Anal. Chem. 2000, 72, 2395-2400
Oxidation-State Speciation of [ReI(DMPE)3]+/ [ReII(DMPE)3]2+ by Voltammetry with a Chemically Modified Microelectrode Zhongmin Hu and William R. Heineman*
Department of Chemistry, University of Cincinnati, P.O. Box 210172, Cincinnati, Ohio 45221-0172
The analytical utility of chemically modified microelectrodes for oxidation-state speciation of redox couples by cyclic voltammetry has been explored. [ReI(DMPE)3]+/ [ReII(DMPE)3]2+, where DMPE ) 1,2-bis(dimethylphosphino)ethane, was studied at carbon-fiber microelectrodes of ∼5 µm in radius coated with Nafion-entrapped solgel-derived silica (Nafion-silica) composite. The results are compared with cyclic voltammetry of [Fe(CN)6]3-/ [Fe(CN)6]4- at bare carbon-fiber microelectrodes. At both microelectrodes, the cathodic and anodic limiting currents are linearly proportional to the concentrations of the reducible and oxidizable species of a redox couple, respectively. The shape of the cyclic voltammogram and the magnitude of the steady-state limiting current are not affected by the potential at which the scan starts. Speciation of both forms of a redox couple could be achieved voltammetrically at the microelectrodes. However, a considerably slower scan rate was required to achieve steady state at the modified electrode because of the smaller diffusion coefficients of [ReI(DMPE)3]+ and [ReII(DMPE)3]2+ in the Nafion-silica composite. The detection limit at the modified electrode was considerably lower (5 × 10-9 M for [ReI(DMPE)3]+) than at the bare electrode (6 × 10-5 M for [Fe(CN)6]3- and [Fe(CN)6]4-) because of the substantial preconcentration of [ReI(DMPE)3]+ by the Nafion-silica composite. Speciation, the identification and concentration determination of different forms in which an ion or compound is found, is important in some environmental and clinical sample analyses. Generally, speciation involves the identification and concentration determination of the forms different in chemical environment or the forms distinct in oxidation state. Electrochemical methods have played a role among the techniques used for speciation of the forms different in chemical environment. For instance, voltammetry has been explored for the speciation of free ion, weak complexes, and strong organic complexes of zinc,1 the speciation of model copper complexes,2 and the speciation of free copper ion and chemically labile copper complexes.3 In addition, voltammetric speciation of copper ions present in different coordination environments was suggested with electrodes modified with ligands (1) Xue, H. B.; Sigg, L. Anal. Chim. Acta 1994, 284, 505-515. (2) Morrison, G. M. P.; Florence, T. M. Electroanalysis 1989, 1, 485-491. (3) Labuda, J.; Korgova´, H.; Vanı´ckova´, M. Anal. Chim. Acta 1995, 305, 4248. 10.1021/ac991201s CCC: $19.00 Published on Web 05/02/2000
© 2000 American Chemical Society
whose formation constants for copper vary over a wide range.4 A few electrochemical speciation analyses involving the concentration determination of the forms distinct in oxidation state have also been performed, as documented for the determination of ferrous and ferric concentrations by using polarography.5,6 Such an oxidation-state speciation is justified by the periodic renewal of the surface of the dropping mercury electrode, since the renewed surface of the electrode senses the original concentration distribution of the species of a redox couple in bulk solution during the course of polarographic measurement. It can be conceived that hydrodynamic voltammetry (HDV) would have a similar or better usefulness for oxidation-state speciation of a redox couple, as the solution in the very vicinity of a rotating electrode is continuously renewed, which makes the electrode continuously respond to the concentration profile of the bulk solution. Both polarography and HDV, nevertheless, would not be appropriate for some situations, especially in the case of in vivo applications. It is of practical importance to be able to determine the concentrations of both species of a redox couple in vivo in metabolism studies of radiopharmaceuticals that are oxidizable or reducible in vivo.7 Many radiopharmaceuticals used in nuclear medicine are electroactive, and the chemical composition of an injected radiopharmaceutical is, in general, known. The specific