Effect of Ca(II), Sr(II), and Ba(II) on the Pulsed Amperometric Detection

The effects of some divalent nonelectroactive cations (DNCs), such as Sr(II), Ba(II), and Ca(II) on the electrochemical oxidation of alditols and carb...
0 downloads 0 Views 306KB Size
Download PDF
Anal. Chem. 1997, 69, 4849-4855

Effect of Ca(II), Sr(II), and Ba(II) on the Pulsed Amperometric Detection of Alditols and Carbohydrates at a Gold Electrode in Alkaline Solutions Tommaso R. I. Cataldi,* Innocenzo G. Casella, and Diego Centonze

Dipartimento di Chimica, Universita` degli Studi della Basilicata, Via N. Sauro, 85, 85100 Potenza, Italy

The effects of some divalent nonelectroactive cations (DNCs), such as Sr(II), Ba(II), and Ca(II) on the electrochemical oxidation of alditols and carbohydrates at gold electrodes in pulsed amperometric detection have been investigated. It seems that in the presence of DNCs in alkaline solutions, two competitive processes are involved: polyhydroxy compound complexation in the following order of metal ion binding affinity, Ca(II) > Sr(II) > Ba(II), and inhibition on the onset of gold oxide formation in the same order. This last effect leads to an increased activity of the electrode surface; at the optimized value of detection potential of D-sorbitol, EDET ) +50 mV vs Ag|AgCl, which is ∼100 mV lower in comparison to the maximum value observed with blank carrier electrolytes (i.e., 0.58 M NaOH), there is an increase in sensitivity of ∼50%, and 30% in the presence of 1.0 mM Sr(II), and 1.0 mM Ba(II), respectively. However, the current response of sample molecules results increased only when Ba(II) or Sr(II) ions were already contained in the alkaline media, that is the experimental condition normally occurring in flow injection and liquid chromatography systems. The voltammetric response, observed upon additions of Ba(II) or Sr(II) to an alkaline electrolyte containing D-sorbitol, showed progressive decrease of the anodic current. Irrespective of the experimental condition adopted, i.e., cation addition to the solution before or after the carbohydrate, the presence of Ca(II) has an adverse effect on the anodic currents. These findings have been explained by a rapid formation of adducts between polyhydroxy compounds and DNCs in sodium hydroxide solutions. Since the introduction of pulsed amperometric detection (PAD) in the early 1980s,1 this detection technique has been increasingly applied to a variety of analytical problems.2-7 The most successful working electrode in basic solutions is gold, which was employed initially for the detection of carbohydrates2-4 but subsequently * Corresponding author: (Fax) 39-971-474223; (E-mail) [email protected]. (1) Hughes, S.; Johnson, D. C. Anal. Chim. Acta 1981, 132, 11. (2) Rocklin, R. D.; Pohl, C. A. J. Liq. Chromatogr. 1983, 6, 1577. (3) Johnson, D. C.; LaCourse, W. R. Anal. Chem. 1990, 62, 589A. (4) Lee, Y. C. Anal. Biochem. 1990, 189, 151. (5) Johnson, D. C.; LaCourse, W. R. Electroanalysis 1992, 4, 367. (6) Johnson, D. C.; Dobberpuhl, D. A.; Roberts, R. A.; Vandeberg, P. J. J. Chromatogr. 1993, 640, 79. (7) Welch, L. E.; LaCourse, W. R.; Mead, D. A., Jr.; Johnson, D. C.; Hu, T. Anal. Chem. 1989, 61, 555. (8) LaCourse, W. R.; Johnson, D. C. Carbohydr. Res. 1991, 215, 159. S0003-2700(97)00471-X CCC: $14.00

