This invention relates to an electrode, and particularly, but not exclusively, to an electrode for use in non-linear dielectric spectroscopy.

In our published PCT Patent Application WO92/04630, a method of analysing biological cell material, substrates therefor, or inhibitors of cell metabolism is disclosed, which method comprises applying an AC electrical potential across the biological material so as to produce a non-linear dielectric spectrum (that is, applying an AC potential at one or more discrete frequencies and determining a response at one or more second frequency absent from the applied potential). The method may be applied to a variety of biological cell suspensions including Saccharomyces Cerevisiae and Micrococcus luteus. Alternatively, as disclosed in our unpublished PCT application GB93/00458, the potential may be applied across living or viable tissue.

In this method, the system is typically excited using two outer (current) electrodes, and the response on a pair of inner (voltage) electrodes determined. The response is normally registered by taking the steady-state waveform and performing a Fourier analysis on it, to extract the frequency components. If the system of interest possesses non-linear dielectric properties, these are manifest either as harmonics of the exciting waveform (if the exciting current is essentially a pure sinusoid) or as beat frequencies (if the exciting current is at more than one frequency). The dependence of these harmonics on the voltage and frequency of the exciting waveform constitutes the non-linear dielectric spectrum, and the biophysical source(s) of the signals may be, and have been, identified by appropriate inhibitor studies, by the analysis of mutant strains, and by purification/reconstitution techniques. In this method, it is known to use gold electrodes.

A problem with this known method using gold electrodes is that cell-free supernatants also appear to generate harmonics when excited with a single sinusoid. This phenomenon is believed to be due to the non-linear interfacial current/voltage relationships of electrodes catalysing faradaic reactions, and which are manifest in electrochemical systems via the well-known technique of 2nd harmonic AC voltammetry. This generation of harmonics by the "background" solution means that, to evaluate the non-linear dielectric properties of the biological systems of interest it is necessary first to subtract those of the

supernatants (whose conductivities were well matched to that of the system of interest). Disadvantages of this known approach are:

(i) a substantial background leads a priori to a degradation in the signal: noise ratio, and (ii) possible changes in the background mean that it is necessary to take a background reading for each experiment.

Apart from the inconvenience involved, and whilst the cell supernatant is suitable for the study of microbial and other cell suspensions, there are systems, such as intact tissues, for which no suitable background exists. In such cases it is virtually impossible to obtain an accurate background reading with which to distinguish the generation of non¬ linear dielectricity by electrodes against that produced via the biological system.

According to the invention there is provided apparatus for analysing cellular biological material, which apparatus comprises: a) means for applying an AC electrical potential at one or more discrete frequencies to biological material; and b) means for determining a response at one or more frequencies which were substantially absent from the applied AC potential, characterised in that the means for applying an AC electrical potential includes at least one electrode comprising silver and/or silver halide.

Electrodes of silver and/or silver halide have both high impedance and high frequency dependence in the impedance compared with electrodes such as gold or platinum black; it was previously thought that the latter type of electrode would have been more suitable. The inventors have thus overcome a technical prejudice, which has resulted in the discovery of electrodes which are highly suitable for non-linear dielectric measurements of biological systems, since they generate minimal harmonics, for example, when immersed simply in electrolyte.

Advantageously, at least one electrode is formed from silver/silver chloride. Alternatively, at least one electrode is formed from silver.

Preferably the means for applying an AC electrical potential to a sample comprises two Ag/AgCl driver electrodes spaced apart from one another and positioned within an electrolyte and forming two outer (current) electrodes, and two inner (voltage) electrodes comprising sensor electrodes positioned between the outer electrodes; such sensor electrodes may be, for example, of gold.

Alternatively, the means for applying an AC electrical potential comprises two gold sensor electrodes spaced apart from one another and positioned within a first electrolyte, and two silver driver electrodes positioned within an outer electrolyte.

Conveniently, the outer electrlyte comprises KC1.

The apparatus according to the invention may further comprise means for comparing the response of the biological material with a stored characteristic of a determinand associated with the cellular biological material (such as, for example, a glucose concentration in living or viable cell material).

The apparatus according to the invention may be used in a method of analysing biological cell material, substrates therefor, or inhibitors of cell metabolism for the cell material, said method comprising applying an AC electrical potential at one or more discrete frequencies to a sample of the biological material, and determining a response at one or more frequencies which were substantially absent from the applied AC potential.

Alternatively, the apparatus may be used in a method of analysing or monitoring a determinand associated with cellular biological material, which comprises applying an AC electrical potential at at least one discrete initial frequency to a sample of the material; measuring a response of the material at at least one response frequency substantially not overlapping with the applied potential; and comparing the response with a stored characteristic of the determinand. In this latter embodiment the electrodes are preferably provided on an adhesive patch or the like for retention of the electrodes to the skin of a subject.

The biological cell material may, for example, be a cell suspension, or living tissue. Examples of such determinands include concentrations of oxygen, glucose, lactic acid or lactate, or the like.