chemical form that is actually responsible for imaging or radiotherapy, however, may be altered due to in vivo redox reaction, as evidenced by clinical results which show that in vivo redox reactions can markedly affect the biodistribution patterns of technetium-99m complex cations.8 Microelectrode sensors have been developed to measure [ReI(DMPE)3]+, where DMPE is 1,2-bis(dimethylphosphino)ethane, a nonradioactive analogue of the imaging agent [99mTc(DMPE)3]+, in vivo.9 Further efforts directed toward in vivo speciation of a redox couple, such as [ReI(DMPE)3]+/[ReII(DMPE)3]2+, with microelectrodes may acquire more information useful for nuclear medicine research. One of the conspicuous properties of microelectrodes is that the current incurred is extremely small while the current density is exceptionally high, because of the small sensing area and high (4) Cha, S. K.; Abrun ˜a, H. D. Anal. Chem. 1990, 62, 274-278. (5) Parry, E. P.; Anderson, D. P. Anal. Chem. 1973, 45, 458-463. (6) Beyer, M. E.; Bond, A. M.; McLaughlin, R. J. W. Anal. Chem. 1975, 47, 479-482. (7) Heineman, W. R.; Swaile, B. H.; Blubaugh, E. A.; Landis, D. A.; Seliskar, C. J.; Deutsch, E. Radiochim. Acta 1993, 63, 199-203. (8) Deutsch E.; Hirth, W. J. Nucl. Med. 1987, 28, 1491-1500. (9) Lee, M. T. B.; Seliskar, C. J.; Heineman, W. R.; McGoron, A. J. J. Am. Chem. Soc. 1997, 119, 6434-6435.
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000 2395
mass-transport rate resulting from nonplanar diffusion. This characteristic of microelectrodes has brought about many electrochemical investigations in poorly conductive media10-16 and investigations regarding kinetics of fast heterogeneous electron transfer.17,18 This unconventional property of microelectrodes should also be applicable to oxidation-state speciation. The extremely small current incurred in measurements with microelectrodes means that only an extremely limited amount of electroactive species is reduced or oxidized, leaving the system virtually intact in the process of determination. Thus, electrochemical speciation of both forms of a redox couple could be carried out to a great advantage by using microelectrodes. To our knowledge, however, microelectrodes have not been exploited for speciation per se, even though the capability is implicit in the nature of the steady-state voltammogram that is characteristic of microelectrode behavior. In this work, the utility of microelectrodes for oxidation-state speciation of redox couples by cyclic voltammetry is explored. Our ultimate goal is to accomplish speciation in vivo of radiopharmaceuticals such as [ReI(DMPE)3]+, which undergoes in vivo oxidation to [Re(DMPE)3]2+. Such an oxidation destroys the monocationic nature of the complex, which is requisite for heart imaging.19 The very low concentrations of [ReI(DMPE)3]+ and the potential interference by other redox species such as ascorbate in this application dictate that the microelectrode be modified with a permselective coating that will impart the requisite detection limit and selectivity by preconcentrating [ReI(DMPE)3]+ and excluding anions such as ascorbate.9 Thus, [ReI(DMPE)3]+/[ReII(DMPE)3]2+ was studied at carbon-fiber microelectrodes of ∼5 µm in radius coated with Nafion-entrapped sol-gel-derived silica (Nafion-silica) composite. We have shown previously that films of pure Nafion, Nafion gel, and Nafion-silica composite all preconcentrate [ReI(DMPE)3]+.7,9,20-22 The results for this system are compared with cyclic voltammetry of [Fe(CN)6]3-/[Fe(CN)6]4at bare carbon-fiber microelectrodes as a representative, wellcharacterized system at an unmodified microelectrode. EXPERIMENTAL SECTION Apparatus. Electrochemical measurements were carried out with a BAS 100B/W electrochemical analyzer in conjunction with a PA-1 low-current module (Bioanalytical Systems, IN). The electrochemical cell consisted of a Ag/AgCl reference electrode, (10) Ciszkowska, M.; Stojek, Z.; Morris, S. E.; Osteryoung, J. G. Anal. Chem. 1992, 64, 2372-2377. (11) Bond, A. M.; Pfund, V. B. J. Electroanal. Chem. 1992, 335, 281-295. (12) Thormann, W.; van den Bosch, P.; Bond, A. M. Anal. Chem. 1985, 57, 2764-2770. (13) Cooper, J. B.; Bond, A. M. Anal. Chem. 1993, 65, 2724-2730. (14) Kulesza, P. J.; Faulkner, L. R. J. Am. Chem. Soc. 1993, 115, 11878-11884. (15) Brina, R.; Pons, S.; Fleischmann, M. J. Electroanal. Chem. 1988, 244, 8190. (16) Jaworski, A.; Donten, M.; Stojek, Z.; Osteryoung, J. G. Anal. Chem. 1999, 71, 167-173. (17) Howell, J. O.; Wightman, R. M. Anal. Chem. 1984, 56, 524-529. (18) Montenegro, M. I.; Pletcher, D. J. Electroanal. Chem. 1986, 200, 371374. (19) Deutsch, E.; Libson, K.; Vanderheyden, J.