Flow-injection analysis of oxidizable species with reverse-pulse amperometric detection

Tahnta, Vol. 29, pp. 901 to 904, 1982

0039-9140/82/l 10901-04$03.00/O

Printed in Great Britain. All rights reserved

Copyright 0 1982Pergamon Press Ltd

FLOW-INJECTION ANALYSIS SPECIES WITH REVERSE-PULSE DETECTION Department

OF OXIDIZABLE AMPEROMETRIC

JOSEPH WANG* and HOWARD D. DEWALD New Mexico State University, Las Owes,

of Chemistry,

(Received

NM 88003, U.S.A.

10 May 1982. Accepted 12 June 1982)

Summary-The technique of reverse-pulse amperometry is applied for the detection of oxidizable organic species at a solid-electrode flow detector. Exploiting the reverse-pulse amperometric waveform gives better sensitivity than d.c. amperometric detection. Species (e.g., phenols) giving responses that are poorly separated from background are easily monitored. Reducible species can be monitored without deaeration of the solution. The effects of flow-rate, waiting time between pulses, precision, and linearity of response are reported. At a flow-rate of I.0 ml/min injection rates of 180 samples per hour and detection limits of a few tenths of a nanogram are obtainable.

proved with the RPA detection mode: sensitivity, detection of species with high redox potentials, and detection of reducible species without interference from dissolved oxygen. The equipment is simple (a modern polarographic analyser with the normal pulse mode and a potential-hold capability) and available in most laboratories.

Continuous analysis in flowing streams with solidelectrode flow detectors has gained increased attention in recent years. ‘J Most solid-electrode detectors utilized in liquid chromatography (LC) or flow-injection analysis (FIA) employ d.c. amperometric detection. Recently, several pulsed-potential waveforms have been applied to LC and FIA detection, mainly in conjunction with cathodic reactions at the dropping mercury electrode. These include differential-pulse detection3 square-wave detection,4 and reverse-pulse amperometric detection of amalgam-soluble metals.’ Applications of reverse-pulse amperometry (RPA) to the flow analysis of oxidizable organic species at a solid-electrode detector have not yet been described. This study characterizes the analytical utility of RPA in a flow-cell with a carbon electrode, as applied to redox reactions in which the (oxidizable) reactant and the (reducible) product coexist in solution. This approach is a variant of normal pulse voltammetry; it is based on the application of an unsymmetrical square-wave potential with a long application of a positive initial potential E, followed by a short pulse at a more negative final potential E,. Oxidizable species are measured by monitoring the reduction (at the end of the more negative pulse) of the oxidation product from the initial potential (in the plateau region):

AAB-A

EXPERIMENTAL Appurutus The electrochemical “wall-jet” detector has been described in detail previously.* The working electrode was a planar glassy-carbon disk (0.25 cm diameter) with a solution stream directed onto it from a solution inlet nozzle (0.34 mm diameter). The distance between the nozzle tip and the surface of the glassy carbon was 0.025 cm. An Ag/ AgCl reference electrode was placed in the cell downstream from the working electrode. The carrier and sample solutions were stored in two Nalgene beakers, with similar hydraulic heads to provide equal flow-rates. These reservoirs were connected, through two Teflon tubes (1.0 mm bore, 0.05 mm wall), to a threeway Teflon stopcock located I2 cm from the detector. The stopcock and the cell were connected by I.O-mm bore Teflon tubing. All measurements were made with a Princeton Applied Research Model 174 Polarographic Analyzer. Detection peaks were recorded on a Houston Omniscribe strip-chart recorder. Reagents The chemicals and reagents used have been described in detail previously,’ except as noted. Stock solutions of NADH, chlorpromazine, phenol and benzoquinone were prepared each day. Aliquots of the stock solution were added to the phosphate buffer supporting electrolyte to give the desired concentration.

