Anal. Chem. 1997, 69, 4045-4050
Electrochemical Behavior of [ReI(DMPE)3]+, Where DMPE ) 1,2-Bis(dimethylphosphino)ethane, at Perfluorosulfonated Ionomer-Modified Electrodes Yingping Deng,† Carl J. Seliskar, and William R. Heineman*
Department of Chemistry, University of Cincinnati, P.O. Box 210172, Cincinnati, Ohio 45221-0172
The perfluorosulfonated ionomer Nafion shows potential utility as a polymer film to enhance the electrochemical detection of [ReI(DMPE)3]+, where DMPE ) 1,2-bis(dimethylphosphino)ethane. [ReI(DMPE)3]+, a nonradioactive radiopharmaceutical analog for heart imaging, partitions strongly into Nafion films on glassy carbon. Well-defined, chemically reversible cyclic voltammograms are obtained for the [ReI(DMPE)3]+/[ReII(DMPE)3]2+ couple with E°′ shifted positively by 60 mV relative to its value on bare glassy carbon. [ReI(DMPE)3]+ partitions into Nafion more strongly than the oxidized form, [ReII(DMPE)3]2+. The detection limit for [ReI(DMPE)3]+ by cyclic voltammetry was improved by 2-3 orders of magnitude by the Nafion film. Differential pulse voltammetry for oxidation of [ReI(DMPE)3]+ at the Nafionmodified electrode has a detection limit of 2.5 × 10-9 M compared to 1.0 × 10-7 M at the bare electrode. A preconcentration factor of 1 × 106 for partitioning of [ReI(DMPE)3]+ from 0.05 M NaCl into Nafion on a glassy carbon electrode was measured.
Extensive development and application of radiopharmaceuticals for use as organ imaging agents and therapy in nuclear medicine have occurred during the past two decades.1,2 Nuclear medicine procedures are based on the tendency of the body to concentrate some chemical form of a particular γ-ray-emitting isotope in the organ of interest; subsequent scanning of the organ with a γ-ray camera provides an image from which diagnostic information can be obtained. Images of the skeleton and of organs such as heart, brain, kidney, liver, lung, and thyroid are routinely obtained for diagnostic purposes. Radioactivity can also be used for therapy when isotopic agents are directed to specific organs by choice of their chemical forms.3 The exact chemical form of the radiopharmaceutical has proven to be an important factor in obtaining the optimum image of the organ of interest.2 Although the chemical form of the radiopharmaceutical may be known, it may be altered after injection due to in vivo reactions. Such changes in chemical †
Current address: Diagnostics Division, Bayer Corp., 1884 Miles Ave., Elkhart, IN 46514. (1) Cox, P. H., Mather, S. J., Sampson, C. B., Lazarus, C. R., Eds. Progress in Radiopharmacy; Martinus Nijhoff Publishers: Boston 1986. (2) Nowotnik, D. P.; Nunn, A. D. Drug News Perspect. 1992, 5, 174-183. (3) Maxon, H. R., III; Thomas, S. R.; Hertzberg, V. S.; Schroder, L. E.; Englaro, E. E.; Samaratunga R.; Scher, H. I.’ Moulton, J. S.; Deutsch, E. A.; Deutsch, K. F.; Schneider, H. J.; Williams, C. C.; Ehrhardt, G. J. Sem. Nucl. Med. 1992, 22 (1), 33-40. S0003-2700(96)01295-4 CCC: $14.00
© 1997 American Chemical Society
composition may alter the imaging properties of a radiopharmaceutical since it is the chemical form of the radiopharmaceutical which targets a specific organ. This is a fundamental problem in nuclear medicine research since in this field one usually measures only radioactivity after injection of a radiopharmaceutical, which provides no information as to chemical form. In vivo monitoring of a radiopharmaceutical with a chemical sensor after injection into a test animal would provide a means of obtaining valuable information about the fate of an imaging agent after injection. Sensors implanted in specific organs of interest in order to determine the chemical form(s) of the radioactive complex that is accumulating in an organ would be especially useful for developing improved radiopharmaceuticals.4 In order to accomplish this goal, sensors for in vivo monitoring of Tc/Re complexes that are used in nuclear medicine are being developed. Since many imaging agents are electroactive and absorb light in the UV-visible range, both electrochemical and optical sensors are being considered.4 The optical sensors are based on the measurement of absorbance with a fiber optic.5 The electrochemical sensors are based on polymer-modified electrodes at which the agent is detected by voltammetry. The polymer film is expected to improve detection limit by extracting the imaging agent from the sample and enhancing its concentration at the electrode surface. It could also improve selectivity by discriminating against possible electroactive interferents. In this paper, we explore the potential utility of Nafion as a polymer film to enhance the electrochemical detection of [ReI(DMPE)3]+, where DMPE ) 1,2-bis(dimethylphosphino)ethane.