-L.; Ketring, A. R.; Maxon, H. R. Nucl. Med. Biol. 1986, 13, 465-477. (20) Deng, Y.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 1997, 69, 40454050. (21) Swaile, B. H.; Blubaugh E. A.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 1998, 70, 4326-4332. (22) Hu, Z.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 1998, 70, 5230-5236.
2396
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000
a platinum wire auxiliary electrode, and a bare or chemically modified carbon-fiber microdisk electrode. The cell was mounted inside a Faraday cage to minimize noise pickup from the environment. Reagents. [ReI(DMPE)3]CF3SO3 was prepared as previously described.19 Nafion perfluorosulfonated ionomer, 5 wt % solution in a mixture of lower aliphatic alcohols and 10% water, and tetraethyl orthosilicate (TEOS, 98%) were purchased from Aldrich and used as received. Phosphate buffer stock solution (Fisher Scientific), after adjustment to pH 7.4 with NaOH, was used to control the pH of [ReI(DMPE)3]+ solutions. As [ReI(DMPE)3]+ in solution is apt to be oxidized by dissolved oxygen,21 [ReI(DMPE)3]+ solutions were freshly prepared prior to measurements by dissolving [ReI(DMPE)3]CF3SO3 in 0.14 M NaCl (pH 7.4) supporting electrolyte solution that had been deoxygenated by argon purging. [ReII(DMPE)3]2+ solution was derived by electrolyzing a freshly prepared [ReI(DMPE)3]+ solution. Ferricyanide and ferrocyanide solutions were prepared from K3Fe(CN)6 (Baker) and K4Fe(CN)6‚3H2O (MC & B), respectively, with potassium chloride (Fisher Scientific) as supporting electrolyte. The pH of the ferricyanide and ferrocyanide solutions was not deliberately adjusted but was found to be 6.5-7.5 with a pH meter. All of the solutions were prepared with water purified from a Barnstead Nanopure water system (Barnstead Sybron Corp.). Solutions were deoxygenated with argon prior to voltammetric measurements and blanketed with argon during measurements. Microelectrode Fabrication. The microelectrodes were constructed by soldering a nominally 10-µm-diameter carbon fiber to a copper wire of 0.25-mm diameter with indium solder. It was suggested that using a small amount of solder instead of a larger amount of silver epoxy reduces stray capacitance as the result of the reduction in surface area of conductive material inside the electrode.23 With indium solder, it is easier to connect the fiber to copper wire than with lead/tin solder. Following the soldering, the assembly was inserted halfway into 1.5-mm-i.d. glass capillary tube. The protruding end of the copper wire was affixed to the capillary with Instant Bonding Adhesive (Archer, Tandy Corp.), with the capillary being unsealed. The capillary containing the assembly was then pulled with a glass pipet puller (Narishige Scientific Instrument). The heat and magnet controllers on the puller were adjusted so that the cone height of the micropipet was short but the glass formed an almost snug fit around the carbon fiber that extends out from the glass tip. Afterward, the micropipet tip was dipped in liquid Teflon (Teflon AF, 6%, DuPont) and the electrode was cured at 70 °C in an oven for 20 min in order to coat the fiber and seal the tip, which may be left unsealed in the above pulling step. After another dipping in liquid Teflon and curing, the protruding carbon fiber was carefully cut under a magnifying glass so that the fiber tip is virtually flush with or slightly sticks out of the glass micropipet tip. Although the carbon fiber, in some cases, slightly protrudes from the seal at the tip, the sensing area of the electrode is expected to be still defined by the radius of the carbon fiber because of the Teflon coating. The tips of the constructed microelectrodes were visually examined under an Optiphot metallurgical microscope (Nikon). It is assumed that, although the surface of the microelectrodes thus constructed might be rather rough, the macroscopical geometry (23) Tschuncky, P.; Heinze, J. Anal. Chem. 1995, 67, 4020-4023.