E,

The current sampled at this point is proportional to the concentration of the species in the flowing stream. The generation/detection capability of RPA yields, at a single electrode, advantages similar to those reported recently for dual-electrode detectors6*’ Different aspects of electrochemical detection can be im-

Procedure The potential limits (Ei and E,) for the RPA were set with the PULSE mode of the polarographic analyser, in a similar way to that described by Maizota and Johnson.’ E, and E, were chosen to be in the limiting current regions for the forward and reverse reactions, respectively. The flow-

*Author for correspondence. 901

JOSEPH WANG and HOWARD D. DEWALD

902

IO nA 30 SW

\ fiLM RPA

DC

a

b

Fig. 1. Typical peaks for detection of lO@f ferrocyanide (a) and NADH (b) by FIA and RPA and d.c. detection modes. Conditions: flow-rate, 1.0 ml/min; Sample flow, 5 set; carrier flow (O.lM phosphate buffer), 25 set; sample size, 83 .ul; pulse repetition time, 0.5 set; low-pass filter, 1 set; potentials, (a) RPA with Ei = + 1.0 V, E, = -0.3 V; d.c. at + 1.0 V; (b) RPA, E, = +0.9 V, E, = f0.2 V: d.c. at f0.9 V.

injection measurements were performed with the three-way valve by alternating periodically, for fixed times, between the carrier solution (the supporting electrolyte) and the sample solution. Details are given in the following section.

RESULTS

AND

DISCUSSION

Figure 1 compares RPA detection peaks with those obtained in conventional d.c. amperometric detection for injections of lo@4 solutions of ferrocyanide (a) and NADH (b). The advantage of the RPA detection mode is evidenced by its higher (by -50%) peak currents. A surprising aspect of the RPA detection is that compounds (e.g., NADH, ascorbic acid) which appear chemically irreversible with “slower” potential-scan techniques (such as cyclic voltammetry), have a significant RPA response because of the different time scale. (Initially, we intended to exploit the expected discrimination against compounds with irreversible redox reactions for improving the selectivity in flowanalysis of mixtures containing “reversible” and “irreversible” compounds.) A similar observation was reported recently for d.c. measurement at a dual-electrode detector with a short spacer between the upstream and downstream electrodes.’ For comparison of the results obtained by RPA with those for d.c. detection, the relative signal, expressed as RPA-peak current/d.c. peak current (i,,,/i,,) may be used. The value of i,, is given by the

equation for the limiting current at the “wall-jet” detector [ref. 10 equation (lo)], and the RPA response can be described on the basis of the mixed hydrodynamic-Cottrell behaviour of pulse voltammetry at convective solid electrodes.’ ’ At low convection rates and short pulse-widths the RPA current will obey the Cottrell equation (and will be independent of convection transport); at higher convection rates, the convection transport will control the current. This behaviour is indicated by the data of Fig. 2a. The relative insensitivity of the RPA response to the Row-rate (below 1.0 ml/min) would be advantageous for making measurements in flowing systems with poorly controlled flow-rates. The mixed hydrodynamic-Cottrell RPA response assumes that the oxidation product is stable and is not adsorbed on the electrode. Therefore, RPA may be used as a tool for confirming the identity of an analyte which gives a product (at EJ that is subject to chemical reaction or adsorption at the electrode (with the iRPA/idcvalue as the criterion in a similar way to the collection efficiency in ring-disk or other dual electrode experiments). Table 1 gives the i,,,/i,, values for several compounds. Though for most compounds a 50% increase in sensitivity is ob- 1.5) for ch\orpromdzine (a comserved (i&i,, pound known to interact with a carbon surface”) there is a 20% decrease (iRPJidc =0.79). The effect of the waiting time between pulses (pulse repetition time) on the RPA response is shown in

Flow-injection analysis of oxidizable Flow

rate,

ml/min

I I

0.5

I

1.5

I

2

I

3(

a c

2< L

.,a

I I

Pulse

I

I

3

5

repetition,

set

Fig. 2. Dependence of the FIA/RPA peak currents on the volume flow-rate (a) and the waiting time between pulses (b). Conditions: (a) 10pM ferrocyanide; sample flow, 5 set; carrier flow (O.lM phosphate buffer) 25 set; Ei = + 1.0 V; E, = 0.3 V; low-pass filter, 1 set; (b) 5 ,uM benzoquinone; sample flow, 5 set; carrier flow (0.1M phosphate buffer) 20 set; flow-rate, 1.0 ml/min; E, = -0.8 V; E, = +0.6 V; low-pass filter, 1 sec.

