CA2783400A1 - Calibration of electrochemical sensor - Google Patents

Abstract

This invention is concerned with electrochemical methods of analysing a fluid for an analyte. It utilises a sensor comprising electrodes and a mediator compound capable of undergoing a cycle of oxidation and reduction in response to varying potential applied to the electrodes, where the compound is also able to undergo a portion of the cycle through chemical reaction with the analyte. As is already known, the sensor is used in a method in which the sensor and the fluid are brought into contact sufficiently for the analyte to react chemically with the mediator, and in a voltammetry procedure, potential is applied to the electrodes while observing current flow with sufficient time for reaction between the mediator compound and the analyte, for thereby observing the concentration of the analyte. The novel characteristic of this invention is a step of calibrating the sensor by applying a rapidly changing potential to the electrodes and observing current flow, the change of this potential and observation of current being sufficiently fast to allow and observe electrochemical oxidation and reduction of the mediator taking place independently of the concentration of analyte, thereby observing the concentration of the mediator compound. The rapid change of potential maybe provided by pulse voltammetry or square wave voltammetry.

Description

Calibration of electrochemical sensor Field of the Invention This invention is concerned with electrochemical analysis of fluids. It may be applied to fluids encountered in many locations but it has particular applicability in connection with electrochemical determination of hydrogen sulfide and dissolved sulfide anions at a location below ground.

Background The analysis of fluid samples from hydrocarbon wells is a significant step in the evaluation of the producibility and economic value of the hydrocarbon reserves. An important factor in determining the economic value of gas and liquid hydrocarbon reserves is their chemical composition, particularly the concentration of gaseous components, such as hydrogen sulfide, carbon dioxide and methane. Therefore, real time gas detection is an important process for downhole fluid analysis.

The presence of sulfide species, principally hydrogen sulfide (H2S) but possibly also encountered as dissolved sulfide anions, has an important impact on the economic value of the produced hydrocarbons and the cost of production operations. Typically, the sulfur content of crude oils is in the range 0.3-0.8 weight percent and the H2S content of natural gas is in the range 0.01-0.4 weight percent, although concentrations of 1-12S in natural gas,of up to 30 weight percent have been reported.

There is a desire to be able to measure concentration of sulfide species in fluids downhole, rather than capturing a sample downhole and transporting it to the surface for laboratory analysis. One incentive for measurement downhole is to avoid any effects of change in temperature and pressure during travel to the surface with the risk that the composition analysed at the surface is not representative of the composition downhole.
Another incentive to carry out measurement downhole is to avoid the delay between collecting a sample downhole and receiving analytical results from ,a surface laboratory facility. However, elevated temperature and pressure downhole, possibly accompanied by corrosive materials, present a challenging environment for an analytical sensor.
There have been a number of proposals for the determination of sulfide species by electrochemical techniques. These include proposals in which electrochemistry is coupled through a mediator compound to sulfide which is the intended analyte.
This mediator compound is present in an electrochemical cell which is exposed to the sulfide.
Both in the presence and absence of the sulfide analyte, an electrochemical oxidation and reduction of the mediator compound can take place when appropriate electrical potential is applied to the electrodes. However, one of the redox reactions of the mediator compound can also be brought about through a chemical reaction with the sulfide, and when this takes place there is a measurable change to the electrochemistry.
A number of compounds including ferrocyanide ion and phenylene diamine derivatives have been proposed as mediator compounds. More recently ferrocene compounds have been used. In the presence of sulfide, the oxidized form of the ferrocene compound or other mediator can be reduced by homogeneous chemical reaction with the sulfide rather than by electrochemical reduction. In consequence, the presence of the sulfide analyte leads to an increase in the observed oxidative current and a reduction in the observed reductive current for the electrochemical reactions. The magnitude of these changes is dependent on the concentration of the sulfide, which is the analyte, and can be used to determine the analyte concentration.

