AU598568B2

AU598568B2 - Method and apparatus for amperometric detection

Info

Publication number: AU598568B2
Application number: AU76425/87A
Authority: AU
Inventors: John Wilson Bixler; Alan Maxwell Bond
Original assignee: ICI Australia Operations Pty Ltd
Current assignee: Orica Australia Pty Ltd
Priority date: 1986-07-04
Filing date: 1987-07-02
Publication date: 1990-06-28
Anticipated expiration: 2007-07-02
Also published as: AU7642587A

Description

METHOD AND APPARATUS FOR AMPEROMETRIC DETECTION

This invention relates to a method and apparatus for amperometric detection. More specifically, the invention is concerned with new methods for electrochemical detection at the trace analysis (sub-picomole) level using microelectrodes in a flow-jet cell and femtoamp current detection methods.

A flow-jet cell, as referred to herein, is an electrochemical cell of the flowthrough type which is characterized by having a fluid jet inlet and a working electrode arranged in such a way that a fluid stream flowing from the inlet jet is directed at the working electrode and is substantially normal to the plane containing the working surface of the electrode.

The significant benefits derived from the application of microelectrodes to the solution of electrochemical problems are becoming well-known. Although the advantages of using microelectrodes under quiescent conditions have been exploited for several years, their application to flow-through cells is more recent. The use of a cylindrical carbon fibre microelectrode inserted into the exit port of an open tubular liquid chromatography (LC) column has been reported, as has the incorporation of an array bundle of carbon fibre microelectrodes in to a thin-layer channel-type flow cell.

Apparently all applications of microelectrodes as sensors in conventional flow cell configurations have been in thin-layer detectors. We have now found that single microdisk electrodes can be used as sensors under free-flow (thick-layer) conditions, particularly in wall-jet cells.

The use of electrochemical detection for LC has generally excluded normal phase applications due to electrical resistance problems. The potential advantage of normal phase liquid chromatography electrochemicaldetection (LCEC) operation with regard to decreased adsorption and organic solubility problems and enhanced working potential range are obvious, but few actual applications have been reported. In cases where normal phase LCEC has been used, eluents were confined to solvent mixtures in which an electrolyte is soluble or else conductance is enhanced by post-column treatment. We have now found that microelectrodes using flow injection to introduce samples into a jet cell with an organic solvent system can be used in liquid systems containing electrolyte concentrations as low as 10^-6M.

Our investigations have shown that disk microelectrodes clearly perform as well as, or better than, a large surface sensor in a jet-type flow cell. Not only are signal-to-noise ratios enhanced, but these microsensors function very well in highly resistive media, which overcomes a major versatility limitation in the application of electrochemical flow cells as liquid chromatographic detectors. This advantage can be further enhanced by optimizing reference electrode and cell configuration design to minimize resistive effects.

The major difficulty in measurements of low concentrations with microelectrodes is that femtoamp electrical currents must be measured to exploit their full analytical capability. Despite suggestions that picoammeters can be used in place of potentiostats with microelectrodes at slow potential scan rates, most workers continue to use a conventional potentiostat, albeit sometimes in the two-electrode mode.

We have investigated the use of picoammeters, as well the use of femtoammeters, as detectors for microelectrode flow-cells. We have found that the elimination of the potentiostat is not only a possibility, but we have shown it to be a distinct advantage. The only instrumentation required is a noise-free, low-resistance off-set voltage and a sensitive ammeter. We have also developed a battery operated voltage source as a means of minimizing noise when measuring very small currents in the femtoampere range.

Thus in accordance with one aspect of the present invention, there is provided a method for electrochemical detection of an electroactive species in which a liquid sample solution containing the species is passed through a flow jet cell provided with a microelectrode as the working electrode.

In accordance with one preferred aspect of the invention, the current flow through the cell is measured using a picoammeter or femtoammeter.

Thus, a preferred apparatus in accordance with the invention comprises the above cell/microelectrode combination in conjunction with a picoammeter or femtoammeter.