of the sensing tip could be best represented by a disk. The actual radius of the microelectrodes was determined electrochemically (vide infra). Microelectrode Modification. Nafion-incorporated silica sol was prepared at room temperature, as described elsewhere.22 The initial molar ratio of water relative to the precursor TEOS was 20:1, and the initial concentration of catalyst HCl was ∼2 × 10-3 M. An aliquot of 1 mL of the formed sol was then mixed with 0.9 mL of the as-received 5 wt % Nafion solution by sonicating for 5 min, with the mass ratio of Nafion ionomer relative to sol-gelderived silica (SiO2) being 2:5. The fabricated carbon-fiber-based microelectrodes were modified with Nafion-silica composite by dipping them in the Nafion-incorporated silica sol and drying under ambient conditions. The modified microelectrodes were conditioned in 0.14 M NaCl (pH 7.4) supporting electrolyte solution overnight, before electrochemical measurements were carried out. RESULTS AND DISCUSSION Voltammetric Characterization of Microelectrodes. Carbon fibers are amenable to the construction of microelectrodes that are suitable for voltammetric applications.17,24 On the other hand, due to possible false soldering of a carbon fiber to the hookup copper wire and the imperfection in cutting the fiber protruding from the glass micropipet tip, some of the fabricated microelectrodes can exhibit no response or non-steady-state response. In addition, the radius of an individual carbon fiber may deviate significantly from the nominal radius. The performance and the radius of each of the constructed bare microelectrodes were thus electrochemically evaluated in 4 × 10-3 M [Fe(CN)6]3- solution of 0.1 M KCl. For a disk-shaped microelectrode, the steady-state limiting current is given by 25
Figure 1. Cyclic voltammograms at a Nafion-silica composite modified carbon-fiber microelectrode in 1 × 10-4 M [ReI(DMPE)3]+ at various scan rates as indicated. Initial potential, -0.4 V; switching potential, 0.45 V; supporting electrolyte, 0.14 M NaCl (pH 7.4). Before the measurements were taken, the modified microelectrode was soaked in the [ReI(DMPE)3]+ solution for 30 min.
