WO1995004272A1

WO1995004272A1 - Reference electrode

Info

Publication number: WO1995004272A1
Application number: PCT/DK1994/000290
Authority: WO
Inventors: Thomas Buch-Rasmussen
Original assignee: Novo Nordisk A/S
Priority date: 1993-07-28
Filing date: 1994-07-19
Publication date: 1995-02-09
Also published as: AU7226994A

Abstract

By determining the concentration of a component in a halide, sulphide or halogenoid containing sample using an electrode unit (1) comprising a silver electrode (2), a working electrode (3), and a counter electrode (4), which electrodes show up at an end of the unit made from an electrically insulating material, the silver electrode being in contact with the sample is transformed into a reference electrode by being coated with a silver halide, sulphide or halogenoid by applying a potential difference over the silver electrode (2), and one of the further electrodes (3, 4) the potential difference having a polarity that causes anodic oxidation of the silver electrode.

Description

REFERENCE ELECTRODE

The present invention relates to a process for preparing in an electrode unit comprising a silver electrode and at least one further electrode, a reference silver electrode to be used by determining the concentration of a certain component in a halide, sulphide, or halogenoid, e.g cyanide, containing sample.

International patent application No. PCT/DK90/00067 - WO 90/10861 relates to a method for measuring the concentration of a constituent in a fluid sample by means of an electrode device comprising a silver halide reference electrode and at least one measuring electrode. The electrodes are embedded in a plastic material to form an elongated electrode unit wherein the electrodes as longitudinally threads stretch in the axial direction of the electrode unit and at one end thereof is connected to an electric circuit. The other end of the electrode unit presents a surface for receiving a sample to be measured and in which surface end the end surfaces of the filamentous electrodes are free to be in electric contact with the sample. The constituent to be measured is glucose in a blood sample. The glucose content in a blood sample can be determined by measuring the current flowing between the measuring electrode and a counter electrode maintained at the reference electrode potential when a certain potential difference is applied over the measuring electrode and the counterelectrode.

An important asset of silver-silver halide electrodes is that they are small, compact, and can be used in any orientation.

The thermodynamic principles of this reference electrode system may be illustrated by the most familia example, the silver-silver chloride electrode:

Ag | AgCI | CI^" which consists of solid silver chloride on silver and in contact with a solution of a soluble chloride salt. In order to keep the activity of the silver chloride in the solution constant, the solution is to be saturated with silver chloride.

According to the above PCT-application a thin slice is cut from the sample receiving end of the electrode unit to lay free a new end part of the electrodes each time a new sample is measured. Thereby, a pure silver surface is laid open at the end of the reference electrode, which is a silver electrode coated by a layer of silver chloride. It is relied on that the free silver surface is very quickly coated by a thin layer of silver chloride formed by reaction between the silver and chloride in the sample. The chloride in the sample may arise from dissolution of silver chloride from the periphery of the silver chloride coated wire and/or from the content of chloride in the sample.

One object of this invention is to prepare reference electrodes which are stable and reproducible.

By a silver electrode in an electrode unit to be used in a halide, sulphide or halogenoid containing sample a stable silver reference electrode may according to the invention be obtained by the free end of the silver electrode which is in contact with the sample being coated with a silver halide, sulphide, or halogenoid in the presence of the sample.

According to this invention this in situ coating of the silver electrode to make it appear as a silver reference electrode is obtained by applying a first potential difference over the silver electrode and a further electrode in the electrode unit, the potential difference made so that it causes anodic oxidation of the silver electrode. According to this invention the reference electrode is prepared immediately before, i.e. less than one hour, preferably less than one minute before, and more preferably less than 20 seconds before the electrodes of the electrode unit are biased to perform the measurement serving the determination of the concentration of said component in the sample. In this way it is ensured that the reference electrode does not chance its performances due to storing.

