determined by repeating the adsorption/stripping process under identical conditions, with the exception that different reductive currents are applied during stripping. By plotting 1/tstrip vs istrip, a straight line is obtained, the intercept on the current axis being equal to ichem. Using this value, the concentration of the complex in the sample can be determined from eq 4 and, obviously, vice versa, if the sample concentration is known, the number of electrons, n, can be determined. EXPERIMENTAL SECTION Instrumentation. A Radiometer Analytical S.A. (Lyon, France) potentiometric stripping analyzer (Tracelab, PSU22) and accompanying personal computer software TAP2 was used for all measurements. By means of this software, fully automated measurement procedures, including control of all experimental parameters, result evaluation, and changing of samples using an automatic sample changer (SAC80, Radiometer Analytical S.A.) could be implemented. Electrodes. A combined 3 mm diameter glassy carbon disk working electrode and Ag and Ag/AgCl counter and reference electrodes (gCC 540, Radiometer Analytical S.A.) built into one cylindrical unit as described elsewhere11 was used for all measurements. When lifted out of a sample, a droplet of reproducible volume (24.8 µL, rsd 1.7% (n ) 20), weight calibration) remained under the electrode, providing electrical contact between the three electrodes through a ceramic plug surrounding the isolated glassy carbon electrode and thus allowing the working electrode potential to be controlled during changing of solutions. The electrode was filled with 3 M KCl, and all potentials given below are vs Ag/ AgCl (3 M Cl-). A dc engine with an unbalanced load was mounted on top of the changer electrode holder, which, when activated, caused the electrode to vibrate at a frequency of ∼100 Hz, thus decreasing the diffusion layer thickness during potentiostatic adsorption of the complexes. Electrochemical Measurements. Prior to use, a film of mercury was plated onto the glassy carbon electrode by means of electrolysis for 4 min at a potential of -1.00 V in a stirred solution containing 1000 mg L-1 mercury(II) and 0.10 M hydrochloric acid. The same mercury film was used for 10-20 measurements. After mercury plating, the electrode was rinsed in distilled water and immersed in samples (20 mL) containing 0.10 mM dimethylglyoxime at pH 9.2 (0.10 M ammonia buffer) and either known concentrations of nickel(II) and cobalt(II) or reference samples certified for the nickel(II) concentration. Stirring was applied for 10 s, and after a quiescent period of 10 s, the electrode, with a sample drop hanging under it, was transferred by the sample changer to an empty beaker which was flushed continuously with nitrogen in order to remove dissolved dioxygen from the sample drop. Vibrational electrolysis was then performed for 3 min followed by a quiescent period of 20 s prior to stripping with a constant reductive current. If not stated otherwise, the adsorption potential for Ni(DMG)2 was -0.30 V, that for Co(DMG)2 was -0.75 V, and the stripping current was 50 µA. RESULTS AND DISCUSSION Adsorption Half-Time. Oxygen Removal. Figure 1(a) shows the stripping curves obtained from samples to which 0, 1, (11) Jagner, D.; Renman, L.; Wang, Y. Electroanalysis 1992, 4, 267-273.
Figure 1. Stripping curves obtained after 180 s of electrolysis followed by reduction with a constant current of 50 µA in 24.8 µL samples containing 0.10 mM dimethylglyoxime (pH 9.2) and to which 0, 1, 2, 4, and 8 µg L-1 (a) nickel(II) and (b) cobalt(II) have been added.
Figure 2. Stripping signals (ms) obtained after 180 s of electrolysis followed by reduction with a current of 50 µA in samples containing 0-400 µg L-1 nickel(II).
2, 4, and 8 µg L-1 Ni(II) has been added, and Figure 1b the corresponding stripping curves for Co(II). Separate experiments, where the vibrational electrolysis time was varied between 20 and 600 s, revealed that the half-time for the adsorption process was ∼35 s, i.e., more than 98% of the complexes were adsorbed during an 180 s period of electrolysis. By varying the electrolysis time, it could also be concluded that dissolved dioxygen was removed completely after ∼100 s of vibration in the nitrogen atmosphere. Determination of the Chemical Current. Samples containing either 1 µg L-1 Ni(II) or 1 µg L-1 Co(II) were analyzed using reductive currents equal to 10, 30, and 50 µA. By plotting the inverse of the stripping times vs the stripping currents, chemical currents equal to 3.1 µA for nickel (r ) 0.9996, n ) 3) and 3.4 µA for cobalt (r ) 0.9997, n ) 3) were obtained by linear extrapolation. Linear Concentration Range. Electroanalytical stripping techniques, exploiting adsorptive enrichment, are surface techniques. Consequently, a linear relationship between metal complex concentration and stripping signals, or stripping signals and electrolysis time, can be expected only at low relative electrode surface loadings, typically less than 5%. That this is the case is shown in Figure 2 , where samples containing 0-400 µg L-1 Ni(II) have been analyzed. After 180 s electrolysis, surface saturation occurs at a Ni(II) concentration equal to ∼200 µg L-1. A linear relationship is, however, obtained in the 0-4 µg L-1 concentration Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
1783
Table 1. Number of Electrons Involved in the Reduction of Ni(DMG)2 and Co(DMG)2 Calculated According to Eq 4a analyte ion
concn (µg L-1)
stripping signal (ms)
chem current (µA)
no. of electrons
Ni(II) Ni(II) Ni(II) Ni(II) Ni(II) Ni(II) Ni(II) Co(II) Co(II) Co(II) Co(II) Co(II)