© 1997 American Chemical Society

was shown to allow the oxidation of glycols, alditols, n-alkanolamines,8-11 amino acids,9 and sulfur-containing compounds as well.12 The electrooxidation of sugars and alditols is controlled primarily by the dependence of the catalytic surface state on the electrode potential and not on the redox potentials of the compounds themselves. Further, it is recognized that polyhydroxy compounds are oxidized on oxide-free Au surfaces because the presence of an oxide film inhibits their oxidation. Indeed, the success of PAD for anodic detection of these compounds results from the periodic reactivation of the electrode surface, within the programmed potential wave form, by the application of positive and negative potential pulses.3 The negative pulses are able to reduce the inert surface gold oxide, formed during the positive potential excursion, back to the pristine metal, followed by a step back to the detection potential. Despite the great interest and significance of PAD at gold electrodes, no report of sodium hydroxide solutions containing alkaline-earth metal ions has hitherto appeared in the literature. There exist only a few examples of diluted Ba(OH)2 solutions used as mobile phases in anion-exchange chromatographic separations of carbohydrates.13,14 Recently, Gartske and Huber15 reported an anomalous behavior of Ca(II) in 0.1 M NaOH solutions using a nickel oxide working electrode under constant applied potential. The observation was that calcium ion strongly inhibited the anodic current of glucose and other sample molecules in conventional constant-potential amperometric detection in flowing streams, whereas the effect became positive when a no-convective detection scheme was employed. We have demonstrated that sodium hydroxide mobile phases containing millimolar concentrations of some divalent nonelectroactive cations (DNCs) such as Ca(II), Ba(II), or Sr(II) provide remarkable improvements in terms of separation and column efficiency in anion-exchange chromatography with integrated PAD (IPAD).16,17 The use especially of Ba(II) or Sr(II) ions offers an improved approach for the determination of alditols and carbohydrates in real samples even when a Cu2O-carbon composite (9) Jackson, W. A.; LaCourse, W. R.; Dobberpuhl, D. A.; Johnson, D. C. Electroanalysis 1991, 3, 607. (10) LaCourse, W. R.; Jackson, W. A.; Johnson, D. C. Anal. Chem. 1991, 63, 134. (11) Dobberpuhl, D. A.; Johnson, D. C. Electroanalysis 1996, 8-9, 726. (12) Polta, T. Z.; Johnson, D. C. J. Electroanal. Chem. 1986, 209, 159. (13) Johnson, D. C. Nature (London) 1986, 321, 451. (14) Johnson, D. C.; Polta, T. Z. Chromatogr. Forum 1986, 1, 37. (15) Gartske, C.; Huber, C. O. Anal. Chim. Acta 1995, 300, 53. (16) Cataldi, T. R. I.; Centonze, D.; Margiotta, G. Anal. Chem. 1997, 69, 4842. (17) Cataldi, T. R. I.; Margiotta, G.; Zambonin, C. G. Food Chem., in press.

Analytical Chemistry, Vol. 69, No. 23, December 1, 1997 4849

electrode was used as an amperometric sensor.18 Such beneficial effects have been explained as due to sugar-metal ion complexes, preferentially formed with those sample molecules that have a favorable steric arrangement of three hydroxyl groups on consecutive carbon atoms.19-21 The attainment of a significant increase in the current integrated signal of chromatographic peaks was also emphasized. However, whereas both Ba(II) and Sr(II) ions exhibited enhancement of the response for all compounds examined, the presence of Ca(II) inhibited the amperometric response of some analytes. The present study was undertaken to investigate the effects of Ca(II), Ba(II), and Sr(II) on the gold electrode activity in alkaline media. Cyclic voltammetry at a gold rotating disk electrode and hydrodynamic voltammetric measurements in flowing streams were employed for the electrochemical investigation of sugars and alditols. Sodium hydroxide electrolytes with and without the presence of DNCs were used with D-sorbitol, D-ribose, D-glucose, D-fructose, and sucrose as model compounds. This study provides the first evidence for inhibition of gold oxide formation by some nonelectroactive cations and additionally confirms that sugar-metal ion complexation occurs even in alkaline solutions. These latter findings have also served to complement the chromatographic results mentioned earlier.16 EXPERIMENTAL SECTION Chemicals. Sodium hydroxide, 50% (w/w) solution in water (d ) 1.515 g/mL), Ca(CH3COO)2 monohydrate 99+%, Sr(CH3COO)2 ∼3% H2O, Sr(NO3)2, Sr(OH)2 octahydrate 96%, Ba(CH3COO)2 99%, D-fructose, and D-ribose were purchased from Aldrich (Milan, Italy); myo-inositol, D-sorbitol, D-mannitol, D-glucose, and sucrose were from Sigma (Milan, Italy) and were used as received. Other chemicals employed were of analytical grade quality and were used without further purification. Stock solutions of sugars and alditols were prepared fresh daily in pure water. All solutions were prepared with double-distilled and deionized water using a deionization system (Cecchinato, Venice Mestre, Italy). Sodium hydroxide solutions of the desired concentration were prepared by diluting the appropriate volume of carbonate-free 50% (w/w) NaOH. Such alkaline solutions were used as the supporting electrolyte as well as the carrier electrolyte for the flow injection experiments. The exact concentration of the hydroxide ion in the mobile phase was determined by titration against a standard solution of hydrochloric acid using phenolphthalein as an indicator in the neutralization reaction. Voltammetric Apparatus. Voltammetric data were obtained at a rotating disk electrode (RDE) using a Model EDI101 rotator from Radiometer Analytical (Copenhagen, Denmark) and a Model 263A potentiostat (EG&G Princeton Applied Research, Princeton, NJ). Data acquisition and potentiostat control were accomplished with a 486/50 MHz IBM-compatible computer running the M270 electrochemical research software (EG&G) version 4.11. Electrochemical experiments were made using a conventional singlecompartment glass cell (PAR). A gold disk electrode (2.0 mm diameter) Model EM-EDI-Au-D2 (Radiometer Analytical) was used for all voltammetric RDE investigations unless otherwise noted. For these experiments, a Pt auxiliary electrode and a Ag|AgCl (saturated KCl) reference electrode were used. The gold working (18) Cataldi, T. R. I.; Centonze, D.; Casella I. G.; Desimoni, E. J. Chromatogr., A 1997, 773, 115. (19) Rendleman, J. A., Jr. Adv. Carbohydr. Chem. 1966, 21, 209. (20) Goulding, R. W. J. Chromatogr. 1975, 103, 229. (21) Angyal, S. J. Adv. Carbohydr. Chem. Biochem. 1989, 47, 1.