H3C
CH3
+
CH3
H3C
P P
P Re
P
P H3C
CH3
CH3 CH3 CH3 CH3
P CH3
CH3
This compound is the rhenium analogue of [TcI(DMPE)3]+, which has shown potential as a myocardial perfusion imaging agent.6 Both [186Re(DMPE)3]+ and [99mTc(DMPE)3]+ show comparable (4) Heineman, W. R.; Swaile, B. H.; Blubaugh, E. A.; Landis, D. A.; Seliskar, C. J.; Deutsch, E. Radiochim. Acta 1993, 63, 199-203. (5) Seliskar, C. J.; Jeng, M. L.; Landis, D. A.; Swaile, B. H.; Blubaugh, E. A.; Heineman, W. R.; Deutsch, E. Proceedings Biosensors 92; Elsevier Science Publishers: Oxford, England, 1992.
Analytical Chemistry, Vol. 69, No. 19, October 1, 1997 4045
uptake by the heart in biodistribution studies performed on rats.7 The experiments reported here were done on the rhenium complex, since it can be synthesized with a nonradioactive isotope, whereas the technetium complex can only be made in radioactive form. Because [Re(DMPE)3]+ is a positively charged and somewhat lipophilic compound, the perfluorosulfonated ionomer Nafion shows promise for this application due to its combination of cation-exchange and hydrophobic properties. This is not surprising given its reported strong affinity for hydrophobic cations, both organic and inorganic.8-11 EXPERIMENTAL SECTION Preparation of [ReI(DMPE)3]+ Solutions. Tris[1,2-bis(dimethylphosphino)ethane]rhenium(I) triflate ([ReI(DMPE)3]CF3SO3) was synthesized and purified using a method described previously.7 For preparation of very dilute aqueous solutions (1.0 × 10-4 M or lower), [ReI(DMPE)3]CF3SO3 was directly dissolved in distilled water which had been purged with N2 gas. For preparation of the more concentrated 1 mM [ReI(DMPE)3]CF3SO3 solution, 4% ethanol was added. Freshly prepared [ReI(DMPE)3]+ aqueous solution was colorless but gradually turned pink due to oxidation to the dication, [ReII(DMPE)3]2+, by dissolved oxygen.12 For reproducible results, a fresh solution was prepared for each experiment. Preparation of Nafion Film-Modified Electrode. Nafion (5 wt %; EW ) 1100) in a mixture of lower aliphatic alcohols and 10% water was purchased from Aldrich Chemical Co. and diluted to 0.5% Nafion with 2-propanol for preparing Nafion film-modified glassy carbon electrodes. A measured amount of diluted Nafion solution was transferred to the surface of the glassy carbon, and the 2-propanol was then allowed to evaporate in air at room temperature for about 20 min. The thickness of the Nafion film was controlled by the volume and concentration of applied Nafion solution (typically 4-7 µL of 0.5% Nafion). The film thickness was calculated from the amount of Nafion applied to the electrode, the surface area of the electrode, and the density of Nafion film (1.35 g/cm3).8,13 Film thicknesses of 600 and 400 nm were used for experiments on [ReI(DMPE)3]+ at higher concentrations (>1.0 × 10-8 M) and lower concentrations (e1.0 × 10-8 M), respectively. The concentrations of [ReI(DMPE)3]+ extracted into the Nafion film on the surface of Nafion-modified electrodes were calculated from the charge required for its quantitative oxidation as measured by coulometry.14 Apparatus and Materials. Cyclic voltammetry and differential pulse voltammetry were performed using a Bioanalytical Systems BAS-100 electrochemical analyzer. The voltammograms were recorded on a Houston Instrument Hiplot digital plotter. Glassy carbon (diameter 3 mm), platinum disk (diameter 1.6 mm), (6) Gerson, M. C.; Deutsch, E. A.; Libson, K. F.; Adolph, R. J.; Ketring, A. R.; Vanderheyden, J.