After the anodic oxidation of the reference electrode it has shown to be appropriate to perform a succeeding cathodic reduction of part of the precipitated silver halide, sulphide, or halogenoid. To obtain this cathodic reduction, the first potential difference may be followed by a second opposite potential difference, the voltage and the duration of this second potential difference being so controlled, that the amount of charge, i.e. the number of Coulombs, transmitted by the first potential difference exceeds the amount of charge transmitted by the second potential difference.

An appropriate relation between the anodic oxidation and the cathodic reduction is obtained, when the amount of charge transmitted by the second potential do not exceed 50% of the charge transmitted by the first potential difference.

According to this invention, the working electrode may be an enzyme electrode containing an enzyme causing an oxidation of the component, the concentration of which is to be measured.

The component to be measured may be glucose, preferably present in a blood sample.

The invention will now be further described with reference to the drawing wherein: Fig. 1 shows the end surface of an electrode unit, i.e. the end on which a sample is placed to be examined,

Fig. 2 shows a cyclogram of applied voltage (1 ) and current (2) during the silver wire electro-chemical pretreatment (left) and silver-silver electrode potential changes as measured after the electrodes pretreatment (right),

Fig. 3 shows kinetics of the silver-silver chloride potential creation in 0.1 M sodium phosphate pH 7.0, 100 mM sodium chloride solution.

Fig. 4A shows the stability of in situ silver-silver chloride electrode potential at 25°C in pH 7.0 0.1 M sodium phosphate buffer, 0.1 M sodium chloride.

Fig. 4B shows the reproducibility of in situ silver-silver chloride electrode potential at 25°C in pH 7.0 0.1 M sodium phosphate buffer, 0.1 M sodium chloride.

Fig. 5 shows the in situ silver-silver chloride electrode potential (vs potential in buffer solution) in solutions comprising miscellaneous physiological compounds.

This drawing is not to be interpreted as limiting this invention. E.g. the number of electrodes in the unit may differ from the number shown in the drawing.

In the following the oxidation of the reference electrode will be described as halogenation and the coating as a halide, although a sulphide or a halogenoid may be precipitated instead of a halide. However, where the thing to be measured is the glucose content in blood, the reference electrode will usually be halogenated using the halide natural present in a blood sample.

An electrode unit comprises a cylindrical, rod-shaped electrode body member. An end surface of such an electrode unit is shown in figure 1. In this surface an end of the electrodes in the unit are laid open to be contacted by an aqueous sample placed on this surface. The electrode body member 1 may be made from an electrically insulating material such as a plastic material. The electrodes comprising a reference electrode 2, a working electrode 3, and a counter electrode 4 are embedded in the electrode body member and stretch longitudinally through the electrode body member from the shown end surface thereof. At the other end of the electrode unit the different electrodes are connected to an electronic circuit providing the different potential differences between the electrodes and measuring the current through the electrodes.

In this example the reference electrode is a silver electrode which is in situ prepared as a halogenated reference electrode as it will be described below. The working electrode may be an enzyme electrode of the kind comprising an enzyme oxidizing the component the concentration of which is to be determined by measuring the current provided by the electrons splitted off by the enzymatic oxidation of said component at the working electrode 3. In the principle, a working electrode and a reference electrode would be sufficient to establish a measuring circuit, but as the precise function of the reference electrode is conditional on no current running through the reference electrode, it is preferred to have a counter electrode which together with the working electrode forms the measuring circuit. The reference electrode only serves the provision of a reference potential in relation to which the potentials of the working electrode and the counter electrode are set to obtain well defined measuring conditions. The reference electrode is of the type comprising a silver electrode coated by a thin layer of a silver halide where the halide ions are present in the sample. The known reference electrodes are coated by the silver halide before they are brought into contact with the sample.

By the reference electrode according to WO 90/10861 the end of the electrode unit is cut off immediately before a measuring is performed relying on a halide layer being established on the pure silver surface laid open by this cutting. This halide layer is established partly by silver halide from the edge of the electrode being dissolved in the sample and precipitated on the pure silver surface and partly by this surface being oxidized by the halide in the sample, that is by simple diffusion conditional chemical processes.