1.0 1.0 1.0 2.0 2.0 4.0 4.0 1.0 1.0 2.0 2.0 4.0
8.789 8.856 8.844 16.978 17.367 35.044 34.989 8.689 8.578 16.689 17.322 34.333
3.1 3.1 3.1 3.1 3.1 3.1 3.1 3.4 3.4 3.4 3.4 3.4
10.1 10.2 10.2 9.8 10.0 10.1 10.1 10.0 9.9 9.6 10.0 9.9
H H3C H3C
range for both nickel (r ) 0.999 97, intercept 0.01 µg L-1, n ) 3) and cobalt (r ) 0.999 98, intercept 0.01 µg L-1, n ) 3). Number of Electrons in the Reduction of M(DMG)2. By repetitive analysis of samples containing 1, 2, or 4 µg L-1 Ni(II) or Co(II) the number of electrons, n, involved in the reduction of Ni(DMG)2 and Co(DMG)2 could be determined using eq 4. The results are summarized in Table 1. The mean n values, 10.1 for Ni(II) and 9.9 for Co(II), with an estimated standard deviation of 0.2, clearly show that there are 10 electrons involved in the reduction. This conclusion is supported not only by the standard deviation but also by the fact that, due to the stoichiometry of the complex and the facts that neither nickel nor cobalt are stable in the monovalent state and that the two ligands are identical, odd numbers of electrons, i.e., 9 or 11, can be excluded. Proposed Mechanism. Most reagents used for adsorptive stripping determination of trace elements provide little stripping potential selectivity between different elements. The reason for this is that the reagent, and not the metal ion in the complex, is reduced. The difference between the reduction potentials of the Ni(II) and Co(II) DMG complexes is ∼0.11 V. It is thus reasonable to believe that the initial step is the reduction of the central ion according to
(5)
followed by a four-electron reduction of each glyoximate ligand. It was shown early by Spritzer and Meites12 that complete electrochemical reduction of DMG results in the formation of 2,3diaminobutane, involving 8 electrons/DMG molecule. Although this supports the conclusions drawn by Vukomanovic et al.,9 the results were obtained in acidic media and are, therefore, not applicable to the alkaline conditions employed in adsorptive stripping applications. Also, the work of Spritzer and Meites12 indicates that, at potentials below the reduction current plateau for DMG, an intermediate product is formed. It has been shown that, under alkaline conditions, reduction of DMG produces a (12) Spritzer, M.; Meites, L. Anal. Chim. Acta 1962, 26, 58-65. (13) Bossa, V.; Morpurgo, G.; Morpurgo, L. Ric. Sci. 1967, 37, 402-407. (14) Jagner, D.; Ma, F.; Wang, Y. Electroanalysis 1996, 8, 952-954. (15) Jagner, D.; Sahlin, E.; Renman, L. Talanta 1994, 41, 515-522.
1784
Analytical Chemistry, Vol. 69, No. 9, May 1, 1997
N C
O–
C
OH N
H3C
+5
H+
+
4–e
N CH
OH
CH
OH
(6) H3C
N H
a Potentiostatic adsorption from a 24.8 µL sample drop in a nitrogen atmosphere for 180 s prior to reductive stripping with a constant current of 50 µA.
M(II)-DMG2(ads) + 2e- f M(0) + 2DMG-(ads)
diffusion-controlled current that is approximately half of that obtained under acidic conditions,13 i.e., that a reaction involving four electrons occurs. It is reasonable to assume that this reduction does not involve the breaking of the N-O bonds in the ligand. Thus, the formation of hydroxylamine groups, viz
is most likely to be the dominating reaction. Consequently, the reduction of the M(II)-DMG2 complex requires 10 electrons, 2 for the metal ion, and 4 for each of the two ligands. Analytical Implications. Determination of Ni(II) in Certified Reference Samples. Coulometric stripping potentiometry, which, since it is based on the fundamental Faraday’s law, is a calibration-free technique, has hitherto only been used in oxidative mode, e.g., for the determination of Pb(II) and Cd(II) subsequent to reductive amalgamation.10,14 Experimentally, reductive coulometric stripping is, however, more complicated than its oxidative counterpart since, in order to obtain maximum sensitivity and reproducibility, dissolved oxygen must be removed from the samples. The results presented here shows that the technique can also be used in the reductive mode, for the calibration-free determination of Ni(II) and Co(II) in the approximate concentration range of 0-5 µg L-1. This was confirmed by analyzing two certified reference samples issued by the Natural Research Council of Canada, NASS-4 seawater and SLRS-2 river water, the certified values being 0.228 ( 0.009 and 1.03 ( 0.10 µg L-1 Ni(II), respectively. The values obtained by reductive stripping coulometry according the procedure described in the present work were, after correction for the reagent blank, 0.238 (sd ) 0.01, n ) 10) and 1.04 (sd ) 0.01, n ) 7) µg L-1, respectively. The instrumental detection limit in stripping potentiometry, using a 90 kHz instrument, is ∼0.2 ms.15 Under the conditions used in the present work, with a sample volume of 24.8 µL and a stripping current of 50 µA, eq 4 shows that this corresponds to Ni(II) and Co(II) concentration limits of ∼0.03 µg L-1 and mass detection limits of ∼0.75 pg. No doubt, this detection limit is relevant for Co(II), but due to nickel impurities in the DMG reagent, the detection limit for this element is ∼0.10 µg L-1. As can be expected for a coulometric technique, the relative repeatability is very satisfactory, a value of ∼2 % being estimated from the data in Table 1. ACKNOWLEDGMENT This work was supported by the Swedish Natural Science Research Council. We are indebted to Prof. David Dyrssen for fruitful discussions regarding the chemistry of dimethylglyoxime metal chelates. Received for review October 3, 1996. Accepted February 4, 1997.X AC961023S X
Abstract published in Advance ACS Abstracts, April 1, 1997.