-L.; Williams, C. C.; Saenger, E. L. Eur. J. Nucl. Med. 1984, 9, 403-407. (7) Deutsch, E. A.; Libson, K. F.; Vanderheyden, J.-L.; Ketring, A. R.; Maxon, H. R. Nucl. Med. Biol. 1986, 13, 465-477. (8) Martin, C. R.; Freiser, H. Anal. Chem. 1981, 53, 902-904. (9) Martin, C. R.; Rubinstein, I.; Bard, A. J. J. Am. Chem. Soc. 1982, 104, 48174824. (10) Martin, C. R.; Dollard, K. A. J. Electroanal. Chem., 1983, 159, 127-135. (11) Szentirmay, M. N.; Martin, C. R. Anal. Chem. 1984, 56, 1898-1902. (12) Swaile, B. H. Ph.D. Dissertation, University of Cincinnati, Cincinnati, OH, 1994. (13) Shi, C. N.; Anson, F. C. J. Am. Chem. Soc. 1991, 113, 9564-9570. (14) Anson, F. C.; Blauch, D. N.; Saveˆant, J.-M.; Shu, C. F. J. Am. Chem. Soc. 1991, 113, 1922-1932.
4046 Analytical Chemistry, Vol. 69, No. 19, October 1, 1997
Figure 1. Cyclic voltammogram of 1.0 mM [ReI(DMPE)3]+ in aqueous 0.15 M NaCl, 4% EtOH; glassy carbon electrode vs Ag/ AgCl; scan rate 100 mV/s.
and Ag/AgCl (3 M NaCl) reference electrodes were purchased from Bioanalytical Systems. Highly ordered pyrolytic graphite (diameter 6.4 mm) was purchased from Union Carbide. Prior to use, glassy carbon electrodes were polished with 50 nm alumina paste (Buehler) on Microcloth (Buehler) followed by ultrasonic cleaning in purified water. Sodium chloride (Fisher, reagent grade) was used for the supporting electrolyte. All solutions were prepared with purified water from a Barnstead Organic Pure water system. Solutions were bubbled with N2 before use and blanketed with N2 during voltammetric measurements. RESULTS AND DISCUSSION Voltammetric Behavior of [ReI(DMPE)3]+ at Bare Electrodes. General Cyclic Voltammetry. The cyclic voltammogram shown in Figure 1 illustrates the electrochemical behavior of [ReI(DMPE)3]+ at bare electrodes in aqueous 0.15 M NaCl with 4% ethanol. Ethanol was used to dissolve the solid compound and to increase the solubility of [ReI(DMPE)3]CF3SO3 in aqueous 0.15 M NaCl solution. Two well-defined, chemically reversible redox couples are observed at E°′ ) +36 mV and E°′ ) +1000, and an irreversible oxidation peak is located at about +750 mV. The results obtained are qualitatively consistent with voltammograms in an organic solvent such as N,N-dimethylformamide.15 The less positive reversible couple is [ReI(DMPE)3]+/[ReII(DMPE)3]2+, which is the couple of interest for the detection of [ReI(DMPE)3]+ by the proposed sensor. It is also the “biologically active” couple in the sense that in vivo oxidation of ReI to ReII can occur. The reversible couple at the more positive potential is [ReII(DMPE)3]2+/[ReIII(DMPE)3]3+. The irreversible oxidation peak with Ep ) +750 mV has been shown to be the oxidation of a ReII species that is a proposed as an intermediate in the dissociation of a DMPE ligand.15 In this paper, we focus on the redox couple at E°′ ) + 36 mV for the purpose of sensing [ReI(DMPE)3]+ concentration. Effect of Solvent. The solubility of [ReI(DMPE)3]CF3SO3 in neutral aqueous solution at room temperature is about 5 × 10-4 M. This low solubility caused by the hydrophobicity of the DMPE ligands undoubtedly plays an important role in its accumulation in the heart as well as in its partitioning into Nafion (vide infra). (15) Kirchhoff, J. R.; Okamoto, K.; Heineman, W. R.; Deutsch, E. Inorg. Chim. Acta, in press.