4850

Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

electrode was polished with 0.05 µm alumina oxide powder on microcloth using water as the lubrificant prior to each series of measurements. For direct comparison of voltammetric results, a well-conditioned gold electrode was used. The current-potential (i-E) curves were the stable profiles recorded after three potential cycles. In this paper, anodic currents are considered to be positive. High-purity N2 gas was bubbled through the cell to remove dissolved O2 prior to the scan sweep, and an atmosphere of N2 was maintained over the test solutions during experimentation. The composition of the various alkaline solutions containing Ba(II), Sr(II), or Ca(II) ions is given in the text or the figure captions. Unless otherwise specified, all experiments were carried out at room temperature, ∼20°C. Flow Injection Apparatus. Flow injection amperometric detection was performed with a Dionex system (Sunnyvale, CA), which consisted of an inert pump, Model IP20, a pulsed amperometric detector, Model ED40, and a Dionex metal-free rotary injection valve with 10 µL injection loop. The flow-through detection cell (Dionex) contained a Au working electrode (1.0mm diameter) and a Ag|AgCl reference electrode. The counter electrode was provided by the titanium upper half of the detection cell. A JJ Instruments X-t chart recorder, Model CR650S, recorded the output signal. As detection mode, we adopted a modified version of pulsed amperometry that involves the digital integration of current, IPAD. The oxidation current is measured and integrated after a delay (tDEL) that allows the double-layer charging current to decay. The oxidation current integrated with respect to time gives a net charge (q) for the detection cycle, so the response is measured in coulombs. Unless otherwise stated, the recommended pulse sequence for IPAD of alditols and carbohydrates in strongly alkaline solutions (e.g., 0.58 M NaOH) modified with DNCs were as follows: oxidative cleaning, EOX ) +650 mV (tOX ) 190 ms), EDET ) +50 mV (delay time before charge integration, tDEL ) 150 ms, and tINT ) 300 ms); and reductive reactivation, ERED ) -150 mV (tRED ) 340 ms). RESULTS AND DISCUSSION Electrochemical Oxidation of D-Sorbitol at a Au-RDE. (1) Effect of Ba(II) and Sr(II) Ions. Current-potential (i-E) curves at a gold rotating disk electrode (Au-RDE) were initially run to evaluate the electrochemical oxidation of D-sorbitol in supporting electrolytes containing Ba(II) or Sr(II) ions. Such cations were added in NaOH solutions as hydroxides or acetate salts. The experiments were first accomplished in 0.58 M NaOH by sweeping the potential of the working electrode between -800 and +650 mV vs Ag/AgCl at a scan rate of 50 mV s-1. The typical i-E behavior of a gold electrode at 900 rpm is illustrated by the voltammogram shown in Figure 1A (dashed line); the solution was bubbled with N2 gas to remove dissolved O2 prior to the scan sweep. The main electrochemical characteristics of gold electrodes in basic media are well known.22-26 During the positive potential scan in the absence of dissolved O2, the major redox process includes a large anodic peak at E > +0.2 V, corresponding to the formation of a passivating film surface oxide, AuO, and a cathodic peak at about +0.06 V, corresponding to the re-formation (22) Kirk, D. W.; Foulkes, F. R.; Graydon, W. F. J. Electrochem. Soc. 1980, 127, 1069. (23) Vitt, J. E.; Larew, L. A., Johnson, D. C. Electroanalysis 1990, 2, 21. (24) Burke, L. D.; Cunnane, V. J. J. Electrochem. Soc. 1986, 133, 1657. (25) Burke, L. D.; Cunnane, V. J. J. Electroanal. Chem. 1986, 210, 69. (26) Burke, L. D.; O’Sullivan, J. F. Electrochim. Acta 1992, 37, 585.

