This invention relates particularly to carbon screen printed electrodes which are for use in particular in the electrochemical determination of chemical species in solution, and to the use of such electrodes in such electrochemical determinations.

The analysis of degradation products derived from critical components in industrial plant and machinery can often be used to assess the overall condition of the plant or machinery and even to predict impending failure of it. Examples of such analytical applications include the measurement of metallic species in lubricating oil in order to monitor bearing wear and in the determination of aldehydes in insulation oils.

Although instrumental methods such as HPLC can be used to monitor organic markers, it is obviously advantageous in several respects to be able to use sensor methods which are suitable for use in the field. However, their present use is often precluded due to lack of sensitivity or selectivity.

A potentially valuable use of portable electrochemical sensors is in the measurement of

2-furaldehyde which is a marker of power transformer degradation.

High power electricity transformers containing conductor windings insulated with paper impregnated with insulating oil are expected to operate for a minimum of 20 years. However, due to the effects of heat, water and oxygen the cellulose based paper is degraded over a period of time. The degradation involves reduction of the polymer molecular chain length of the cellulose which in turn reduces the

mechanical strength of the paper. The life of the transformer is partly dependent on the life of the insulator. Accordingly, an analysis of the degradation products of paper insulators can be used to monitor the degree of depolymerisation of cellulose and this provides a simple method of predicting transformer failure, as reported by Emsley AM, Stevens GC in IEE Proc. Sci. Meas. Technol., 1994, Vol. 141, No. 5 at pages 324 to 334. As is explained, 2-furaldehyde is one of the degradation products, as also are 2-acetyl furan and 5-methyl-2-furaldehyde. Other degradation products which are possible useful markers are phenol, m-cresol and xylene which are degradation products of phenol-formaldehyde insulating resins present in transformer windings. However, as a specific example of a chemical species in solution which needs to be determined and which is a marker of power transformer degradation 2-furaldehyde is taken and a description of its determination is exemplified in this specification.

It is known that 2-furaldehyde and related compounds undergo electrochemical reduction at a mercury electrode at between around -0.9 V and -1.6 V against a standard caramel electrode, depending on the value of pH. It has been conjectured that a mercury coated screen printed carbon electrode could be used for 2-furaldehyde determinations. However, it has been found that mercury deposition on printed carbon electrodes gives rise to various problems including uneven distribution of the mercury, variation in size of mercury droplets, coalescence, and various other disadvantageous effects which make it difficult if not impossible to produce a satisfactory electrode for the purpose. However, it has now been discovered that if the

carbon of the electrode contains silver particles dispersed therein the mercury tends to plate on the silver particles at the surface of the carbon rather than at the carbon surface itself, due to the silver particles being more electroactive. In consequence, the electroactive sites on the carbon are in effect better controlled, and the separation of mercury micro droplets and the percentage mercury coverage of the electrode can also be better controlled. Additionally, it is found that the maximum mercury coverage of such modified electrodes is significantly higher than that of unmodified carbon screen printed electrodes and other advantages are a greatly improved reproducibility of mercury electrode areas, and less disparity in mercury micro drop deposit sizes. Also, as compared with the possible use of mercury plated solid silver electrodes there is lower long term contamination of the mercury surface with silver ions due to the small size and the degree of separation of the silver particles.

According to the present invention there is provided a mercury-coated carbon screen printed electrode for use in the electrochemical determination of chemical species in solution wherein the carbon of the electrode contains silver particles dispersed therein.

Also provided by the present invention is a carbon screen printed electrode suitable for coating with mercury to provide a mercury-coated carbon electrode for use in the electrochemical determination of the chemical species in solution wherein the carbon of the electrode contains silver particles dispersed therein.

Additionally the present invention provides the use of such mercury-coated carbon screen printed

electrodes in the electrochemical determination of chemical species in solution, for example and in particular in the analysis of 2-furaldehyde. However, it is envisaged that other chemical species in solution can be determined using the electrodes of the invention for example vitamins e.g. ascorbic acid (Vitamin C) ; heavy metal pollutants such as lead and cadmium, and other inorganic species such as nitrates. The procedure for producing conventional carbon screen printed electrodes is well known and will not therefore be described here. The way in which the modified electrodes of the present invention can be manufactured requires an adaptation of the conventional procedures by including a silver ink in the carbon ink which is used for the printing process. This enables the printing of an electrode pad with a nearly homogeneous distribution of silver particles in the carbon. The optimum concentration of silver in the carbon is readily determined by routine experimentation. Although the subsequent formation of the amalgam on the modified carbon electrode is quite slow the mercury deposit is sufficiently stabilised at the end of the electrodeposition for immediate use as a sensor. Examples will now be described showing how carbon ink based sensors can be fabricated in accord with the present invention and also showing how the electrodes perform.