Table 1. Voltammetric Parameters of [ReI(DMPE)3]+/ [ReII(DMPE)3]2+ in Aqueous and Aqueous/Ethanol Solutionsa solution (% EtOH)
electrodeb
∆Ep (mV)
E°′ (mV)
ipa/ipc
D (cm2/s)
0 4 4 4 4
GC GC Pt Au PG
67 61 67 60 78
30 36 34 37 39
1.09 1.06 1.03 1.05 1.02
5.0 × 10-6 5.1 × 10-6
a E°′ taken as the midpoint of cathodic and anodic peak potentials vs Ag/AgCl (3 M NaCl); 1.0 mM [ReI(DMPE)3]+ in 0.15 M NaCl; scan rate 100 mV/s. D of [ReI(DMPE)3]+ calculated from ipa ) (2.69 × 105)n3/2ACυ1/2D1/2. b GC, glassy carbon; PG, pyrolytic graphite.
However, this low solubility has also made characterization by cyclic voltammetry difficult because of the low associated Faradaic currrent. In order to partially overcome this experimental handicap, a small amount of ethanol was added to solutions to make higher [ReI(DMPE)3]+ concentrations where improved signal-to-noise measurements could be made. Cyclic voltammetry of the [ReI(DMPE)3]+/[ReII(DMPE)3]2+ redox couple was performed on a bare glassy carbon electrode in solutions with ethanol (at high concentrations of complex) and without ethanol (at low concentrations of complex). Voltammograms obtained in both solutions showed similar reversible electrochemical reactions when the potential was scanned between +500 and -500 mV, although the higher concentrations of complex in the ethanol solutions gave much better defined voltammograms. Some electrochemical parameters obtained from these experiments are listed in Table 1. The separation of anodic/ cathodic peak potentials is close to the theoretical value 59/n mV, which shows a Nernstian reaction involved in both solvent systems. A slight positive shift in E°′ with addition of alcohol is observed. In the case of DMF solution, this couple is shifted to +333 mV vs Ag/AgCl.15 This positive shift is attributed to stabilization of the more hydrophobic [ReI(DMPE)3]+ by the ethanol or DMF. The peak potential was independent of the scan rate, and the peak current was proportional to the square root of scan rate over a scan rate range from 0.010 to 0.500 V/s in both water and water/ethanol solvents for both the anodic and the cathodic peaks. The diffusion coefficients of [ReI(DMPE)3]+ in both solvent systems were calculated from the slopes (from linear regression analysis) of ipa vs υ1/2 plots by the Randles-Sevick equation assuming a one-electron process. The results in Table 1 demonstrate that all of the electrochemical features of [ReI(DMPE)3]+/[ReII(DMPE)3]2+ in alcoholic aqueous solution very closely resemble those in pure aqueous solution and it appears that the addition of a small amount of ethanol does not affect the basic electrochemical behavior of the redox couple. Thus, 4% EtOH was added to those solutions containing 1.0 mM [ReI(DMPE)3]CF3SO3 but not to more dilute solutions. Effect of Electrode Material. In order to determine the effect of electrode materials on the electrochemical behavior of the rhenium complex, cyclic voltammograms of [ReI(DMPE)3]+/[ReII(DMPE)3]2+ were measured on glassy carbon, platinum, gold, and a highly ordered pyrolytic graphite. Well-defined, reversible voltammograms were observed at glassy carbon, platinum, and gold. In the last case, a large anodic background current beyond +500 mV was caused by oxidation of the gold electrode due to
Figure 2. Cyclic voltammograms of (a) 1.0 × 10-3 M [ReI(DMPE)3]+ in 0.15 M NaCl, 4% EtOH and (b) 1.0 × 10-5 and (c) 1.0 × 10-7 M [ReI(DMPE)3]+ in 0.15 M NaCl at bare glassy carbon electrode; scan rate 100 mV/s.