Figure 2. Voltammetric responses of 1.0 mM D-sorbitol at a AuRDE in 0.58 M NaOH (dashed line, a) and upon addition in the supporting electrolyte of (b) 0.5 , (c) 1.0, (d) 1.5, and (e) 2.0 mM Ba(II). Other conditions as in Figure 1.

Figure 1. (A) Voltammograms at a Au-RDE in 0.58 M NaOH (dashed curve) and 0.58 M NaOH + 1.0 mM Ba(CH3COO)2 (solid curve). (B) Voltammograms of 0.5 mM D-sorbitol in the presence and absence of Ba(II) ions, solid and dashed lines, respectively. Conditions: degassed solutions; 900 rpm (15 Hz) as rotation speed; 50 mV s-1 as scan rate.

of a clean surface of Au0. This behavior is essentially the same as that reported by LaCourse and Johnson in 0.10 M NaOH.27 Upon addition of Ba(II) ions (1.0 mM) to the NaOH solution, two aspects of the background scan, solid line in Figure 1A, are noteworthy. First the anodic peak at ∼+312 mV in the blank supporting electrolyte, which is due to gold oxide formation, is slight shifted at about ˜+336 mV and attenuated in the presence of Ba(II) ion. The second feature is the shift of the reduction peak of AuO, during the reverse scan, toward less positive potentials with a difference in peak potentials of ∼22 mV. This indicates that the extent of oxide formation on the forward scan is decreased in Ba(II)-containing NaOH supporting electrolytes. Notice also a slight lowering of the AuO reduction peak, albeit the total peak area, which is proportional to the charge involved for reduction, remains almost unchanged. The oxidation mechanism of simple carbohydrates at Au electrodes has been discussed extensively5,26-29 and is concluded to involve weak association (adsorption) of the reacting molecules with the electrode surface. It is believed that adsorbed hydroxyl species, AuOHads,22,23 or hydrous oxide, AuOH, 24-26 formed on the gold surface have catalytic properties for oxidation reactions of polyhydroxy compounds including sugars and alditols, which involve transfer of oxygen from water to the oxidation products. If other surface-active species adsorb more strongly, thereby interfering in the oxidation mechanism, there is a severe attenuation of the faradic process.5 Figure 1B compares the voltammetric responses of 0.5 mM D-sorbitol at the Au-RDE in the presence (solid curve) and absence (dashed curve) of 1.0 mM Ba(II) in solution. Most significant is the observation that, upon the addition of D-sorbitol in the Ba(II)-containing supporting (27) LaCourse, W. R.; Johnson, D. C. Anal. Chem. 1993, 65, 50. (28) Larew, L. A., Johnson, D. C. J. Electroanal. Chem. 1989, 262, 167. (29) Neuburger, G. G.; Johnson, D. C. Anal. Chem. 1987, 59, 150.