More specifically, the present invention provides apparatus for the electrochemical detection of an electroactive species in a flowing stream characterized in that it comprises a flow-jet cell having:

(i) a working electrode with a surface of nominal diameter less than or equal to 50 micrometre;

(ii) a fluid jet inlet having an orifice with a nominal diameter less than 0.3 mm aligned with the surface of the working electrode so as to direct substantially all of the fluid stream flowing from the orifice into the cell in a direction substantially normal to the surface of the working electrode, and so that the surface of the electrode is wholly within the diameter of the said fluid stream; and

(iii) a reference electrode.

The working electrode may consist of a single element or it may comprise an array of electrode elements each with a surface of nominal diameter less than or equal to 50 micrometre.

The working electrode(s) may be composed substantially of metal or of carbon. The preferred metal is platinum but any other suitable metal may be used.

The invention, and various specific and preferred aspects thereof, will be further described and elucidated in the following non-limiting examples. The following abbreviations are used:

AN = Acetonitrile

Fc = Ferrocene

Feoc = Potassium ferrocyanide

TEAP = Tetraethylammonium perchlorate.

Except where otherwise stated reference to the size of an electrode signifies the nominal diameter of the electrode.

Reference will be made to the accompanying drawings which are identified as follows:

Figure 1 and 1a

Circuit diagram of battery based voltage offset device for applying constant potential for DC amperometricdetection with flow injection analysis. Figure la shows how the device of Figure 1 is used with a cell and current measuring device.

Figure 2

Cyclic Voltammograms of ImM Fc in AN 0.01M TEAP (Oi is zero current). Three electrode potentiostatic mode, scan rate = 10mV/s.

(a) 5μm carbon microdisk.

(b) 50μm platinum microdisk.

Figure 3

Flow-Injection Hydrodynamic Voltammogram for 50μm Platinum Microdisk. Sample: 20μL ImM Fc in acetonitrile 0.01M TEAP. Solvent flowrate : 1.4mL/min. (o) peak height, (Δ) peak area function.

Figure 4

Flow rate Dependency of Peak Height for oxidation of Feoc in aqueous 0.05M KNO₃ E = + 1.0V vs Ag/AgCl (3M aq. NaCl), (o) 4.6mm Pt disk, ImM Feoc; (Δ) 50μm Pt disk, ImM Feoc; (□) 5μm carbon disk, 20mM Feoc.

Figure 5

Flow-Injection Peaks for Platinum Disk Sensors. E = +1.1V VS Ag/AgCl (3M aq. KCl), Flow rate: 1.4mL/min., X-axes: 50s. in three electrode mode.

(a) 4.6mm Pt disk, 0.1μM Fc in AN (ImM TEAP).

(b) 50μm Pt disk, 0.lμM Fc in AN 10 μM TEAP.

(c) 50μm Pt disk, potentiostat only, 2μM Fc in AN (0.01 M TEAP).

(d) 50μm Pt disk, preamplified potentiostat, 2μM Fc in AN (0.01 M TEAP). X marks analytical peaks; other peaks are injector rotation spikes.

Figure 6

Concentration Dependency of Flow-Injection Peak Current for a 5μm Carbon Disk with Ammeter Detector. Sample: 30μL Fc in AN (ImM TEAP). E = + 1.016V vs Ag/AgCl (3M aq. KCl), flow rate 1.4mL/min.

Figure 7

Flow Injection Peaks for the 5μm Carbon Disk Sensor. E = + 1.0 V vs Ag/AgCl (3 M aq. KCl), 1.4mL/min. flow rate. Sample: 20μL Fc in TEAP + AN, X-axis: 50s/in.

(a) 0.lμM Fc in ImM TEAP, three electrode mode: preamplified potentiostat.

(b) 20 pM Fc in ImM TEAP, two electrode mode: ammeter detector.

(c) 0.lμM Fc in 10μM TEAP, three electrode mode: preamplified potentiostat.

(d) 20pM Fc in 10μM TEAP, two electrode mode: ammeter detector.

(e) 0.4μM Fc in lμM TEAP, three electrode mode: preamplified potentiostat.

(f) 0.4μM Fc in lμM TEAP, two electrode mode: ammeter detector.

X marks analytical peaks; other peaks are injector rotation spikes.