This approach to the electrochemical determination of sulfide was described in W02001/063094 and W02004/011929. Subsequently, ferrocene carboxylate and sulfonate have been suggested as as mediator compounds in Electroanalysis Vol pp1658-63 (2006) and in Electrochimica Acta Vol 52 pp499-50 (2006). A number of ferrocene sulfonates for possible use in this way have been described in Journal of Organometallic Chemistry Vol 692 pp5173-82 (2007). Experimental work in this area has, however, generally been confined to laboratory experiments at ambient room temperature.

Determining sulfide downhole by means of this kind of electrochemical procedure entails exposing a sensor, including the compound which acts as the mediator, to the aggressive environment of the subterranean temperature and fluid. It may be necessary to accept that the sensor will have a limited working life, perhaps no more than a day.
The magnitude of the analytical signal, deriving from reaction between the analyte and the mediator compound, is dependent on the concentration of the mediator compound as well as on the concentration of the analyte. However, at working temperatures likely to be encountered, the mediator compound may undergo significant decomposition within the limited working lifetime of an electrochemical sensor located downhole.
One way to overcome this is to provide two sensors, only one of which is exposed to the analyte whilst the other is not exposed to the analyte and serves as a reference.
Apart from requiring two sensors and thereby doubling a cost, a difficulty with this approach is that the sensors cannot be exposed to identical conditions and the accuracy of the reference is therefore open to question.

The invention The present invention provides a way to calibrate a sensor whilst it is in position below ground, so that the mediator compound and the sensor containing it have a useful lifetime even though the concentration of the mediator compound is progressively decaying with time.

Although this problem of a decomposing mediator compound has been recognized in the context of determining sulfide species - to which the invention does indeed have particular applicability - it is also applicable with other analytes. This invention can be applied where an electrochemical analytical process entails the utilisation of a mediator compound whose concentration needs to be determined.

In a first aspect this invention provides a method of analysing a fluid for an analyte species, comprising:

i) providing a sensor comprising electrodes and a mediator compound capable of undergoing a cycle of oxidation and reduction in response to varying potential applied to the electrodes, where the mediator compound is also able to undergo a portion of the cycle through reaction with the analyte;
ii) bringing the sensor and the fluid into contact sufficiently for the analyte to react with the mediator, iii) applying potential to the electrodes and observing current flow with sufficient time for reaction between the mediator compound and the analyte, for thereby enabling observation of the concentration of the analyte species present, characterized in that the method also comprises iv) calibrating the sensor by applying rapidly changing potential to the electrodes and observing current flow, the change of this potential and observation of current being performed sufficiently quickly to observe electrochemical oxidation and reduction of the mediator compound taking place independently of the concentration of analyte, for thereby observing the concentration of the mediator compound.

The changing potential required for this calibrating part (iv) of the invention may be varied continuously or changed in discrete steps or pulses.

In the above method, point (iv) has the effect of separating baseline conditions, as would be observed in the absence of analyte, from the conditions with analyte present. This allows calibration of the sensor through observation of the current concentration of the mediator compound which enables the concentration of analyte to be derived from observations made in point (iii) of the method.

This invention is not limited to a specific mediator chemical nor to a specific electrochemical reaction of the mediator. However, the electrochemical change may be oxidation and/or reduction and such a redox reaction may be a change in oxidation state of the mediator brought about by electron transfer.

The method of this invention may be carried out at a subterranean location, notably downhole in a wellbore. Possibilities for the fluid which is subjected to analysis include subterranean water or brine and fossil hydrocarbon such as natural gas or crude oil.
However, the method could also be carried out at the surface, possibly to analyse produced hydrocarbon or to analyse an effluent stream.

The sensor used in the method of this invention may be of a type which is already known. It may be an electrochemical cell of the Clark type in. which the electrodes are in contact with a fluid electrolyte and the mediator compound is in solution in the electrolyte. In such an arrangement, the electrolyte may be separated from other fluid 5 by means of a membrane which is permeable to the analyte so as to allow the analyte to pass from the subterranean fluid into the electrolyte. A sensor of this character, intended for analysis of sulfide downhole, is the subject W02004/063743. A
cell of this type using microelectrodes with a thin layer of electrolyte separated by a membrane of polytetrafluoroethylene (ptfe) from the fluid to be analysed has been described in Anal. Chem. vol.75 pp 2499-2503 (2003).