electrolyte, the peak current on the forward scan increased by more than 25%, suggesting the possible involvement in the catalytic oxidation process of the divalent ion may be as barium hydroxide species or as complexed with D-sorbitol. Additionally, whereas the oxidation current of D-sorbitol in the blank alkaline solution steadily drops as gold oxidation continues to increase, indicating that the oxide-covered Au surface inhibits oxidation of the sample molecule, in the presence of Ba(II), a further increase in current occurs superimposed on the potential wave corresponding to AuO formation. Indeed, during the forward scan, a second wave form with a peak potential at ˜+400 mV appears. With barium ion in solution, the peak potential for oxidation of D-sorbitol is shifted to less positive values, during either forward or reverse scans, further evidence that the gold electrode has become more actived. The observation that sugars and alditols are oxidized on a reduced Au surface and that their oxidation is inhibited by the formation of gold oxide provides a rational basis for interpreting the above results. Perhaps the inhibition effect on the onset of formation of gold oxide, when Ba(II) is present in solution, leads to an increased activity of the electrode surface. Moreover, the anionic oxidation product in basic media is probably more rapidly desorbed in the presence of Ba(II), allowing the adsorption site of the gold surface to be recycled in the oxidation mechanism. It is worthwhile mentioning that the experimental results illustrated above in Figure 1B were successfully obtained upon the addition of D-sorbitol to the alkaline supporting electrolyte containing the divalent cation. Surprisingly, when the sequence was reversed and the i-E curves for D-sorbitol were recorded before and after the addition of Ba(II) in solution, a different behavior was observed (see Figure 2). The addition of barium ions (0.5, 1.0, 1.5, and 2.0 mM) to a sodium hydroxide solution containing 1.0 mM D-sorbitol decreased progressively the anodic response at the gold electrode, curves b-e, respectively. A similar phenomenon has been reported for the voltammetric response of D-sorbitol in the presence of 0.1 mM propylamine in 0.1 M NaOH, where the probable cause has been suggested to be the competitive adsorption of propylamine on the gold electrode surface with the consequent attenuation of the anodic signal.5 Yet, on the basis of the earlier results, we suggest here that the lowering of response might be related to a rapid formation of adducts between D-sorbitol and free Ba(II) ions in solution. The Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

4851

Figure 3. (A) Voltammograms at a Au-RDE in 0.58 M NaOH (dashed curve) and 0.58 M NaOH + 1.0 mM Sr(CH3COO)2 (solid curve). (B) Voltammograms of 0.5 mM D-sorbitol in the presence and absence of Sr(II) ions, solid and dashed lines, respectively. Other conditions as in Figure 1.

overall effect is that barium species are no longer available for modifying the electrode surface, most likely because D-sorbitol complexation in solution is more rapidly established. Such a tentative explanation agrees with the chromatographic findings reported previously in which the formation of sugar-metal ion complexes was also invoked.16 In addition, such sugar-Ba(II) complexes probably hinder the availability of oxidation sites in the sample molecule, thus reducing the anodic signal. Virtually the same results were obtained when strontium acetate was used to test the effect on the oxidation of D-sorbitol at a gold RDE (see Figure 3). The residual i-E curve (panel A) on the positive scan for the supporting electrolyte containing 1.0 mM Sr(II) exhibits a positive shift (∼30 mV) of the anodic wave in the gold oxide formation region. The behavior is similar to that observed in the presence of Ba(II) and the same considerations are valid, though the catalytic oxidation during the positive scan ceases with the concurrent formation of surface oxide. The above observations are consistent with the conclusion that both strontium and barium ions or their corresponding hydroxides play a role in the electrochemical oxidation process of D-sorbitol, either inhibiting the onset of gold oxide formation or enhancing the interaction with the anionic form of D-sorbitol. This last species might be expected to have a higher affinity for the electrode surface in the presence of these alkaline-earth metal ions. (2) Effect of Ca(II). In sharp contrast to the effect of Ba(II) and Sr(II) ions, inhibition rather than enhancement was observed with Ca(II) ion in sodium hydroxide solutions. Considering that in high basic media the solubility of calcium ion, as well as that of other divalent cations, is limited by formation of barely soluble metal hydroxides, more diluted solutions of Ca(II) (e.g., 0.2 mM) in 0.58 M NaOH were employed. The solubility product (pKs)30 in water of Ca(OH)2 is 4.2, which corresponds to ∼0.18 mM Ca(30) The Handbook of Chemistry and Physics, 67th ed., Weast, R. C., Ed.; Chemical Rubber Co.: Cleveland, 1986.