The reference electrode according to the invention is in its initial condition a pure silver electrode, i.e the surface coming into contact with the sample is a silver surface which, however, may be plated on another material. When a sample is placed on the end surface presented by the electrode unit as shown in figure 1 it covers the end surfaces of the electrodes of this electrode unit establishing a liquid current path between these electrodes. By applying over the reference electrode 2 and another electrode, e.g. the counter electrode 4, a voltage with such a polarity that it causes anodic oxidation of the reference electrode 2, this electrode will be oxidized by the halide in the sample to provide a halogenated reference electrode. The amount of halide generated may be measured by measuring the charge transported through the circuit to provide the anodic oxidation. This measurement is important as it is proved useful to succeed the oxidizing potential difference by a second potential difference with an opposite polarity causing retransformation into silver of some of the halide generated at the reference electrode surface . Of course only part of the of the halide generated should be reduced back into silver and this may be obtained by controlling the amounts of charge transported in the first and the second direction in the reference electrode preparing circuit.

This unsymmetrical charge transportation is obtained by making the first applied potential difference larger and of a longer duration than the second applied potential difference.

As an example in a chloride containing sample the first potential difference may be 2 volts for 10 seconds succeeded by a second potential of -0.5 volts for 5 seconds. Integration of the time/current graph representing the charge transmission have proved that by the above mentioned potential differences and durations the charge transported during the reduction of the reference electrode is about 30% of the charge transported during the oxidation of this electrode. Herby a stable silver chloride layer is formed on the part of the reference electrode coming into contact with the sample and an reference electrode providing a stable reference potential for a long time is provided.

In the following is described experiments estimating the performances of a reference electrode prepared in situ.

EXPERIMENTAL PART

Electrodes, Apparatus and Procedures

Two types of electrodes were used for in situ silver-silver chloride electrode preparation. The electrode potential optimization was done employing electrodes with two silver wires (thickness 0.125 mm) glued by an epoxy compound into two channels (i.d. 1.5 mm) of rods (o.d. 6 mm; length 4 cm) prepared from high density polyethylene. The electrodes for physiological solutions were formulated using silver wires of thickness 0.20 mm and polyurethane glue.

All in situ electrode preparations and potential measurements were done using new renovated surfaces of silver wires obtained by polishing on 250 mesh emery paper or (in the study of physiological solutions) by cutting slices (thickness 0.3 mm) of polyethylene rod together with silver wires and glue using a special device with a stainless steel knife.

The electrolysis was carried out and the electrode potential was measured using a computerized self made potentiostat and the Fluke 45 Dual Display Multimeter (John Fluke MFG. Co. NC). The program is written in "C" for IBM compatible computers. A/D converter resolution is 16 bits, integration time of sampling is 20 ms.

The potential of the in situ prepared silver-silver chloride electrode was investigated by thermostating the electrode, buffer solution (100 ml) and a salt bridge in a thermostat (HETOFRIG, Denmark) chamber. The salt bridge was 3 cm length, 0.3 cm i.d. silicone rubber tube filled with buffer solution. The saturated calomel electrode (SCE) (K401 , Radiometer, Denmark) temperature (24.6°C) in these measurements was kept constant using a polyurethane container as thermoinsolator.

The stationary silver-silver chloride electrode potential dependence on temperature was measured using the same thermostat at isothermal conditions. The 100 ml of buffer solution, stationary silver-silver chloride (E9001 P0₂ electrode, Radiometer) and SCE (K401 , Radiometer) were thermostated at fixed temperature for at least 1 h. The temperature in the buffer solution was controlled using a KEITHLEY 871 Digital Thermometer. After establishment of stable potential (+/- 0.02 mV) 5-6 potential values were recorded with intervals between measurements of 10-12 seconds and the average potential value was used for calculations. The action of chloride ions on the in situ silver-silver chloride electrode was investigated at the same conditions at 25°C.