The invention will now be further described by way of example only with reference to the accompanying drawings in which:

Figure 1 is a schematic representation of the generation of harmonics by known gold electrodes;

Figure 2 is a schematic representation showing the generation of harmonics by known platinum black electrodes;

Figure 3 is a schematic representation showing the generation of harmonics using silver electrodes in accordance with the present invention, wherein the electrodes are immersed in potassium chloride (KC1);

Figure 4 is a schematic representation of the generation of harmonics by silver electrodes in accordance with the present invention, in which the electrodes are immersed in potassium nitrate (KNO ₃ ) solution;

Figure 5 is a schematic representation of an electrode configuration according to a first preferred embodiment of the invention;

Figure 6 is a schematic representation of the generation of harmonics by the electrode configuration of Figure 5;

Figure 7 is a schematic representation of an electrode configuration according to a second preferred embodiment of the invention;

Figure 8 is a schematic representation showing linear impedance of terminal gold electrodes;

Figure 9 is a schematic representation showing linear impedance with two terminal platinum black electrodes;

Figure 10 is a schematic representation showing linear impedance of two terminal electrodes in accordance with the present invention;

Figure 11 is a schematic representation showing linear impedance of two terminal Ag/AgCl electrodes using the configuration of Figure 5;

Figure 12 is a schematic representation showing the linear impedance of two terminal Ag/AgCl electrodes having the configuration of Figure 7;

Figure 13 is a schematic representation illustrating the effect of the amount of AgCl deposited on the generation of harmonics by Ag/AgCl electrodes having the configuration of Figure 5;

Figure 14 is a schematic representation illustrating the generation of harmonics by Ag/AgCl electrodes having the configuration of Figure 5;

Figure 15 is a schematic illustration of the effect of conductivity on the generation of harmonics by Ag/AgCl electrodes having the configuration of Figure 7;

Figure 16 illustrates the non-linear dielectric spectra of yeast cells as measured with Ag/AgCl electrodes using the configuration of Figure 5;

Figure 17 illustrates the non-linear dielectric spectra of yeast cells as measured with Ag/AgCl electrodes using the configuration of Figure 5;

Figure 18 shows the effect of glucose metabolism on the non-linear dielectric spectra of yeast cells as measured with Ag/AgCl electrodes using the configuration of Figure 7.

Referring to the Figures, non-linear dielectric measurements were carried out using the four terminal system disclosed in our PCT patent application WO92/04630, as described above. Unless otherwise stated, spectra were averaged over ten blocks, were taken using a lOOmM KG sample (with a conductivity at the temperature used of approximately 12mS/cm) at 100Hz with the voltage across the inner electrodes set to 50mV. Silver electrodes were chlorided using a Princeton Applied Research Model 174 A Polarographic Analyser, in potentiostatic mode but at a voltage that was set to obtain a particular current. The reference electrode was Ag/AgCl, and the counter electrode was graphite.

In our PCT patent application WO92/04630, data were displayed which showed the generation of harmonics by cell free supernatants for a small number of excited voltages and frequencies. In addition it is known to display either different spectra (cell minus supernatants) or simply the power in particular harmonics within the different spectra.

Referring to Figure 1, the generation of harmonics by gold electrodes is shown. Measurements were performed using an excited applied voltage across the outer electrodes in the four terminal system herinabove described of one volt (zero-to-peak) at the frequency indicated. The power in each of the harmonics was taken across the inner electrodes. The electrodes were immersed in lOOmM KC1 with a conductivity of about 12mS/cm. Figure 1 illustrates the dependence of this known four terminal gold electrode system on exciting frequency of the power generated at each of the first ten harmonics across the inner electrodes. The exciting frequencies were spaced logarithmically, as 2.154435 ⁿ Hz (n = 0 - 9), giving four points per full decade.

Previously, the many studies of the interfacial impedance of electrodes, both from a faradaic electrochemical standpoint and from the view of their use of linear dielectric spectroscopy, have either used very low exciting voltages such that the electrical properties are essentially linear, or have simply ignored any non-linearities. The display of the data is usually in terms of the equivalent electrical circuits consisting simply of linear, passive elements such as resistors and capacitors.

Referring to Figure 2, the harmonics generated using platinum black electrodes rather than the gold electrodes of Figure 1 are shown. Similarly, Figure 3 shows the generation of harmonics using silver metal outer electrodes (but gold inner electrodes), whilst Figure 4 shows the generation of harmonics using silver electrodes emersed in a solution of potassium nitrate.

The configuration of Figure 5 comprises a pair of outer Ag/AgCl driver electrodes 2, a pair of inner gold sensor electrodes 4, spaced by electrolyte 6.

The electrolyte used in the configuration of Figure 5 is preferably potassium chloride (KCl). Substantial concentrations of Cl ^" are not compatible with all biological systems which may be of interest, and may not exist in natural environments in which it might be desired to study the generation of non-linear dielectric effects.