4852 Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

Figure 4. (A) Voltammograms at a Au-RDE in 0.58 M NaOH (dashed curve) and 0.58 M NaOH + 0.2 mM Ca(CH3COO)2 (solid curve). (B) Voltammograms of 0.5 mM D-sorbitol in the presence and absence of Ca(II) ions, solid and dashed lines, respectively. Other conditions as in Figure 1.

(II) ion in 0.6 M NaOH. However, the concentration of the calcium acetate salt was normally used in slight excess (i.e.,, up to 0.5 mM) of the calcium hydroxide solubility value (see below) in order to ensure a sink of Ca(II) ions in solution. Figure 4 (panel A) shows the residual i-E curves, obtained in the absence and presence of Ca(II) ions, dashed and solid lines, respectively. As can be seen, despite that the behavior of Ca(II) in sodium hydroxide solutions at the Au-RDE is more pronounced than that of Sr(II) and Ba(II) ions, the decrease in the catalytic oxidation wave of D-sorbitol is quite dramatic in this case (panel B). Indeed, the onset of gold oxide appears more shifted in the presence of calcium ions, but the overall effect is opposite in terms of anodic current response. Consequently, on the basis of these experimental data, the effectiveness with which the DNCs inhibit the onset of gold oxide formation decreases in the order Ca(II) > Sr(II) > Ba(II). These findings further support the suggestion that metal ion complexation occurs in alkaline solutions, following the same order of binding affinity reported in aqueous solutions by Angyal,21 Ca(II) > Sr(II) > Ba(II). At this time, however, it is not clear whether the decrease of anodic signal is caused by strong adsorption of insoluble Ca(OH)2 onto the electrode surface with consequent blockage of the catalytic sites, stronger complexing capability of Ca(II) compared to other DNCs,20,21 or both. Electrochemical Oxidation of D-Ribose at a Au-RDE. Effect of Ba(II), Sr(II), and Ca(II) Ions. In view of the results shown above for D-sorbitol, the voltammetric behavior of D-ribose in the same experimental conditions was also investigated. Such a reducing sugar was chosen because of its reported behavior in anion-exchange chromatographic runs using alkaline mobile phases containing DNCs:16 improved peak symmetry, higher number of column plates, and increased peak (area) intensity. Figure 5 shows the current-potential curves of 0.5 mM D-ribose in 0.60 M NaOH in the absence of added ions (curve a, dashed line) and in the presence in the supporting electrolytes of 1.0 mM Ba(II), 1.0 mM Sr(II), or 0.5 mM Ca(II), curves b-d, respectively.

Figure 5. Voltammetric responses at a Au-RDE of 0.5 mM D-ribose in 0.60 M NaOH (dashed line, a) and 0.60 M NaOH + 1 mM Ba(II) (b), 0.60 M NaOH + 1.0 mM Sr(II) (c), and 0.60 M NaOH + 0.5 mM Ca(II) (d). Other experimental conditions as in Figure 1.

Similarly to other reducing sugars, during the positive scan, D-ribose exhibits an oxidation wave beginning at ∼-600 mV that corresponds to oxidation of the aldehyde functionality. A much larger anodic signal results from the combined oxidations of the alcohol and aldehyde groups in the region between -0.2 and +0.2 V (curve a). As with D-sorbitol, the anodic signal is sharply inhibited by the concurrent formation of surface oxide for E > +0.2 V. All these features closely resemble those reported for D-glucose.8,28,31 As indicated in Figure 5, the presence of Sr(II) or Ba(II) but not Ca(II) induces an increase of response. Our examination of the current-potential curves includes some other sample molecules such as D-mannitol, myo-inositol, D-glucose, and D-fructose. Of more importance is the observation that on the reverse scan (negative) the oxidative current peak, which is located at the same potential for all three cations, is slightly greater than the current observed at the same potential for the positive scan. Apparently, such a phenomenon is more pronounced in the presence of Ba(II) ions. These data also support the hypothesis that upon stripping of the gold oxide in the presence of DNCs, surface sites made free of oxide are immediately active for anodic oxidation of the sample molecule. To interpret the remarkable difference in the calcium effect and that of other divalent ions, the configuration of D-ribose and its ability to orient the hydroxyl groups toward the metal ion are of major significance.20 Flow Injections with Integrated Amperometry. According to LaCourse and Johnson,27 the optimal wave form in 0.1 M NaOH solutions should consist of an oxidation potential (EOX) in the range from +600 to +1000 mV vs Ag/AgCl with recommended values equal to +800 mV and 180 ms, as EOX and oxidation time (tOX), respectively. The positive potential pulse has to be chosen to give maximum oxide coverage, which correlates directly with cleaning of the electrode surface. Clearly, in stronger alkaline conditions, like those we employed, an EOX ) +650 mV and tOX ) 190 ms were considered suitable for the efficient oxidative cleaning of the electrode surface. Albeit higher EOX could be used, the choice of +650 mV avoids problems related with the anodic evolution of O2 bubbles on the electrode surface. The values of (31) LaCourse, W. R.; Mead, D. A., Jr.; Johnson, D. C. Anal. Chem. 1990, 62, 220.