With the measured steady-state limiting current ilim, the known value of bulk concentration C*, and the documented diffusion coefficient D of 7.62 × 10-6 cm2/s for 4 × 10-3 M [Fe(CN)6]3- in 0.1 M KCl aqueous solution,26 the radii of the microelectrodes can be determined. Only those electrodes whose electrochemically determined radii were ∼5 µm, which is the nominal radius of the carbon fiber used in this work, were used for further studies. Figure 1 shows typical cyclic voltammograms of Nafion-silica composite modified microelectrodes in [ReI(DMPE)3]+ solution at various scan rates. The film on the electrode was first loaded with [ReI(DMPE)3]+ by soaking in solution for 30 min. Compared to the cyclic voltammetric response of bare carbon-fiber microelectrodes in [Fe(CN)6]3- solution at various scan rates (figure not shown), the modified microelectrode exhibited steady-state response at a slower scan rate. The achievement of steady state depends not only on the radius of an electrode but also on the diffusion coefficient of the electroactive species involved. The diffusion that hydrophobic cations undergo in Nafion is ap-
proximately 1000-fold slower than in aqueous solution.27-29 The diffusion of [ReI(DMPE)3]+ has been shown to be faster in Nafion-silica composite than in Nafion,22 but it is still significantly slower than in aqueous solution. In 0.15 M NaCl aqueous solution, the diffusion coefficient of [ReI(DMPE)3]+ is 5.0 × 10-6 cm2/s.20 Due to the decreased diffusion coefficients of the Re complex cations in the Nafion-containing film, a potential scan slower than the one usually used to obtain steady-state voltammograms with bare microelectrodes is needed to achieve a steady-state response at the modified microelectrode. For oxidation-state speciation, it is desirable to reach truly steady-state conditions. However, the time needed might be unaffordable. As can be seen from the voltammogram shown in the upper left panel of Figure 1, at the scan rate of 1 mV/s there exists a gap between the forward and reverse voltammetric tracks, indicating the incompleteness of the steady state. At this scan rate, the deviation from pure steadystate response was estimated to be ∼7% by calculating the maximum error due to the incompleteness of the steady state at the half-wave potential.30 Speciation of Redox Couples. For speciation, an important criterion is that measurements should be made with negligible perturbance to the original distribution of species. With voltammetry carried out at a conventional-sized electrode, when an initial potential, usually several hundred millivolts more positive or negative than the apparent formal potential of a redox couple, is applied, the original distribution of species of the redox couple in the vicinity of the electrode is perturbed by the electrolysis occurred upon application of the potential. One might run voltammetry in the anodic region and in the cathodic region, respectively, by initiating the potential scan at a carefully prede-
(24) Dayton, M. A.; Brown, J. C.; Stutts, K. J.; Wightman, R. M. Anal. Chem. 1980, 52, 946-950. (25) Wightman, R. M.; Wipf, D. O. In Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, 1989; Chapter 3, Vol. 15. (26) von Stackelberg, M.; Pilgram, M.; Toome, V. Z. Elektrochem. 1953, 57, 342-350.
(27) Martin, C. R.; Rubinstein, I.; Bard, A. J. J. Am. Chem. Soc. 1982, 104, 48174824. (28) Porat, Z.; Rubinstein, I.; Zinger, B. J. Electrochem. Soc. 1993, 140, 25012507. (29) Martin, C. R.; Dollard, K. A. J. Electroanal. Chem. 1983, 159, 127-135. (30) Zoski, C. G. J. Electroanal. Chem. 1990, 296, 317-333.
ilim ) 4nFDrC*
(1)
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000
2397
Figure 2. (A) Cyclic voltammograms of [Fe(CN)6]3-/[Fe(CN)6]4- in 0.1 M KCl at a carbon-fiber microelectrode with scan rate of 10 mV/ s: (a) 1.0 mM [Fe(CN)6]3-; (b) 0.75 mM [Fe(CN)6]3- and 0.25 mM [Fe(CN)6]4-; (c) 0.5 mM [Fe(CN)6]3- and 0.5 mM [Fe(CN)6]4-; (d) 0.25 mM [Fe(CN)6]3- and 0.75 mM [Fe(CN)6]4-; (e) 1.0 mM [Fe(CN)6]4-. (B) Dependence of cathodic limiting current (measured from (A) at -0.55 V) and anodic limiting current (measured from (A) at 0.95 V) on [Fe(CN)6]3- and [Fe(CN)6]4- concentrations, respectively. The abscissa is the concentration of [Fe(CN)6]3- for cathodic limiting current and the concentration of [Fe(CN)6]4- for anodic limiting current. For the linear regression equations shown, when the concentration of the electroactive species is in millimolar, the limiting current is in nanoampere.