the presence of chloride in the electrolyte. Obvious adsorption of [ReI(DMPE)3]+ occurs on pyrolytic graphite as evidenced by prominent shoulders on the major cathodic and anodic waves. This result is in agreement with that observed at spectroscopic graphite.12 The predominant electrochemical reaction of [ReI(DMPE)3]+/[ReII(DMPE)3]2+ at the surface of platinum and gold was a diffusion-controlled process. Table 1 lists cyclic voltammetric parameters at the four different working electrodes. No significant differences are observed. Detection Limit on Bare Electrodes. The effect of concentrations ranging from 1.0 × 10-7 to 1.0 × 10-3 M [ReI(DMPE)3]+ on voltammograms at a bare glassy carbon electrode is illustrated in Figure 2. The detection limit is about 10-6 M, which is typical for this technique at a bare solid electrode. The detection limit can be improved with electrochemical techniques such as differential pulse voltammetry with which a detection limit of 1 × 10-7 M [ReI(DMPE)3]+ could be reached. However, further improvement in detection limit by electrode modification with a preconcentrating polymer film is necessary for the proposed application of in vivo detection. Electrochemical Behavior of [ReI(DMPE)3]+ at NafionModified Electrodes. In the present work, Nafion film was used to modify electrodes for extraction of the Re complex since this polymer has been extensively studied and has shown remarkable affinity for cationic complexes with hydrophobic ligands.8-11 Partitioning of [ReI(DMPE)3]+ into Nafion Film. The incorporation of [ReI(DMPE)3]+ into a Nafion film-modified glassy carbon electrode can be observed voltammetrically by continuously cycling the potential between +500 and -500 mV vs Ag/ Analytical Chemistry, Vol. 69, No. 19, October 1, 1997
4047
AgCl. The peak current increases as the electroactive species is incorporated into the film. A steady-state voltammogram was obtained after cycling for about 15-100 min, depending on parameters such as film thickness, concentration of [ReI(DMPE)3]+, and transport conditions in solution (i.e., stirred or quiescent). The electrode with the incorporated redox couple could be taken out of the [ReI(DMPE)3]+ solution, rinsed with water, and placed in pure supporting electrolyte solution, and a comparable cyclic voltammogram obtained on the trapped complex. Comparison voltammograms of the complex obtained at the bare glassy carbon electrode and at a Nafion-coated glassy carbon electrode give cyclic voltammograms that are similar in shape. The E°′ was shifted positively by about 60 mV at the Nafionmodified electrode, which indicates that [ReI(DMPE)3]+ has been stabilized by Nafion relative to [ReII(DMPE)3]2+. This is attributed to a strong hydrophobic interaction between [ReI(DMPE)3]+ and the fluorocarbon phase of Nafion (vide infra). The Nafion film also causes the separation of peak potential to increase to 130 mV compared to 61 mV at the bare electrode. Effect of Oxidation State on Retention in Nafion. Retention of the Re complexes in the Nafion film after transfer to pure supporting electrolyte solution was examined under several experimental conditions in order to evaluate the effects of applied potential (i.e., oxidation state of the trapped complex). There was no measurable current loss when a Nafion-modified electrode loaded with [ReI(DMPE)3]+ was transferred to a supporting electrolyte solution and soaked for about 2 h without application of a potential to the electrode before recording a cyclic voltammogram. This strong retention of [ReI(DMPE)3]+ is consistent with the report that Ru(bipy)32+ is also strongly retained such that 70% remains in the film after soaking for several days in supporting electrolyte.9 However, if the electrode potential were continuously cycled after transfer into pure supporting electrolyte, the peak current after 50 cycles (about 20 min) decreased to about 67% of the peak current of the first voltammogram recorded after the transfer. This result suggests that the oxidized form of the complex, [ReII(DMPE)3]2+, is less strongly retained than [ReI(DMPE)3]+. Since a substantial amount of the complex would exist in the dicationic form during continous