The preparation in situ of silver-silver chloride electrodes in phosphate buffer solution pH 7.4 comprising 0.1 M sodium chloride and in physiological solutions was done using a drop (ca 0.03 ml) of mixture. The electrode was kept in vertical position and the electrode pretreatment started immediately after the application of a drop to the electrodes at room temperature.

The silver-silver chloride electrodes were prepared in situ by sequentially applying anodic (2 V) and cathodic potentials (-2, -1 and -0.5 V) to two silver wires. In the experiments of the electrode potential optimization, the anodic potential was applied 5 s. The cathodic potential and relative charge (in comparison to charge during the anodic process) were varied in the intervals from -2 V to -0.5 V and from 10 to 90%, respectively. A typical cyclogram of electrode pretreatment is presented in Fig. 2. where applied voltage (1) and current (2) during the silver wire electro-chemical pretreatment are shown to the left and silver-silver electrode potential changes as measured after the electrodes pretreatment are shown to the right. t_b and t_e indicates the beginning and the end time of potential measurement. The kinetics of the electrode potential change was measured immediately after the cathodic procedure. For characterization of the electrodes potential stability it was measured after 2, 5 and 10 s of delay from the beginning of measuring time (t_b) and after 30 s or 45 seconds (t_e). The stability of the potential at these times (E_b and E_e, respectively) was calculated using the expression: (Σ(E_b - E_e)²)^'Λ*100/Σ E_e, %. The number of measurements in the optimization procedure was typically 10. Twenty parallel measurements were performed in biological solutions. Ten parallel measurements were done using venous blood. The action of different compounds was investigated in three parallel measurements. The relative standard deviation of E_e was used for electrode potential reproducibility characterization.

Buffer Systems Used

The measurements of electrode potential reproducibility, stability and thermal effects were performed in 0.1 M sodium phosphate buffer solution pH 7.0 containing 0.1 M sodium chloride. Chloride ion action was investigated in the same buffer solution comprising additionally 20 mM potassium chloride.

As physiological solution 0.1 M sodium phosphate buffer solution pH 7.4 containing 0.1 M of sodium chloride was used. Polyethylene glycol solution comprises additionally 2% of poly(ethylene glycol) - PEG. As plasma a dried samples of K-89E (Danish Society of Clinical Chemistry) diluted with 10 ml of water was used. Albumin solution was prepared by diluting 20% Albumin Nordisk

(Novo Nordisk A/S) with buffer solution. The final albumin concentration was 5%. As artificial blood two mixtures were prepared prior to measurements. One mixture (blood-a) was prepared by diluting 5 ml of SAG-M with 5 ml of plasma

(Novo Nordisk A/S). The glucose concentration in this mixture was 8.1 mM, haematocrit was 0.3. Another mixture (blood-b) was prepared following a standard laboratory method which includes centrifugation of SAG-M and mixing of erythrocytes with plasma. The final haematocrit of this mixture was 0.46, and the glucose concentration was 4.1 mM at the beginning and 3.8 mM after 4 h of experiment. Samples of capillary blood were supplied from volunteers. Chemicals

Silver wires of purity 99.9% and thickness 0.125 mm (type "soft") and 0.2 mm (type "hard") was purchased from Dansk Hollandsk Edelmetal. Buffer components, sodium and potassium chloride and physiological compounds were of analytical grade. The molecular weight of the PEG used was 20 kDa. All chemicals were used as received.

RESULTS AND DISCUSSION

Reproducibility and Stability of Silver-Silver Chloride Electrode Potential

In buffer solution, when a potential of 2 V was applied for 5 s between two silver wires, the silver electrode potential approached 65 mV and decreased with time (Fig.3). The sequential application of symmetrical potentials of 2 V and -2 V to the electrodes decreases the initial electrode potential to 37 mV, but it is not stable at the beginning of measurement. The electrode potential stability increases when anodization and cathodization is carried out using non-symmetrical potentials of 2 and -0.5 V, respectively (Fig.3).