Fig. 2b. The four repetition times available with the PAR 174 analyser were examined. As the waiting time increases, from 0.5 to 5 set, the FIA/RPA peak current decreases by more than 50%. This is because as the waiting time increases, the frequency of sampling the current decreases, i.e., fewer points are sampled during the passage of the short peak-shaped sampleplug profile through the detector (a peak-shaped response was obtained at the different times employed). Different effects on the peak current are expected if larger volumes are injected or if the product formed at Ei is not stable. From the data found, a repetition time of 0.5 set was selected. The main application of RPA detection is not to measurement of easily-oxidized analytes. AmperoTable 1. Values of

i,,,/i,,,

Ei, Compound Ferrocyanide Chlorpromazine NADH Phenol Benzoquinone

V

f0.9 + 0.9 +0.9 + 1.4 -0.8

I(A)--8

metric d.c. detection of compounds which react at high potentials (> 1.O V), where background processes (water or mobile-phase oxidation) occur, suffers from detector drift, increased background and noise levels, and the need for high current gain.6*7.13 These compounds may be better detected by using the generation/detection capability of RPA, i.e., by detecting the oxidation product at lower potentials. A reversepulse procedure was exploited in a similar way to deal with the interfering hydrogen-evolution reaction, in batch analysis at the dropping mercury electrode.14 Among the important compounds that are oxidized at potentials near that for solvent decomposition is phenol, the oxidation of which reaches a current plateau at potentials higher than + 1.2 V (US. Ag/AgCl electrode). I5 Problems associated with d.c. amperometric detection of phenolic compounds have been reported. 15*16Figure 3a illustrates the RPA detection peaks for successive injections of 2.6pM phenol in phosphate buffer solution (corresponding to 20 ng in the injection volume used). The sampling rate is 180/hr. Well-defined and reproducible peaks are observed. A detection limit near 50nM (0.4 ng) is expected (signal/noise = 2). Two separate experiments were performed to estimate the precision of the results. For a series of 10 repeated injections of a 25pM ferrocyanide solution, the average value of the peak current was 0.364 PA (with a range of 0.3540.370 PA), and the relative

a

5 nA (01 I IO nAlb)

for several compounds*

Ef, V

0.0 0.0 f0.2 -0.1 +0.6

iRPA/idc

1.55 0.79 1.46 1.54 1.41

*Concentration, 10 PM. FIA conditions and instrumental parameters as for Fig. 1. The d.c. measurements were made at potential Ei. TN_. 29jl

903

species

b

Fig. 3. Detection peaks in flow-injection analysis of a 2.6pM phenol solution (a) and a 5pM benzoquinone solution (b). Conditions: flow-rate, 1.0 ml/min; repetition time, 0.5 set; low-pass filter, 1 set; sample flow, 5 set; carrier flow (0.1M phosphate buffer), 15 set (a); 20 set (b); potentials, Ei = + 1.4 V, E, = +0.6 V (a).