Figure 6. Flow injection peaks of 0.25 mM D-sorbitol as a function of Ba(II) concentration at a gold working electrode with IPAD using 0.59 M NaOH as the carrier electrolyte. Peaks refer to carrier electrolytes containing Ba(CH3COO)2 at the following concentrations: (a) 0.0, (b) 0.5, and (c) 1.0 mM. Flow rate, 0.5 mL min-1.

reductive potential (ERED) and time (tRED) have to be chosen in order to guarantee the complete reductive dissolution of the surface oxide, and the value of ERED is in the range -800 to +100 mV vs Ag/AgCl in 0.1 M NaOH. A further important requirement is that the ERED should be negative enough to minimize oxidation of analytes, especially carbohydrates, and avoid at the same time a massive oxygen reduction with H2O2 formation, prior to application of EDET, and a reduction time large enough to ensure the adsorption of analytes. Thus, the values ERED ) -150 mV with tRED ) 340 ms were normally employed using 0.5-0.6 M NaOH solutions. The triple-step IPAD wave form described in the Experimental Section was considered as optimal, as confirmed by the rapid establishment of baseline stability accompanied by constant analytical signals on repetitive injections. Subsequent analytical work was mostly carried out using such potential and time settings. Before examining the effect of EDET on the IPAD response in the presence of DNCs, we illustrate the action of barium ion in the carrier electrolyte. Using a flow-through thin-layer cell with a Au working electrode, the electrochemical oxidation of D-sorbitol in an alkaline carrier electrolyte was studied as a function of Ba(II) concentration. As reported in Figure 6, it was experimentally verified by systematically changing the concentration of Ba(II) from 0.0 (a) to 0.5 (b), and 1.0 mM (c), and vice versa in 0.58 M NaOH as carrier electrolytes that the specific presence of such cation determines an enhancement of response in integrated pulsed amperometry as its concentration increased. Well-defined peaks were always observed using a detection potential of +0.05 V. Interestingly, the contribution of surface roughness at the Au electrode can be ruled out as responsible for the increase in the anodic signal intensity. Analogous results were obtained when strontium ion was used (data not shown). Effect of Detection Potential. To validate the voltammetric results with the Au-RDE, the response of several polyhydroxy compounds in 0.58 M NaOH carrier electrolyte containing Ba(II), Sr(II), or Ca(II) was monitored in flowing streams with IPAD. Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

4853

Figure 7. Effect of DNCs on the detection potential of 0.25 mM D-sorbitol in flow injection using 0.58 M NaOH as the carrier electrolyte containing 1.0 mM Ba(CH3COO)2, 1.0 mM Sr(CH3COO)2, or 0.5 mM Ca(CH3COO)2. Dashed line represents the response obtained in blank 0.58 M NaOH as a carrier electrolyte. Flow rate, 0.5 mL min-1; loop, 10 µL.

Figure 8. Effect of Ba(II), Sr(II), and Ca(II) on the detection potential of 0.25 mM D-glucose in flow injection with IPAD. Other conditions as in Figure 7.