For optimization of the reproducibility and the stability of electrode potential 3 parameters were varied: The cathodic potential(CP), the charge of the cathodic process (CC) and the delay time from the end of the cathodic process to the electrode potential measurement (DT).

From the experimental results it follows that the electrode potential stability and reproducibility changes in the range 0.5-3.1 and 0.15-1.4%, 0.17-0.60 and 0.19- 1.5% and 0.19-0.64 and 0.15-1.6% at 2, 5 and 10 s delay, respectively (Figs 4 A,B). The best electrode potential stability (0.19-0.21%) was indicated at 10 s of delay and 10-90% of charge. The electrode potential reproducibility practically does not depend on cathodic process potential and delay time, but it is sensitive to charge; the best reproducibility (0.15-0.27%) was obtained at CC 30%. At optimal conditions, when the multiplied stability and reproducibility parameters have minimum values (CP -0.5 V, CC 30%, DT 10 s), the prepared in situ electrode potential was 45.0 +/-0.1 mV. The best reproducibility of the potential can be achieved at CP -2 V using the same CC and DT. However, in this case the stability of the electrode is less and the potential is 43.35+/-0.06 mV.

The potential of a silver-silver chloride electrode prepared in situ is determined by the silver ion activity (a^ . Following the Nernst equation the electrode potential can be expressed:

E = E ^ + RT/F In a_Ag+

It is evident that all factors that influence on the silver ion activity at the electrode surface will change the electrode potential.

When a potential of 2 V is applied to two silver wires the anode is covered by silver chloride due to reaction:

Ag + CI^"— > AgCI + e

On the other electrode, oxygen reduction and hydrogen liberation proceed:

0₂ + 2 e + 2 H⁺ — > H₂0₂

H⁺ + e --> ¹/₂ H, In a very simplified analysis the accumulation of silver ions near the anode surface cause a high electrode potential at the beginning of measurements and it decreases in time due to ion diffusion into solution and reaction with chloride ions. The subsequent application of a cathodic potential causes reduction of silver ions and silver chloride and generates a naked metallic silver surface:

AgCI + e — > Ag + CI^"

Concurrently, on the anode predominantly silver oxidation takes place as described previously. Due to the high rate of silver dissolution by anodic process the applied potential mainly drops on the cathode where the additional process of water reduction can take place. Possibly, the last reaction and hydrogen accumulation is the reason for the lower electrode potential stability when -2 V is applied to the electrodes. The small dependence of the electrode potential reproducibility on cathodic potential means that current density and time of dechlorination are not critical. The main factor is the ratio between oxidation and reduction charges.

The correspondence of the silver-silver chloride in situ electrode potential to previous findings can be established using the potential (at 25°C) of a saturated silver-silver chloride electrode (197 mV vs NHE), the SCE potential (244 mV vs NHE) and the 0.1 M KCI calomel electrode potential (337 mV vs NHE). From these values follows that the potential of a 0.1 M KCI silver/silver chloride electrode should be 290 mV vs NHE (46 mV vs SCE). The experimentally determined value of our in situ silver-silver chloride electrode potential is close to this value. Temperature Dependence of Silver-Silver Chloride Electrode Potential

The temperature dependence of the in situ prepared silver-silver chloride electrode was correlated with same dependence of a stationary silver-silver chloride electrode. In contrast to the saturated silver-silver chloride electrode, data concerning the temperature dependence of a 0.1 M sodium chloride silver-silver chloride electrode potential have not been reported. For this reason the stationary silver-silver chloride electrode potential dependence on temperature was investigated in 0.1 M sodium phosphate buffer solution comprising 100 mM sodium chloride. It was shown that in isothermal conditions the potential determined against SCE of the stationary silver-silver chloride decreases when temperature increases. In the interval of 15-35°C the potential change was 7.4 mV (Table 1).