Figure 8

Preferred form of flow cell in accordance with the invention. Reagents. The reagents and apparatus used in the examples were as follows:

Reagent grade ferrocene (Fc), potassium ferrocyanide (Feoc) and potassium nitrate, and HPLC-grade acetonitrile (AN) were used as received. The water was double-distilled and deionized. Tetraethylammonium perchlorate (TEAP) (Southwestern Analytical Chemicals, Inc.) was dried for several hours in a vacuum desiccator over phosphorus pentoxide prior to use.

Electrochemical Cell and Microelectrodes.

Two microdisk electrodes were prepared to fit the working electrode port of a Metrohm EA 1096 wall-jet type flow-through detector cell. The Metrohm cell design required that the disk be located at the centre of the flat end face of a 7mm o.d. insulating rod and both electrodes were designed to meet this criterion.

(i) A glass-sealed 50μm diameter platinum microdisk was prepared by flame-sealing a 50μm platinum wire into a 5mm segment of 5mm o.d. soda glass capillary tubing. A fine copper wire was soldered to one end of the platinum wire and the platinum-glass segment was sealed into one end of a 6cm length of thin-walled 7mm o.d. glass tubing using Torr Seal to make the seal and strengthen the solder joint inside the tube. The electrode end was cut off flat with a diamond-bladed saw, polished and cable connections made as described below.

(ii) The second micro-disk detector was constructed from a Magnamite carbon fibre (Hercules Corp.) with a nominal diameter of 5μm. A 1cm conical segment was cut from a piece of heavy-wall capillary tubing which had been drawn slightly so that about half of the conical segment would fit inside a 7mm o.d. glass tube. The edges of the glass cone were fire-polished and a piece of carbon fibre was fed through the capillary orifice by micro-manipulation so that the fibre protruded from both ends. The small end of the glass cone was fitted to a transparent vacuum hose, a drop of Torr Seal was placed on the wide end of the cone and gentle suction was applied, holding the fibre in place, until Torr Seal visibly filled the capillary. A fine copper wire was attached to the carbon fibre at the narrow end of the cone using silver conducting epoxy. The narrow end of the cone was then sealed into a 6cm length of 7mm o.d. glass tubing with Torr Seal. After the epoxy hardened, the cone end was cut off flush with the outer tube. The copper wire attached to each microelectrode was soldered to the lead of a coaxial cable. The cable was rigidly attached to the electrode body with heat-shrinkable plastic tubing. The electrodes were polished on successively finer grades of wet emery paper, ending with P1200, and then with successively finer grades of alumina slurries (5, 1, 0.3, 0.05μm) on microcloth. Each electrode was repolished briefly with 0.05μm alumina at the beginning of a day's use.

The Metrohm silver-silver chloride reference electrode body designed to fit the flow cells was filled with aqueous 3.0 M NaCl. The bridge between the reference electrodes and the cell was filled daily with fresh aqueous 0.1 M TEAP for experiments in non-aqueous media and 0.05 M KN03 for aqueous media. The auxiliary electrode was a Metrohm glassy carbon electrode, which was also used to plug the auxiliary electrode port during two-electrode operation. The cell was enclosed in a grounded Faraday cage.

Figure 8 shows a preferred form of flow cell specifically designed for use in accordance with the invention.

The components of the cell are identified as follows;

1. Main cell body

2.* Reference electrode

3. Working electrode support

4.* External contact for working electrode

5. O-ring seal

6,7 Keeper plate and set-screw for locking support 3 to body 1

8. Inlet insert, incorporating inlet jet

9. Outlet insert 10.* Microelectrode

* These components are metal or other electrically conductive materials. The remaining components are composed of insulating material(s), except 6 and 7 which are metal.

In the arrangement depicted, the working microelectrode 10 is concentric with the silver plated stainless steel reference electrode 2. This arrangement minimises the resistance and capacitance of the cell and ensures a uniform electric field and an equipotential surface across the working electrode.

This results in a current density distribution on the working electrode which is determined solely by the cell hydrodynamic pattern.

Detector Instruments

Four different systems were used to apply a potential to the working electrode and detect the cell current.