Examples Apparatus

Experiments were performed using a EG+G Model 273 potentiostat/galvanostat run by EG+G Model 270 Version 4.10 electrochemical software (EG+G Instruments, Princeton Applied Research, Princeton, NY, USA). The

electrochemical cell was a 2 ml polypropylene vial (BDH, Merck Ltd, Lutterworth, England, UK) .

Sensor Fabrication Carbon ink based sensors were fabricated using a semi-automatic screen printing machine (MPM Model TF-100, MPM, Franklin, MA, USA). Carbon/silver inks were prepared from the carbon graphite ink and Acheson Electrodag 418SS silver based polymer thick film ink (Acheson Colloids Co., Plymouth, England, UK) for approximately 33% silver content.

The reference electrode was Metech 2539 Ag/AgCl ink (Metech Polymers Corporation, Carson City, NV, USA) , the auxiliary electrode was also carbon/silver ink and the insulating ink was Metech P7192M blue dielectric.

The sensor design required four individual printing stages each requiring a separate screen. The screens were prepared by a photolithographic patterning method from photopositive images. Screen specifications depend on the nature of the inks, and the required thickness and resolution of the cured ink coating. Inks vary according to the pigment and binder. This affects solids content, viscosity and emulsion thickness. The screens chosen were of a stainless steel standard mesh with cellulose acetate backing. Mesh counts are tabulated against ink emulsion thicknesses and the ones chosen were 200 counts/inch for screens 2 and 3, 250 counts/inch for screen 1 and 280 counts/inch for screen 4. The mesh angle for each screen was 45°, and the screen thicknesses were 15 mm.

Printing was carried out on 0.5 mm thick PVC substrates, pre-baked for 1 hour at 85°C to minimise shrinkage at printed substrate curing stages.

Screen 1; Carbon/silver ink printed, cured

83°C % hour. Screen 2; Silver/silver chloride ink printed, cured 83°C hour. Screen 3; Modified carbon ink printed, cured

83°C hour. Screen 4; Dielectric ink printed, cured 83°C hour. A schematic for each printing stage and the final sensor assembly is shown in the accompanying Figure 1 in which are shown from left to right the respective screens 1, 3, 2 and 4 mentioned above and the complete assembly.

Platinum SPEs for comparison were fabricated by ESL Europe/Agmet Ltd. (Reading, UK) using a fritted platinum ink (Product 5545) printed onto an alumina substrate.

Reagents

All solutions were prepared with deionised water from a Whatman R050 reverse osmosis, ion-exchange system (Whatman Labsales Ltd, Maidstone, England, UK) . Reagents used were sodium perchlorate (BDH AnalaR. BDH) , sodium hydroxide (Rhone-Poulenc Analytical reagent, Rhone-Poulenc Ltd, Manchester, England, UK) , sulphuric acid (BDH ConvoL, BDH) , and 2-furaldehyde (Fisons 98% min, Fisons Scientific Equipment, Loughborough, England, UK) . The mercury plating solution was prepared using a 1000 ppm mercury in 1 M nitric acid standard metal solution (Fisons Scientific Equipment) .

Mercury Plating Carbon/silver ink SPEs were immersed in a 1000

ppm mercury in 1 M nitric acid/0.6 M sodium perchlorate solution, at open circuit, prior to stepping the potential to -0.8 V vs a screen-printed Ag/AgCl reference electrode using the chronoampero etry experimental option on the EG+G Model 270 Version 4.10 electrochemical software. In this way the current could be monitored against time to find the optimum plating potential and time for mercury deposition. Also, chronoamperograms could be compared for each plated electrode. Plated electrodes were viewed using an optical microscope (xl28) , through which it could be seen that the mercury deposit was in the form of tightly packed 'micro-droplets' which covered the entire surface of the electrode.

Electrodes were preconditioned prior to plating by a single CV scan between -1100 V and 800 mV at 100 mV/s in mercury plating solution.