Table 1

The dependence on temperature of the stationary silver-silver chloride potential vs SCE (ΔE) and half cell potential (E) of silver-silver chloride electrode in 0.1 M sodium phosphate buffer solution pH 7.0 containing 0.1 M sodium chloride.

Temperature, °C ΔE (+/-sd), E, mV vs SCE mV vs NHE

15 44.70 (0.02) 294.0

20 46.33 (0.02) 292.9

25 48.40 (0.01) 292.2

30 50.29 (0.03) 291.3

35 52.11 (0.03) 290.4 Due to isothermal conditions, the potential change is a result of potential changes of both the silver-silver chloride and the SCE electrodes. To calculate the half-cell potential of the silver-silver chloride electrode at different temperatures, the potential of SCE at the same temperature was added. For calculations an equation was used for the temperature dependence of the SCE potential:

E (mV) = 243.8 - 5.5 10^'1 (t - 25)

where E (mV) - SCE half-cell potential vs NHE, t - temperature in °C.

The silver-silver chloride electrode potential temperature dependence can be approximated by equation (R = 0.9984):

E (mV) = 296.0 - 0.205 t + 5.7 10^"4 f

From this equation it follows that in the interval 15-35°C the potential of the stationary silver-silver chloride electrode in 100 mM sodium chloride solution decreases in average 0.18 mV when the temperature increases 1°C, and the temperature coefficient at 25°C was 0.18 mV/°C. The determined silver-silver chloride electrode potential dependence on temperature is close to the 0.1 M calomel electrode dependence. However, the temperature sensitivity of SCE is higher; it changes 0.55 mV/°C.

The potential of the in situ silver-silver chloride electrode prepared using optimized conditions was more sensitive to temperature than the potential of the stationary electrode. In the interval 10-35°C it decreases 7.5 mV and the temperature coefficient was 0.3 mV/°C. The higher temperature sensitivity indicates some non-stationary state of the silver electrode interface in reliable concordance with the nature of the electrode potential. Silver-Silver Chloride Electrode Potential Dependence on Chloride Ion Concentration

The influence of chloride ions on the electrode potential is important for the silver- silver chloride electrode application in biological solutions. Adding potassium chloride into the buffer solution changes the stationary electrode potential (Table 2).

Table 2

Stationary silver-silver chloride electrode potential dependence on potassium chloride concentration at 30°C in 0.1 M Na-phosphate buffer pH 7.0, containing 0.1 M of sodium chloride and 20 mM of potassium chloride

ΔC.mM ΔE_det, mV ΔE^, mV ΔΔE, mV

0 0 0 0

3.9 -0.81 -0.835 0.025

7.8 -1.67 -1.644 -0.026

11.7 -2.40 -2.428 0.028

15.6 -3.15 -3.190 0.040

19.5 -3.91 -3.930 0.020

23.4 -4.67 -4.650 -0.020

As can be seen from Table 2, addition of 23.4 mM KCI to the phosphate buffer solution containing a fixed concentration of chloride ions decreases the potential of the electrode 4.67 mV, i.e. 0.202 mV/mM in average. In the same chloride ion concentration range the potential of the in situ electrode prepared using optimized pretreatment conditions changes in average 0.19 mV/mM. At physiological conditions the chloride concentration varies 9 mM (95% range). Such a variation of chloride concentration can change the silver-silver chloride electrode potential 1.7 mV.

The experimentally determined electrode potential change was compared to the prediction of the modified Nernst equation:

E = E⁰^ + RT/F In K^,

where K. - AgCI dissociation constant, a_cι. - chloride ion activity

At constant temperature and relatively low concentrations of chloride and other ions the transformed Nernst equation involving the concentration instead of the ion activity can be applied to electrode potential calculation:

ΔE (mV) = RT/F [ln(120) - ln(120 + ΔC)]

where RT/F at 30°C is 26.1 mV.

The close correlation between determined and calculated electrode potential dependence on chloride ion concentration is clearly demonstrated by Table 2.