termined “rest” potential where, ideally, neither oxidation nor reduction would occur. The choice of this initial potential would be critical in order to keep the original concentration profile unperturbed. On the other hand, with voltammetry carried out at a microelectrode, the original species distribution of a redox couple in the vicinity of the electrode should be only negligibly disturbed by the process of a determination performed at slower scan rates. Bare Microelectrode for [Fe(CN)6]3-/[Fe(CN)6]4-. Our previous work with aqueous [ReI(DMPE)3]+ solution shows strong adsorption of [ReI(DMPE)3]+ on bare solid electrodes.21 Adsorption of [ReI(DMPE)3]+ does not occur at Nafion-silica composite modified electrodes.22 Adsorption causes a mixed mechanism for an electrode process that would prevent a direct comparison of the voltammograms observed. Consequently, ferro-ferricyanide couple was used as the model system to establish properties of the bare electrode. As shown in Figure 2A, when the solution contains only the reducible form, [Fe(CN)6]3-, or the oxidizable form, [Fe(CN)6]4-, of the redox couple, only cathodic limiting current resulting from the reduction of [Fe(CN)6]3- (curve a) or anodic limiting current resulting from the oxidation of [Fe(CN)6]42398 Analytical Chemistry, Vol. 72, No. 11, June 1, 2000
(curve e) is found. When the solution contains both species of the redox couple, for instance, an equal concentration of [Fe(CN)6]3- and [Fe(CN)6]4- (curve c), both cathodic and anodic limiting currents are found, and the magnitude of the cathodic limiting current is comparable to that of the anodic limiting current. When the concentration ratio of these two electroactive species is varied, with the total concentration of them being kept constant, the ratio of the cathodic and anodic limiting currents is correspondingly changed (curves b and d), with the magnitude of the cathodic and anodic limiting currents being linearly proportional to the concentrations of [Fe(CN)6]3- and [Fe(CN)6]4-, respectively (Figure 2B). The dashed line drawn at zero current in Figure 2A serves as the baseline for measuring the cathodic and anodic steady-state limiting currents from the voltammograms. As can be seen from Figure 2A, in the absence of the oxidizable form, [Fe(CN)6]4-, and in the anodic potential region where no [Fe(CN)6]3- is reduced, the cyclic voltammetric track superimposes on the baseline (cf. curve a). Similarly, when the solution contains merely the other moiety, [Fe(CN)6]4-, of the redox couple, the voltammetric track in the cathodic region overlaps the baseline (cf. curve e). Thus, it is convenient and appropriate to measure both cathodic and anodic steady-state limiting currents of the voltammograms against the zero-current baseline. The small deviation of the voltammetric track from the assumed baseline in the far cathodic region was most likely caused by the existence of residual oxygen in the solution. In addition, it was observed that the shape of the cyclic voltammogram and the magnitude of the limiting current are independent of the potential at which the scan starts, whether the scan initiates at the positive potential end, the negative end, or in between. This independence would be of advantage in practical applications, since the potential scan can be started at any potential without the need to take account of the species composition of the redox couple in the target solution. As Figure 2B shows, the slope of the linear dependence of the cathodic limiting current on the concentration of [Fe(CN)6]3is larger than the absolute value of the slope of the linear dependence of the anodic limiting current on the concentration of [Fe(CN)6]4-. This can be attributed to the difference in diffusion coefficients of the reducible and the oxidizable species of the redox couple, as eq 1 defines. Individual calibration curves are thus