cycling, poorer retention of this form could have a significant influence on retention, even though the film was initially loaded with the monocationic form. Consequently, the loss of peak current for a loaded electrode was further studied with experiments in which the potential was held at -300 mV where the Re complex was maintained as [ReI(DMPE)3]+ and at +300 and +500 mV where the complex is oxidized to [ReII(DMPE)3]2+. Figure 3 shows the difference in retention at these applied potentials as determined by periodically recording a cyclic voltammogram and measuring the peak height. About 20% signal loss was observed when the negative potential of -300 mV was applied for 2 h to hold the complex in the reduced form. Presumably the 20% loss in comparison to essentially no loss when no potential was applied is due to the [ReII(DMPE)3]2+ formed as a result of the periodic measurement scan. At the positive potentials, the signal loss was substantially greater; 93% of the signal was lost after 45 min, at which time the signal ceased decreasing. It is apparent that the reduced, monocationic form of the complex is retained much more strongly than the oxidized, dicationic form. This result reveals a lot about the mechanism of retention of the complex by Nafion. The dicationic form has the 4048 Analytical Chemistry, Vol. 69, No. 19, October 1, 1997
Figure 3. Effect of potential on retention of [ReI(DMPE)3]+ and [ReII(DMPE)3]2+ in Nafion-modified glassy carbon electrode: [ReI(DMPE)3]+, Eholding ) -300 mV; [ReII(DMPE)3]2+, Eholding ) +300 for 10 min and then switched to + 500 mV; 0.05 M NaCl as supporting electrolyte.
greater charge density and hence would have the greater electrostatic interaction with negatively charged SO3- groups. The monocationic form is the more hydrophobic and would therefore be expected to have the stronger hydrophobic interaction with the fluorocarbon phase in the Nafion film. Retention in Nafion involves a cooperative effect of ion exchange and hydrophobic interactions. It is apparent from these results that the main effect is hydrophobicity, since the more hydrophobic monocation is retained more strongly than the dication with greater charge density. This result is consistent with the positive shift in E°′ with the Nafion film, which indicates that the film renders the reduced form more difficult to oxidize. A similar effect has been reported for [RuII(bipy)3]2+, which is somewhat hydrophobic and is retained more strongly than its oxidized form [RuIII(bipy)3]3+ with the greater charge density.9 However, with hydrophilic ions such as [RuII(NH3)6]2+/[RuIII(NH3)6]3+ the ion with the greater charge density, [RuIII(NH3)6]3+, partitions more strongly into Nafion, as expected for an interaction that is strictly ion exchange.11 The plots shown in Figure 3 are analogous to the “unloading curves” described by Porat et al. for multicharged viologen derivatives.16 These viologens provide an interesting example of the balance between the hydrophobicity/hydrophilicity of compounds and their interactions with Nafion. Partitioning Rate. The partitioning of [ReI(DMPE)3]+ into the Nafion film was measured by immersing a Nafion-modified electrode into a stirred solution of [ReI(DMPE)3]+ and measuring the peak current from cyclic voltammetry. This was done for two [ReI(DMPE)3]+ solutions, 1.0 × 10-5 and 1.0 × 10-8 M, in 0.05 M NaCl. Figure 4a shows the results of normalized peak current vs immersion time for an electrode immersed in 1.0 × 10-5 M [ReI(DMPE)3]+. Maximum uptake occurred at about 45 min as evidenced by the plateauing of the current plot. Similar behavior occurred for a more dilute solution, 1.0 × 10-8 M, although a longer time was necessary to reach equilibrium (curve b). Effect of Potential Range. The sweep potential window should be chosen carefully when electrochemical detection of [ReI(DMPE)3]+ is performed. A positive scan of a Nafion-modified electrode loaded with [ReI(DMPE)3]+ initiated at -1000 mV shows both oxidation waves corresponding to [ReI(DMPE)3]+/[ReII(DMPE)3]2+ and [ReII(DMPE)3]2+/[ReIII(DMPE)3]3+, as shown in (16) Porat, Z.; Tricot, Y.-M.; Rubinstein, I.; Zinger, B. J. Electroanal. Chem. 1991, 315, 217-223.