(i) The first system was a Metrohm E611 Detector (potentiostat) which was also equipped with a Metrohm E612 Scanner for cyclic voltammetry.

(ii) The second system was the Metrohm E611 potentiostat with a Keithley 480 picoammeter connected as a preamplifier at the working electrode input. The characteristics of this system have been described previously (Bixler J.W., Bond A.M., Lay P.A., Thormann W. , Van der Bosch P., Fleischmann M. , Pons S. Anal. Chem. Acta. V187, 1986, p67).

(iii) The third system was a Keithley 614 digital electrometer (used in the current mode) and a miniature standard Weston cell, both connected in series with the two-electrode cell. The Weston cell was physically located inside the Faraday cage with the analytical cell and provided a noise-free low resistance + 1.016 V emf applied to the working microelectrode. (iv) Finally, a variable voltage battery operated power source was constructed according to the circuit shown in Figure 1 and used as an alternative to the Weston cell in the third system.

In Figure 1, the circuit components and their values are indicated by the usual symbols. The AD584 is a precision voltage reference programmed to produce 5.0V at pin 1 relative to pin 4. The potentiometer circuit formed by R1 and RV1 produce a variable 0 to 3.5V voltage at pin 3 of the 7611 operational amplifier which is configured as a voltage follower circuit with a gain of unity. The polarity of the output (0-3V DC) can be made either positive or negative depending on the position of the reversing switch, SW2. SW1 is the on/off switch. The circuit is powered by 9-18V battery B.

Figure la shows the connections between the voltage source V of Figure 1, the reference and working electrodes (RE,WE) of the cell and the Keithley 614 digital electrometer (E) (which is represented diagrammatically) and is operated in its current to voltage converter mode. The voltage source V sets the potential of the reference electrode and thus the solution potential relative to the common potential (C). The feedback action of the input amplifier (A) contained within the electrometer (E) is such that the potential of the working electrode WE is maintained close to the common potential. The output of the electrometer (E) appears at the terminals (T₁, T₂).

Flow-injection peaks were recorded with a Houston 2000 recorder in the y-t mode. The recorder was connected to the undamped output of the Metrohm detector or the normal analog output of the multimeter. The response characteristics of the detector-recorder systems were evaluated by generating and recording synthetic triangular peaks. The detector-recorder systems showed less than 1.5% peak truncation for full-scale rise times of one second or greater. Peak areas were measured using a Waters Data Module.

Flow Injection System

Except where the use of the single-piston BAS LC-6 pump is specifically indicated, all flow-injection experiments were undertaken with a double-piston Waters 6000A pump coupled to a Rheodyne 7125 injector equipped with a 20μL sample loop. The injector was connected to the jet nozzle of the cell with a 18cm length of 0.3mm i.d. teflon tubing.

Example 1

Microelectrode Characteristics

The non-linear diffusion properties of the microelectrodes are illustrated by the cyclic voltammograms shown in Figure 2, which were recorded in quiescent solutions. They both exhibit the sigmoidal shape at slow to moderate scan rates which is characteristic of situations where spherical diffusion predominates. The effective radii of the two microdisks can be compared, since Equation (1) describes the limiting current of a microdisk electrode under conditions of spherical diffusion. i_d=4nFDCr (1)

where id = diffusion-controlled limiting current n = number of electrons involved in the electron transfer

F = Faraday constant

D = diffusion coefficient

C = concentration r = electrode radius.

From the data in Figure 1 and assuming the diffusion coefficient of ferrocene in acetonitrile to be 2.3 x 10 ^-5cm²/sec, the effective radii of the carbon and platinum microdisk electrodes are estimated to be 2.4μm and 27μm respectively, compared to nominal values provided by the manufacturers of 2.5 and 25μm.

Flow-injection hydrodynamic voltammograms were obtained by repeated injection of a ferrocene sample into the flow system and recording the peak information at a series of fixed potentials over the range of interest. Figure 3 shows both peak current and peak area plotted as a function of applied potential for the 50μm platinum microdisk electrodes. The plots show distinct sigmoidal dependence upon applied potential and permit optimum selection of an applied potential for analytical applications.