904

JOSEPHWANGand HOWARDD. DEWALD

Fig. 4. Reverse-pulse amperometric detection for injections of 1, 2, 3, 4, 5 and 6 pcM chlorpromazine solutions. Flow conditions, repetition time and filter, as for Fig. 1. Ei = f0.9 V, Er = 0.0 V.

standard deviation was 1.3% (conditions as in Fig. la, except that the carrier flowed for 10 set). A series of 8 repeated injections of an 11pLM phenol solution yielded an average value for the peak current of 139 nA (range 134142 nA), and a relative standard deviation of 1.0% (conditions as in Fig. 3a, except that the carrier flowed for 25 set). Another benefit of RPA detection is that low concentrations of reducible species can be monitored without interference from dissolved oxygen. The classical d.c. amperometric detection of reducible species usually requires the removal of oxygen from the test solution. MacCrehan and Dursth employed serial dual-electrode detection for eliminating oxygen interferences. In RPA detection, a similar advantage is obtained with a single electrode. Species are reduced at Ei (where reduction of oxygen occurs), and the species generated are oxidized and measured at E, (which is insufficient to oxidize the hydrogen peroxide produced in the reduction of oxygen). This advantage of RPA detection was demonstrated’ for detection of amalgam-forming metal ions by use of a dropping mercury electrode. We have exploited this capability for detecting reducible organic species at a solid electrode. Typical peaks for QtM benzoquinone obtained by FIA and RPA are shown in Fig. 3b. A defined anodic response is obtainable through the oxidation, at E, = +0.6 V, of the hydroquinone generated (by reduction) at Ei = -0.8 V. To demonstrate the utility of RPA for measurements of low concentrations of organic compounds, FIA cufrentttime data were recorded (Fig. 4) for chlorpromazine in the l-6pLM concentration range (corresponding to 299174 ng for the 83-~1 injection). Well-defined peaks and low noise level are observed. The estimated detection limit (signal/noise = 2) is 57 nM, which corresponds to 1.6 ng for the sample volume injected. The data of Fig. 4 yielded a linear calibration plot, the slope of which corresponds to a sensitivity of 5.9 nA.l.pmole’ (correlation coefficient

0.995, intercept -0.2 nA). The tailing of the peak and the longer (25 set) wash time needed in the chlorpromazine experiment may indicate that for compounds like chlorpromazine, ‘* that interact with the carbon surface the RPA involves a slow “stripping” of the analyte from the surface. In view of the results presented here, RPA flowdetection of electroactive species may be considered as a rival approach to d.c. amperometric detection. The low detection limit is a result of combining a sensitive detection mode with an effective “wall-jet” detector. The improved sensitivity. and detection of compounds with extreme redox potentials or of reducible species without the need for deaeration, indicate great promise and applicability for flowthrough detectors.

REFERENCES 1. R. J. Rucki, Tuluntu, 1980, 27, 147. 2. K. Stulik and V. Padkova, J. ElectrounuL Chmt.. 1981, 129, I. 3. W. A. MacCrehan, And. Chern., 1981. 53, 74. 4. J. Wang, E. Ouziel, Ch. Yarnitzky and M. Ariel, Anal. Chim. Acta, 1978, 102, 99. 5. P. Maizota and D. C. Johnson, ibid., 1980, 118, 233. 6. W. A. McCrehan and R. A. Durst. Ad. Chem., 1981. 53, 1700. 7. D. A. Roston and P. T. Kissinger, ibid., 1982, 54, 429. 8. J. Wang and H. D. Dewald, Talanta, 1982, 29, 453. 9. J. Wang, Anal. Chim. Acta, 1981, 129, 253.

10. J. Yamada and H. Matsuda. J. ElectroanaL Chem.. 1973,44, 189. 11. D. J. Myers, R. A. Osteryoung and J. Osteryoung, Anul. Chem., 1974, 46, 2089. 12. T. B. Jarbawi and W. R. Heineman, A&. Chim. Actu, 1982, 135, 359. 13. D. E. Weisshaar. D. E. Tallman and J. L. Anderson. Anal. Chem., 1981, 53, 1809. 14. J. Osteryoung and E. K. Eisner, ibid., 1980, 52, 62. 15. D. N. Armentrout, J. D. McLean and M. W. Long, ibid.,, 1979, 51, 1039. 16. R. C. Koile and D. C. Johnson, ibid.. 1979, 51. 741.