The resulting hydrodynamic voltammograms (HDVs) obtained for D-sorbitol, D-glucose, and sucrose under flow injection conditions are shown in Figures 7-9, respectively. Each compound was injected and detected in consecutive runs in which the detection potential (EDET) in IPAD was changed incrementally, leaving unchanged both the oxidation (EOX) and reduction potentials (ERED) values. The EDET was examined in the potential window from -100 to +300 mV, that is, in the region where the divalent ion effect was particularly relevant in terms of electrochemical oxidation of sugars and alditols. The FI responses in 0.58 M NaOH are qualitatively similar for all compounds tested at the Au electrode. In agreement with the voltammetric data, the resulting HDVs in the presence of DNCs in the alkaline carrier electrolyte provide more useful information about the electrooxidation process of analytes. Moreover, these cations impart some noticeable differences in the anodic signal of sample molecules; albeit the responses in 0.58 M NaOH (dashed lines) and 1 mM Ba(II) + 0.58 M NaOH (solid lines) of D-sorbitol, D-glucose, and sucrose develop into broad peaks with the maximum located at the same potential, signal increments of 18, 24, and 28%, respectively, were evaluated. Enhancement of D-sorbitol sensitivity is 4854 Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

Figure 9. Effect of Ba(II), Sr(II), and Ca(II) on the detection potential of 0.25 mM sucrose in flow injection with IPAD. Other conditions as in Figure 7.

consistent with the prediction based on the voltammetric results described earlier. It is especially noteworthy that the location of the maximum response for D-sorbitol in the presence of Sr(II) ions is shifted to a less positive detection potential by ∼100 mV (i.e., EDET ) +50 mV) with an increment of ∼50% in the signal intensity. In the case of Ba(II), under the same experimental conditions, the anodic signal enhancement for the same analyte relative to the blank alkaline electrolyte is ∼30%. These findings agree very well with the significant improvements previously reported in the chromatographic separations of sugars and alditols using DNC-containing alkaline mobile phases. An examination of the responses of D-sorbitol, D-glucose, and sucrose in the presence of Ca(II) ions clearly shows that these compounds are oxidized at a lower peak potential compared to the HDVs recorded with the blank carrier electrolyte. The only exception is the response of sucrose, which is considerably attenuated in the potential region investigated. It may be possible to detect sucrose by choosing EDET < 0 mV, but it should be noted that O2 is reduced at these potentials. Indeed, the variation of background current with potential must also be considered when a choice of optimum detection potential is made. The background current (not shown) is essentially constant and very near zero for the region from -50 to +250 mV and increases dramatically at more negative potentials where oxygen is reduced, thereby making worse the signal-to-noise (S/N) ratio. Moreover, the background current increases when EDET is shifted to more positive values because a greater amount of surface oxide is produced. An optimum value of EDET between 0 and +50 mV vs Ag/AgCl corresponds to the largest S/N ratio. For this value, there exists only a very small contribution from reduction of dissolved O2 and virtually no contribution from surface oxide formation. CONCLUSIONS The enhanced electrocatalytic activity of gold electrodes in the presence of Ba(II), and Sr(II) toward the oxidation of polyhydroxy compounds in basic media is attributed to changes in the adsorption features of the surface and, particularly, to the restrained formation of gold oxide. It is also believed that the oxidation products are more rapidly desorbed from the Au surface, allowing the surface sites to be recycled within the electrocatalytic mechanism. The most substantial difference between Ca(II) and the other two cations is the extent to which the calcium ions

suppress the onset of gold oxide formation. While the electrochemical results provide further evidence for the existence of alkaline-earth metal complexes with polyhydroxy compounds, further work is required to establish the surface effect of these DNCs. These findings confirm, however, that the divalent ions investigated here are very useful, at a millimolar concentration, as alkaline mobile-phase components in anion-exchange chromatography applications with pulsed amperometric detection.

provided partial funding for this research for developing a laboratory of synthesis and characterization of innovative materials “LaMI”. This work was also supported by the National Research Council of Italy (CNR) and Ministero dell’Universita` e della Ricerca Scientifica e Tecnologica (MURST).

ACKNOWLEDGMENT The authors express their gratitude to Dr. G. Margiotta for her help with preliminary experiments. The Regione Basilicata

AC970471C

Received for review May 7, 1997. 1997.X

X

Accepted July 29,

Abstract published in Advance ACS Abstracts, October 15, 1997.

Analytical Chemistry, Vol. 69, No. 23, December 1, 1997

4855