required for determining the concentrations of both species of a redox couple. For the [Fe(CN)6]3-/[Fe(CN)6]4- couple, the detection limits were found to be ∼6 × 10-5 M for both forms in the existence of 1 × 10-3 M of the counter form of the redox couple. Modified Microelectrode for [ReI(DMPE)3]+/[ReII(DMPE)3]2+. Figure 3 shows the steady-state cyclic voltammograms obtained in [ReI(DMPE)3]+, in [ReII(DMPE)3]2+, and in a mixture of [ReI(DMPE)3]+ and [ReII(DMPE)3]2+ solutions at a Nafion-silica composite modified microelectrode after a 30-min soak in the solution of interest. After each measurement, the modified microelectrode was regenerated by soaking it in 0.5 M HCl for 30 min to expel [ReI(DMPE)3]+ and [ReII(DMPE)3]2+ and then reconditioned in 0.14 M NaCl (pH 7.4) solution for 15 min. It is evident that, when the solution nominally contains only [ReI(DMPE)3]+ or [ReII(DMPE)3]2+, anodic limiting current resulting
Figure 3. Cyclic voltammograms of the [ReI(DMPE)3]+/[ReII(DMPE)3]2+ couple at a Nafion-silica composite modified carbonfiber microelectrode with a scan rate of 0.2 mV/s: (a) 1 × 10-4 M [ReI(DMPE)3]+; (b) 5 × 10-5 M [ReI(DMPE)3]+ and 5 × 10-5 M [ReII(DMPE)3]2+; (c) 1 × 10-4 M [ReII(DMPE)3]2+. Initial potential, -0.4 V; switching potential, 0.45 V; supporting electrolyte, 0.14 M NaCl (pH 7.4); open-circuit quiescent preconcentration time, 30 min. The inset shows the dependence of anodic limiting current (measured at 0.44 V) and cathodic limiting current (measured at -0.39 V) on [ReI(DMPE)3]+ and [ReII(DMPE)3]2+ concentrations, respectively. The abscissa is the concentration of [ReI(DMPE)3]+ for anodic limiting current and the concentration of [ReII(DMPE)3]2+ for cathodic limiting current.
from the oxidation of the monocation (a in Figure 3) or cathodic limiting current resulting from the reduction of the dication (c in Figure 3) is found. Whereas the nominal presence of both [ReI(DMPE)3]+ and [ReII(DMPE)3]2+ gives rise to both anodic and cathodic limiting currents (b in Figure 3). In the nominal presence of only [ReI(DMPE)3]+, the observed cathodic limiting current, although negligible in comparison with the anodic limiting current, is slightly greater than zero. This is attributed to the oxidation of a small portion of the monocation by residual oxygen in the solution during the preconcentration and voltammetric scan. The inset of Figure 3 shows that under the experimental conditions the anodic and cathodic limiting currents are linearly proportional to the concentration of [ReI(DMPE)3]+ and the concentration of [ReII(DMPE)3]2+, respectively. However, the slopes of the plots are very different. In the case of a chemically modified electrode, the slopes of these analytical calibration curves depend not only on the diffusion coefficients of the two species but also on their partition coefficients. The larger slope for the anodic limiting current is indicative of greater partitioning of [ReI(DMPE)3]+ into the film and/or a larger diffusion coefficient of [ReI(DMPE)3]+ in the film. Because of the preconcentration effect of the Nafion-silica composite, we expected to achieve lower detection limits at the modified electrode than for ferro-ferricyanide at a bare electrode. Shown in Figure 4 are representative cyclic voltammograms observed at the modified microelectrode in 1 × 10-6 (voltammogram a) and 1 × 10-8 M (voltammogram b) [ReI(DMPE)3]+ solutions. At these lower concentrations, a relatively larger portion of [ReI(DMPE)3]+ was oxidized to [ReII(DMPE)3]2+, as indicated
Figure 4. Representative cyclic voltammograms at Nafion-silica composite modified carbon-fiber microelectrodes in (a) 1 × 10-6 M [ReI(DMPE)3]+ and (b) 1 × 10-8 M [ReI(DMPE)3]+ solutions. Initial potential, -0.4 V; switching potential, 0.45 V; scan rate, 0.2 mV/s; open-circuit quiescent preconcentration time, 40 min; supporting electrolyte, 0.14 M NaCl (pH 7.4).