Example 2

Flow Rate Dependencies

Data illustrating the dependency of flow-injection peak heights upon the flow rate of the running solvent for the 4.6 mm diameter platinum disk and the carbon and platinum microdisks are presented in Figure 4 for the oxidation of [Fe(CN)₆]^4- in aqueous media. Each plotted data point is the average of five or more replications. The relative standard deviation for all three electrodes is typically between 3 and 4% for these experiments and independent of electrode, suggesting that the peak height reproducibility is limited by the performance of the sample injector.

The exponential flow rate dependencies for the peak height data from Figure 4 are given in Table I. The flow rate dependency with the 4.6mm platinum disk is not far from the theoretical 0.75 power flow rate dependency predicted for constant electroactive sample flow into a wall-jet detector. The 50μm platinum and 5μm carbon disks are certainly impinging jet electrodes, rather than wall-jet electrodes, since their electrode radius-to-nozzle radius (approximately 0.3mm) ratio is large and hence both electrodes experience turbulent rather than laminar flow.

These results show the 5μm disk electrode is clearly superior to the 50μm and 4.6mm peaks which have much larger flow rate dependencies.

The flowr-rate dependency of the 50μm platinum and 5μm carbon electrodes was also examined under steady-state, rather than injected, sample flow conditions by continuously pumping ferrocene solution containing 0.01 M TEAP in acetonitrile through the flow cell. Under these conditions the continuous current for the 50μm platinum and 5μm carbon disks had exponential flow rate dependencies of 0.74 ± 0.01 and 0.188 ± 0.003, respectively, which are markedly lower than those obtained with 20 μL injected samples of K₄Fe(CN)₆ in aqueous media of ferrocene in acetonitrile. Thus, the flow rate dependency needs to be carefully evaluated with respect to specified conditions of use. However, the 5μm carbon disk electrode clearly shows the desired low flow rate dependence under both steady-state and non-steady-state flow conditions.

Furthermore, the thickness of the boundary layer and hence the current was found to be, be independent of location near the stagnation point at the center of the impingement. This was conveniently tested with the 5μm carbon disk, since microscopic examination clearly indicated that the carbon fiber was located slightly off-center. Rotation of the electrode 120° in either direction in its mounting had no effect upon injection peak heights, verifying independence of current upon position in this region. From this point of view, the 5μm disk electrode is clearly superior to the 50μm and 4.6 mm electrodes, which have much larger flow rate dependencies of peak height.

Example 3

Analytical Considerations

The concentration-dependence of the microelectrodes were examined using ferrocene sample concentrations ranging from one millimolar down to sub-micromolar under a range of experimental conditions. In each case, flow-injection hydrodynamic voltammograms were used to select an applied potential on the limiting current region (+1.0 or +1.1V) and a running solvent flow rate of 1.4ml/min. was used with all electrodes. In all cases. the linearity with concentration and experimental precision was as good, and in some cases better, when peak heights rather than peak areas were used as the analytical signal. Thus, there appears to be no advantage to employing integration in cases where the detector system response is sufficiently fast to accurately record the peak height.

A summary of the results of concentration-dependence studies with the micro-sensor is given in Table II.

Experiments with the 4.6mm platinum disk are also included for comparison. As expected the relative analytical sensitivity (amplitude of actual current measured) increases as the electrode size increases. Variation in sensitivity for a given electrode reflects different solution and detector amplification characteristics as well as area changes which result from daily polishing.

However, the relative sensitivity is of much less importance than linearity of response and signal-to-noise ratio. Both of the microelectrodes exhibit a linear response range as low or lower than the convential 4.6mm platinum disk. Moreover, the 4.6mm disk exhibits some undesirable features seen only to a lesser degree or not at all with the microelectrodes. Very large injection spikes are obtained in millimolar electrolyte with the

4.6mm disk upon interrupting the flow momentarily when rotating the injector valve (Figure 5-a) and baseline stability is restored very slowly after injection. This makes the measurement of small peaks difficult and precludes useful peak integration below the micromolar concentration level. The 4.6mm electrode was useless in dilute (10^-5M) electrolyte, since the compliance voltage of the potentiostat was insufficient to maintain potential control in this resistive medium for ferrocene concentrations greater than 4 micromolar (or current <0.1μA). Also, a linear response was not obtained at lower concentrations where potentiostatic control was restored because of uncompensated resistance. Thus the conventional macrodisk clearly is unsuitable for use in highly resistive solutions.