Figure 4. Current reponse as a function of extraction time: (a) 1.0 × 10-5 M [ReI(DMPE)3]+; (b) 1.0 × 10-8 M [ReI(DMPE)3]+; 0.05 M NaCl as supporting electrolyte; solution stirred; scan rate 100 mV/s.
Figure 1 for a bare electrode. However, a substantially smaller reduction wave for [ReIII(DMPE)3]3+ and no reduction wave for [ReII(DMPE)3]2+ were found on the reverse (negative) scan. Voltammograms obtained within the potential window +700 to -1000 mV in 1.0 mM [ReI(DMPE)3]+ solution after the electrode had been cycled between +1200 and -1000 mV in the same solution exhibit no redox waves whatsoever for the [ReI(DMPE)3]+/ [ReII(DMPE)3]2+ couple. We speculate that this loss of redox activity is related to instability of the ReIII oxidation state, as has been observed for DMF.15 In DMF, ligand substitution by solvent occurs when ReIII is electrogenerated. The similarity in the voltammograms obtained in aqueous 4% EtOH (Figure 1) and in DMF suggests that the same solvent substitution is occurring in water. However, when ReIII is generated within the Nafion film, SO3- groups from the the Nafion could be functioning as coordinating ligands to form a new Re-DMPE-sulfonate complex that is electroinactive. Another possible, and in our opinion more likely, explanation involves the hydrophobicity of the DMPE ligands. When the labile [ReIII(DMPE)3]3+ complex is electrogenerated, the hydrophobic DMPE ligands are stripped from the complex by strong interaction with the fluorocarbon phase of Nafion. Loss of the stabilizing DMPE ligands results in reaction of ReIII with water to form ReO2, which precipitates within the film and would be electroinactive. This latter explanation is consistent with the known inorganic chemistry of Re and Tc coordination compounds in which removal of ligands from the metal ions in their lower oxidation states invariably results in a fine precipitate of the metal oxide. Whatever the reason, it is clear that cycling the potential into the [ReII(DMPE)3]2+ oxidation wave has a disasterous effect on the sensor. The positive potential must be limited to +700 mV or less for the uptake of [ReI(DMPE)3]+ at the Nafion-modified electrode to be consistently observed. Reproducibility. In order to evaluate the reproducibility of Nafion-modified electrodes, the standard deviation of oxidation and reduction peak currents in cyclic voltammograms (-700 to +700 mV potential window, scan rate 100 mV/s) from three individual electrodes was measured. The electrodes coated with 600 µm Nafion films were prepared by the same procedure and soaked in a 1.0 × 10-5 M [ReI(DMPE)3]+ solution for 1 h prior to measurement. The relative standard deviation for the peak
Figure 5. Cyclic voltammograms of (a) 1.0 × 10-5, (b) 1.0 × 10-7, and (c) 1.0 × 10-8 M [ReI(DMPE)3]+ in 0.05 M NaCl at Nafionmodified glassy carbon electrode after extraction time of 60 min; scan rate 100 mV/s.