Results and Discussion

SPE Characterisation

The Applicants' believe that by modifying the carbon ink with silver ink, better control in terms of number and distribution of equivalent active surface sites for mercury deposition is obtained. Evidence for this is found from cyclic voltammograms for carbon, carbon/silver, and platinum SPEs cycled in mercury plating solution (Figs 2-4) , and from chronoamperograms produced during mercury electrodeposition (Fig 5). Consecutive CV's for carbon/silver SPEs in mercury plating solution show the elimination of a silver oxidation peak around 300 V with an increase of a mercury oxidation peak around 200 mV, after the first cycle, suggesting complete blocking of the silver sites due to mercury

deposition. Also, the peak current density for mercury oxidation at the silver/carbon SPE is twice that for the carbon SPE. This suggests there is a higher number of active surface sites, hence the electrode is more amenable to mercury deposition.

Cycling the platinum SPE in mercury plating solution leads to the formation of a thin layer of HgPt ₂ , HgPt, and Hg ₂ Pt compounds, accounting for the oxidation peaks around 0 mV and -400 mV in Figure 4. Chronoamperograms recorded during mercury deposition for each electrode show the silver/carbon electrode reaches a steady state current in a shorter time than either the carbon or platinum electrode (Fig 5) . The current densities are much higher for the silver/ carbon electrode compared to the carbon electrode, and are close to those observed at the platinum electrode.

Mercury Coated Silver/Carbon SPE Sensor Alkaline Media CV was performed in 0.1M NaOH/O.lM NaC10 ₄ - reference solution and also 2-furaldehyde/reference solutions in the range 71-113 ppm respectively, between -1 and -1.8V at 100 mV/s. The second consecutive scan is shown for each solution (Fig 6) . 2-furaldehyde suppressed background cathodic currents up to around -1.57 V and enhanced the cathodic currents between -1.57 V and -1.8 V, with the occurrence of a peak around -1.63 V. The peak increased with increasing 2-furaldehyde and was attributed to the non reversible reduction of

2-furaldehyde to furfuryl alcohol at the mercury surface. It was noted over a period of time that the electrode response diminished and this was attributed to electroche ically generated derivatives of 2-furaldehyde which fouled the electrode surface.

Acid Media

CV was performed in 0.5 M H ₂ S0 ₄ reference solution and also 2-furaldehyde/reference solutions in the range 107-555 ppm respectively, between -0.6 and -1.6 V at 100 mV/s. The second consecutive scan is shown for each solution (Fig 7) . 2-furaldehyde did not suppress the background cathodic currents but enhanced the currents between around -1.1 V and -1.6 V, with the occurrence of a peak around -1.25 V. The peak increased with increasing 2-furaldehyde and was attributed to the non-reversible reduction of 2-furaldehyde to 2-furfuryl alcohol at the mercury surface. The electrode did not foul and could be reused after rinsing in deionised water to give similar results. A current/concentration calibration was taken from Fig 7 at -1.25 V (Fig 8).

CV is not a good choice of method for quantitative determinations since it introduces a large capacitive background current, hence, it is difficult to detect the faradaic current at lower concentrations, so, square wave voltammetry (SWV) measurements were investigated. SWV is a pulsed technique where the current response to the potential excitation is sampled once on each forward pulse and once on each reverse pulse generating three possible current potential plots; forward current vs. potential, reverse current vs. potential, or difference current vs. potential. Thus, SWV can discriminate against the charging current and eliminates the drawbacks of CV, thus allowing the detection of 2-furaldehyde at lower concentrations. Also, it is known that SWV allows voltammetric measurements to be made in the presence of dissolved oxygen. SWV was used to detect 2-furaldehyde at lower

concentrations, in the range 0.9-36.4 ppm. Difference current vs. potential curves for SWV are shown in Fig 9, for 2-furaldehyde concentrations in the range 8.9-36.4 ppm and in Fig 11 for 2-furaldehyde concentrations in the range 0.9-8.9 ppm. A current/ concentration calibration was taken from Fig 9 at -1.225 V (Fig 10), and from Fig 11 at -1.225 V (Fig 12) .

Although the foregoing description is oriented towards producing electrodes by screen printing techniques it will be apparent that it is possible that the carbon containing the dispersed silver particles may be deposited on the electrode substrate by other means, for example by means of an offset lithographic process and thus still enable the benefits of the invention to be obtained. Accordingly, it will be understood that in its broadest aspect the invention is not limited to screen-printing as the means for depositing the carbon containing the silver particles on the electrode substrate, although screen-printing is the preferred method.