The 50μm platinum disk shows significantly improved response over the 4.6mm disk, giving a very acceptable response even in 10^-5M electrolyte. Figure 5-b shows that although injection spikes are seen on low current ranges, a flat baseline is restored quickly after injection and the baseline noise level is very low. Figures 5-c and 5-d illustrate the difference in the signal-to-noise ratio when a 50μm platinum electrode is used with a potentiostat alone in contrast to a potentiostat with a picoammeter preamplifier. The use of the preamplifier results in about a 6-fold improvement in the signal-to-noise ratio consistent with what has been reported for similar detector configurations used to record cyclic voltammograms in static solutions.

The cell response employing the 5μm carbon disk sensor was examined at three different support-electrolyte concentrations using both the picoammeter-preamplified potentiostat in the three-electrode mode and the electrometer (ammeter mode) without potentiostatic control in the two-electrode mode. The battery voltage supply of

Figure 1 was used in the latter mode to minimize noise.

An example of the concentration-dependent response is illustrated in Figure 6. The linear response in 10^-3 and 10^-5 M TEAP extended to slightly lower ferrocene concentration in the two-electrode than in the three-electrode mode, as is reported in Table II. Moreover, a comparison of Figures 7a and 7b clearly indicates that the background noise is substantially lower in the two-electrode mode. This background noise reduction was also observed in 10 -5 M TEAP. Figure 7b shows that 4 x 10 ^-13 mole samples are easily detected in the unpotentiostated two-electrode mode even with Dilute electrolyte. Note that currents being measured are in the femptoampere range and electronic noise is therefore of concern, so that all experiments were performed in a

Faraday cage. It was observed that the connecting cables were much less sensitive to mechanically induced noise in the two-electrode mode with the potentiostat absent from the circuit.

Claims

1. Apparatus for the electrochemical detection of an electroactive species in which a liquid sample solution containing the species is passed through a flow jet cell, characterized in that a microelectrode is used as the working electrode.

2. Apparatus for the electrochemical detection of an electroactive species in a flowing stream characterized in that it comprises a flow-jet cell having:

(ii) a fluid jet inlet having a orifice with a nominal diameter less than 0.3 mm aligned with the surface of the working electrode so as to direct substantially all of the fluid stream flowing from the orifice into the cell in a direction substantially normal to the surface of the working electrode/ and so that the surface of the electrode is wholly within the diameter of the said fluid stream; and

(iii) a reference electrode.

3. Apparatus as claimed in Claim 1 or Claim 2, characterized in that the working electrode is an array of electrode elements each with a surface of nominal diameter less than or equal to 50 micrometre.

4. Apparatus as claimed in any one of Claims 1 to 3, characterized in that the working electrode(s) is/are composed substantially of metal.

5. Apparatus as claimed in any one of Claims 1 to 3, characterized in that the working electrode(s) is/are composed substantially of carbon.

6. Apparatus as claimed in any one of the preceding Claims, characterized in that the current, flow through the cell is measured using a picoammeter or femtoammeter.

7. Apparatus as claimed in any one of the preceding Claims, characterized in that the solution potential relative to the working electrode is provided by a voltage source connected between common potential and the reference electrode.

8. Apparatus as claimed in Claim 7, characterized in that the potential of the working electrode is maintained close to common potential by the action of an operational amplifier configured as a current to voltage converter.

9. Apparatus as claimed in any one of the preceding Claims, characterized in that the reference electrode is located coaxially with respect to the working electrode.

10. Apparatus as claimed in any one of the preceding Claims, characterized in that the cell is enclosed within a Faraday cage.

11. A method for the electrochemical detection of an electroactive species, characterized in that a sample containing the species is added to a solvent stream for injection into the apparatus of any one of the preceding Claims. 12 A method as claimed in Claim 11, characterized in that the solvent is a solvent system containing electrolyte concentrations lower than 10^-4 molar.