currents was 0.2% for the anodic peak and 1.2% for the cathodic peak. Analytical Detection Limits and Calibration Curves. Figure 5 shows cyclic voltammograms for Nafion-modified electrodes after immersion in three solutions of different [ReI(DMPE)3]+ concentrations for 60 min. A well-defined cyclic voltammogram was obtained even for the lowest concentration. Comparison of these results with those shown in Figure 2 for a bare electrode shows the detection limit from cyclic voltammetry to be about 2-3 orders of magnitude lower at the Nafion-modified electrode than at the bare electrode. Even lower detection limits were reached when differential pulse voltammetry was employed at Nafion-coated glassy carbon electrodes. Figure 6 shows differential pulse voltammograms at Nafion-modified glassy carbon for the 10-9 M concentration range. The calibration curve over this low range is linear. Preconcentration Factor. The approximately 2-3 orders of magnitude decrease in detection limit for cyclic voltammetry and differential pulse voltammetry at the Nafion-modified electrode compared to a bare electrode points to a significant partitioning of [ReI(DMPE)3]+ into Nafion. The extent of partitioning into Nafion has been described by the preconcentration factor, or distribution coefficient.11 The preconcentration factor is taken as the ratio of the concentration of [ReI(DMPE)3]+ in the Nafion film to that in solution. Figure 7 shows the partition isotherm for the solution concentration range of 10-8 M in 0.05 M NaCl. The slope of the plot, or the preconcentration factor, is 1 × 106. This large value is comparable to those obtained by Martin et al. for similar hydrophobic cations such as (ferrocenylmethyl)trimethylammonium hexafluorophosphate (1.1 × 106) and [RuII(bipy)3]2+ (2.1 × 107).11 Analytical Chemistry, Vol. 69, No. 19, October 1, 1997
4049
Figure 6. Differential pulse voltammograms of [ReI(DMPE)3]+ in 0.05 M NaCl at Nafion-modified glassy carbon electrode in 10-9 M concentration range after extraction time of 60 min.
It is interesting to note that the preconcentration factor is about 1 millionfold, whereas the detection limit improvement is only about 1 thousandfold. This difference is made up for by the approximate thousandfold decrease in diffusion coefficient that hydrophobic cations undergo in Nafion.9,17 Thus, in a voltammetric detector, a substantial amount of the concentration enhancement that the Nafion provides is lost because of the role that diffusion plays in defining the current output. However, this would not be the case for an optical sensor, where even greater improvements in detection limit should be realized. CONCLUSIONS These results indicate that the Nafion film-coated electrode would be a good candidate for the electroanalytical determination of [ReI(DMPE)3]+, and perhaps other heart imaging agents that are electroactive hydrophobic cations, down to concentrations as low as the 10-9 M range. The successful extension of this sensor
Figure 7. Partition isotherm for [ReI(DMPE)3]+ at Nafion-modified glassy carbon electrode in 0.05 M NaCl.
to in vivo measurements in biological media depends on retention of the large preconcentration factor observed in saline solution and adequate selectivity for that more complex environment. ACKNOWLEDGMENT The authors thank Mr. Huailiang Chen for assistance in experiments and Dr. Elmo Blubaugh for helpful discussions during this work. Financial support provided by DOE Grant 86ER60487 and NIH Grant CA42179 is gratefully acknowledged.
Received for review December 30, 1996. Accepted June 27, 1997.X AC961295V
(17) Porat, Z.; Rubinstein, I.; Zinger, B. J. Electrochem. Soc. 1993, 140, 25012507.
4050 Analytical Chemistry, Vol. 69, No. 19, October 1, 1997
X
Abstract published in Advance ACS Abstracts, August 15, 1997.