by the more appreciable cathodic limiting current in the nominal presence of only [ReI(DMPE)3]+. In addition, at the lower concentrations, the current-voltage curve of the reverse scan noticeably deviates from that of the forward scan, even at the very low scan rate of 0.2 mV/s. The more appreciable difference between the current-voltage curves of the forward and reverse scans than the incompleteness of steady state might be ascribed to the gradual decrease in the concentration ratio of [ReI(DMPE)3]+ to [ReII(DMPE)3]2+, manifested in the time scale of a voltammetric scan at this slow scan rate, because of the oxidation of [ReI(DMPE)3]+ by residual oxygen. The detection limit was found to be ∼5 × 10-9 M [ReI(DMPE)3]+ for a 40-min preconcentration period. Diffusion Coefficient Assessment. At a voltammetric microelectrode, the diffusion coefficient of an electroactive species can be determined if the electrode radius and the number of electrons, n, associated with an electrode reaction are known.31,32 When the radius of a microelectrode is not exactly known, the ratio of diffusion coefficients of the reducible and the oxidizable species of a reversible redox couple can be obtained from the linear relationship between the concentration of the electroactive species and the corresponding steady-state limiting current, as defined in eq 1. It is apparent that, when the same electrode is used, the ratio of diffusion coefficients of the species of a redox couple equals the ratio of the slopes of cathodic and anodic calibration curves. In the case of the [Fe(CN)6]3-/[Fe(CN)6]4- couple, the ratio of the diffusion coefficient of [Fe(CN)6]3- to that of [Fe(CN)6]4- in solution was determined from the slopes of the calibration curves, as shown in Figure 2B, to be 1.16. The determined ratio agrees with the ratio calculated from the diffusion coefficients of [Fe(CN)6]3- and [Fe(CN)6]4- in the literature.26 With modified electrodes, the above assessment of diffusion coefficient ratio by the ratio of the slopes of calibration curves might not be justified, because the concentration ratio of the two (31) Baur, J. E.; Wightman, R. M. J. Electroanal. Chem. 1991, 305, 73-81. (32) Denuault, G.; Mirkin, M. V.; Bard, A. J. J. Electroanal. Chem. 1991, 308, 27-38.
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000
2399
species of a redox couple in the modifying medium might not be the same as the concentration ratio of the species in solution, due to the probable difference in partition coefficients. In the case of the [ReI(DMPE)3]+/[ReII(DMPE)3]2+ couple, this difference in partition into the Nafion-silica composite might be substantial, as indicated by the large difference in slopes of the calibration curves shown in Figure 3. Hence, it might not be correct to use the slope ratio for the assessment of diffusion coefficient ratio. CONCLUSIONS Speciation of both forms of a redox couple at chemically modified microelectrodes was shown to be feasible as demonstrated with a Nafion-silica composite modified carbon-fiber microelectrode for the [ReI(DMPE)3]+/[ReII(DMPE)3]2+ couple. Since the concentrations of both moieties of a reversible redox couple can be determined simultaneously, without the need of meticulously predetermining the “rest” potential of the couple in a particular sample, the convenience and relative versatility of this proposed approach can be appreciated. In addition, this approach may open a way to access in vivo speciation of redox couples,
2400
Analytical Chemistry, Vol. 72, No. 11, June 1, 2000
such as the [ReI(DMPE)3]+/[ReII(DMPE)3]2+ couple. Surface modification would not only provide a microelectrode with selectivity but also enhance sensitivity and improve detection limit. On the other hand, as the diffusion of an electroactive species in a film is, in general, slower than in solution, a slower scan rate or smaller radius of the electrode tip is necessary to achieve a steadystate response. With a slow scan rate, the time needed to acquire a voltammogram would be longer, leading to the prolongation of an experiment. In this regard, the use of smaller microelectrodes for maintaining a steady-state condition would be of advantage. ACKNOWLEDGMENT Support provided by the Department of Energy (Grant 86ER60487) is gratefully acknowledged. Helpful comments on the manuscript from Prof. Harry B. Mark, Jr. are also acknowledged.
Received for review October 18, 1999. Accepted February 17, 2000. AC991201S