Figure 6 shows the generation of harmonics using Ag/AgCl electrodes in the configuration of Figure 5. In order to carry out this experiment the outer electrodes are coated by making them the anode in an electrochemical cell (with a graphite cathode containing lOOmM KCl) and passing a current of 1mA (approximately ImA/cm ² ) for 60 seconds.

A second electrode configuration is shown in Figure 7, in which the outer electrodes 10 are formed from Ag/AgCl and are each immersed in an electrode 12 comprising KCl in high concentration (3N). They are separated from the inner electrodes 14, which might contain a sample of interest, via porous glass frit 18 (LKB 4990-55). In this case, virtually no harmonics were generated by the arrangement. Indeed, at 1 Hz and at 1 kHz the largest harmonics are respectively 53 dB and at least 57 dB below the fundamental, the latter value again lying within the background. However, the magnitude of the fundamental as measured across the inner electrodes, whilst more or less frequency independent, this is subject to very substantial linear impedance, and at 10 Hz the imposition of 10 Volts (zero-to- Peak) across the outer electrodes led to the appearance of only 20mV across the inner electrodes.

Figures 8 to 12 illustrate the impedance of various electrodes described above, i.e gold, platinum black, silver and Ag/AgCl with the electrode configuration of Figure 5 and Ag/AgCl with the electrode configuration of Figure 7.

The inventors have made the surprising discovery that the electrodes which have both the lowest impedance and the greatest frequency dependence in the impedance,

na ely gold and platinum black, and which have been previously expected by those skilled in the art to be the most suitable for non-linear dielectric measurements, in fact generate the greatest harmonics. Conversely, of those surveyed, those electrodes with the greatest linear impedance are in fact the most suitable for non-linear dielectric measurements of biological systems, since they generate the smallest level of harmonics when immersed simply in electrolyte.

This discovery is in direct contradiction with the general prejudice existing in this area of technology.

In terms of the stability of Ag/AgCl electrodes over time when immersed in lOOmM KCl, the lack of generation of harmonics was stable to within 2 dB over a period of at least a day, typically deteriorating over a period of some 10 days.

Figure 13 illustrates the effect of the amount of AgCl deposited on the electrodes. The layer was deposited with a Princeton Applied Research Model 174 A Polarographic Analyser, the counter electrode was graphite, the reference electrode was Ag/AgCl, and the silver electrode to be coated was the working electrode. Spectra averaged over 10 blocks were taken using a lOOmM KCl sample (conductivity approx 12mS/cm) at 10Hz, with the voltage across the inners set to 50mV. Figure 13 shows that the harmonics reduce with an increasing thickness of the AgCl layer until some 600 C/m ² of charge have been passed, after which there is no further advantage. Visually this corresponds to a thin dark grey/brown layer. This layer must be complete (and the electrodes must be initially cleaned - preferably by abrasion) and the charge must be supplied at a sufficiently low current density to prevent bubbling at the electrodes, 1-2 mA/cm ² being fine for these purposes.

Similar experiments with the electrode configuration of Figure 7, which was also time-stable, indicate that the best results (in terms of lowered harmonics) can be obtained by cleaning the outer electrodes by mild abrasion and using them with no predeposition of AgCl. A sufficient layer of AgCl builds up automatically under the AC signal, presumably due to rectification at the interfaces, and is in fact just visible as a light grey patina. An Ultra-Low-Frequency sweep (Fig 14) from 1-10 Hz (linear) at 8V (-80mV across inners) and with an electrolyte of lOOmM KCl Shows that the harmonics drop below background (at - 55dB) at ca.4Hz. The conductivity of the sample in the electrode chamber (in the range 3-15 mS/cm) has no noticeable effect on the levels of the harmonics (not shown).

The effect of the conductivity of KCl in the outer electrode channels of the

configuration of Figure 7 on the generation of harmonics is shown in Figure 15. This shows a small but noticeable improvement to better linearity as the strength of the KCl around the outer electrodes is increased, presumably due to a continuing thickening of the AgCl layer. Data on the non-linear dielectric spectra of yeast cells was measured with Ag/AgCl electrodes (in the configuration of Figure 5; similar data were observed with the configuration of Figure 7, as shown in Figures 16 and 17). Figures 16 and 17 respectively show the voltage and frequency windows of a suspension of yeast cells (minus their supernatant), illustrating that very clean non-linear dielectric spectra may be obtained, whose features are qualitatively similar to those observed previously but with a much improved signal:noise. The effects of glucose on the non-linear dielectric spectra of yeast cells (with the configuration of Figure 7) are shown in Figure 18 which shows the time-dependence of the non-linear dielectric spectra, indicating that some 10 minutes after the addition of glucose the third harmonic is replaced predominantly by a second harmonic, and that after the glucose has been consumed the third harmonic reappears. This suggests that the apparatus according to the invention may be useful as a non-invasive sensor for glucose or other metabolic substrates.