Electrode Potential in Physiological Solutions

The in situ silver-silver chloride electrode potential in phosphate buffer solution pH 7.4 containing 0.1 M sodium chloride is similar to that in a solution with pH 7.0 (Table 3). The reproducibility and stability of the electrode potential, when only a drop of buffer solution is applied to the electrodes, changes insignificantly, too. Table 3

The potential of in situ silver-silver chloride reference electrode in miscellaneous media at 22.9°C.

Conditions of in situ chlorination Electrode Potential

Media potential stability

Chlorination, Dechlorination, (+/-sd), %

V; s V; % mV vs SCE

Buffer solution 2; 5 -0.5; 30 46.0 (0.2) 0.15

PEG solution 2; 5 -0.5; 30 42.9 (0.3) 0.14

Albumin solution 2; 5 -0.5; 30 46.6 (0.5) 1.26

Plasma 2; 10 -0.5; 30 46.8 (0.3) 1.54

Plasma 2; 5 -0.5; 50 45.7 (0.8) 1.27 Blood-a 2; 5 -0.5; 30 46.2 (1.4) 1.10

Blood-b 2; 5 -0.5; 30 44.3 (0.6) 0.75

Capillary blood 2; 5 -0.5; 30 43.7 (0.8) 3.7

The compounds which were selected as potential interferant to the in situ reference electrode potential were such which exist in physiological solutions at rather high concentrations and which can react electrochemically or chemically with the silver ions. Among the investigated substances only reduced glutathione, paracetamol and uric acid interfere with the electrode potential (Fig.5). In figure 5 the following abbreviation is used for (concentration in mM) of compounds: ASC (0.1) - ascorbic acid, ASP (0.2) - aspirin, CYS (0.033) - L-cysteine, GLU (1.0) - reduced glutathione, HIS (0.085) - L-histidine, LYS (0.15) - L-lysine monohydrat, PAR (0.2) - paracetamol, SAL (0.02) - salicylic acid, URE (4.5) - urea, URI (0.27) - uric acid. The most significant was the action of reduced glutathione. The small increase of electrode potential after the short delay time in the case of paracetamol and uric acid indicate electrolysis of these compounds at high electrode potential during the anodization process followed by product adsorption on the electrode. However, the adsorption is reversible and the electrode potential approaches that in buffer solution fast. In the case of the glutathione the initial potential of the electrode was close to that in buffer solution. This indicates that the electrode potential initially is determined by the silver-silver chloride couple. The decrease in the 50 s period of the electrode potential to 36.8 mV (vs SCE) can be caused by glutathione reaction with silver ions. It was demonstrated that the potential of a silver electrode being treated by hydrogen sulfide is about 36 mV (vs NHE). To liberate completely glutathione in blood, hemolysis of erythrocytes must be performed. This indicates that the glutathione concentration in plasma is low and thus the interference to the electrode potential in fresh blood should be unimportant. The insignificant interference of L-cysteine to electrode potential can be explained also by its low concentration in the samples.

PEG was added to the buffer solution to compare the potential in buffer solution with more viscous physiological solutions. The results presented in Table 3 show that the addition of 2% PEG to buffer solution decrease the electrode potential by 3.1 mV. This indicates a complexation of silver ions rather an effect of viscosity changes.

The electrode potential in albumin solution and in plasma are close to buffer solution containing the same amount of chloride, but its reproducibility was less than in the PEG solution (Table 3). The major difference between PEG solution and plasma is the electrode potential stability. The conditions of in situ chlorination changes this parameter little. It was noticed that the immense potential change at the beginning of measuring and at the end of measuring time stabilizes.

In blood-a holding a large amount of diluted SAG-M the reproducibility of electrode potential was 3%, and the stability was very similar to that in albumin solution and plasma (Table 3). In blood-b which was prepared from plasma and erythrocytes the electrode potential reproducibility was 1.4% and the stability was better than in plasma. In both blood samples the potential changes at the beginning of measurements and later stabilizes as in the case of plasma.

The decrease of electrode potential reproducibility and stability in blood samples can be caused by physical and chemical action. Water evaporation from the applied drop and the consequent chloride concentration increase and decrease of temperature can be factors which influence the statistical parameters. Among chemical factors the adsorption of native and denaturated proteins on the silver electrode can be dominating. It was shown that the adsorption of proteins on silver proceeds by the formation of coordination bonds between silver or silver ion and amino acid residues containing hetero atoms. However, it seems that the main factor which determines the rather low reproducibility and stability of the electrode in clotted or simulated blood is adsorption of denaturated protein agglomerates on the electrode. This conclusion follows from the fact that the lowest potential reproducibility was determined in diluted SAG-M, in which the amount of denaturated proteins is the largest, and that the agglomerates of denaturated protein adsorb on silicon surfaces very strongly. In "old" or artificial blood with erythrocytes reduced glutathione can be liberated into the plasma and cause some interference to the electrode parameters, too.

In capillary blood the reproducibility of the in situ electrode potential was 1.2, 1.3 and 1.8% at 5, 20 and 50 seconds of delay. The stability of potential during the measurement time (30 s) was 3.7% in average. It seems that the coagulation is important to establish reliable reproducibility and stability of electrode potential. However, during the short time of measurement (30 s), when coagulation essentially does not proceed, the in situ electrode potential reproducibility and stability satisfy analytical requirements.

In conclusion, the procedure described in this study provides direct evidence that by using a rapid and simple way of electrochemical silver electrode pretreatment, it is possible to prepare in situ silver-silver chloride electrodes that show improved reproducibility and stability of the potential. We believe that such electrodes can be prepared from metallic silver or silver in different compositions, materials of different sizes, in water, organic solvents or in melted salts containing chloride, bromide, iodide or sulphide ions. The suggested method of the reference electrode preparation is suitable for the electrochemical measurements, especially for electrodes with renewable surfaces, for measurements in vivo and for measurements with microelectrodes. The practical conditions under which the solid-state reference microelectrodes potential is expected to be invariant to surrounding were analyzed very recently by Collins.

Any novel feature or combination of features described herein are considered essential to this invention.

Claims

1. A process for preparing in an electrode unit comprising a silver electrode and at least one further electrode, a reference silver electrode to be used by determining the concentration of a certain component in a halide, sulphide, or halogenoid containing sample characterized in that the free end of the silver electrode which is in contact with the sample is coated with a silver halide, sulphide, or halogenoid in the presence of the sample.

2. A process according to claim 1 , characterized in that the coating of the silver electrode to make it appear as a silver reference electrode is obtained by applying a first potential difference over the silver electrode and a further electrode in the electrode unit, the potential difference having a polarity that causes anodic oxidation of the silver electrode.

3. Process according to claim 2, characterized in that during the reference electrode preparation the first potential difference is succeeded by a second potential difference with an opposite polarity causing cathodic reduction of the silver halide, sulphide or halogenoid caused by the anodic oxidation, the voltage and duration of the potential differences being controlled so that the amount of charge transmitted by the first potential difference exceed the amount of charge transmitted by the second potential difference.

4. Process according to claim 3, characterized in that the amount of charge transmitted by the second potential difference do not exceed 50% of the charge transmitted by the first potential difference.

5. Process according to anyone of the preceding claims characterized in that the working electrode is an enzyme electrode containing an enzyme causing the substance, the concentration of which is to be measured, to be oxidised.

6. Process according to claim 5, characterized in that the component to be measured is glucose, preferably present in a blood sample.

7. Process according to anyone of the preceding claims, characterized in that the coating of the reference electrode is provided less one hour, preferably less than 5 minutes, more preferably less than 1 minute, and most preferably less than 20 seconds before the electrodes of the electrode unit are biased to perform the measurement serving the determination of the concentration of said component in the sample.