Another possibility for sensor construction is that the mediator compound and electrolyte may be contained within a porous electrode, as described in W02004/011929, where the porous electrode may be separated from subterranean fluid by a membrane permeable to the analyte.

It is also possible that electrodes can be screen printed onto an insulating support, as has also been discussed in W02004/011929. Such an arrangement has also been described in Anal. Chem. vol.75 pp 2054-2059 (2003) where it is proposed that the mediator compound should be an insoluble compound mixed with a carbon-based electrode.
The electrolyte may then be provided by constituents present in the fluid for analysis.

The application of potential at part (iii) of the method set out above could be application of a fixed potential, for example the potential at which the current flow associated with oxidation is at a maximum. Alternatively, it may be application of a varying potential with observation of the current as the potential is varied. However, the variation and observation of current must allow time for reaction of the mediator and the analyte, whereas the change of potential in the calibrating part (iv) of the method must be considerably faster than any variation of potential at part (iii) so that electrochemical oxidation and reduction of the mediator compound take place independently of the concentration of analyte.

More specifically, the application of potential at part (iii) of the method may be carried out as cyclic voltammetry in which the potential applied to a working electrode is cycled over a sufficient range to bring about the oxidation and reduction reactions while recording the current flow as the potential is varied. Such cyclic voltammetry has been described and exemplified in Electroanalysis vol 12 page 1453 (2000) and in later documents including W02004/063743. The recorded current shows peaks at the potentials associated with the reduction and oxidation reactions.

Cyclic voltammetry is customarily performed with a continuously varying potential, changing sufficiently slowly that electrochemically oxidised mediator compound is able to come into contact with analyte within the electrolyte. Potential which changes in steps rather than continuously can possibly be employed as an alternative, provided the steps are long enough for steady-state conditions to be established before a subsequent step in potential.

It is also possible that this variation in potential whilst recording current flow could be carried out over only a portion of the reduction and oxidation cycle. This would be classed as linear scan voltammetry.

The characterising feature which is the calibration referred to as point (iv) of this invention as stated above may be implemented as one of the various forms of pulse voltammetry, with square wave voltammetry being the preferred technique. This will now be explained further and the invention will be exemplified with reference to the following drawings.

Brief Description of the Drawings Fig. 1 shows the waveform applied in cyclic voltammetry;

Fig. 2 shows the results of cyclic voltammetry applied to solutions containing varying concentrations of sulfide and also shows a plot of peak current against sulfide concentration;

Fig. 3 shows sequences of voltammograms taken at intervals at two different temperatures;

Fig. 4 is a graph in which current flow is plotted against time at various temperatures;
Figs 5, 6 and 7 are schematic representations of waveforms applied in pulse voltammetry together with illustrations of typical observations;

Fig. 8 shows the results of square wave voltammetry applied to solutions containing varying concentrations of sulfide;

Fig. 9 shows the results of square wave voltammetry applied to solutions containing varying concentrations of t-BuFcSO3 and varying concentrations of sulfide;

Fig. 10 is a schematic representation of a wellbore tool which is positioned in a wellbore;
Fig. 11 is a schematic cross sectional view of the electrochemical sensor within the tool of Fig. 10; and Fig. 12 shows the electrodes on one face of an electrode assembly within the sensor of Fig. 11.

Detailed Description Cyclic voltammetry is normally carried out using an electrochemical cell with three electrodes: a working electrode, a counter electrode and a reference electrode. A
varying potential relative to the reference electrode is applied to the working electrode.

This potential is varied, usually linearly, over a period of time in a cycle from a lower limit value to an upper limit value and then back again after which the cycle may be repeated.
This linearly varying waveform, with the cycle being repeated, has the form shown schematically in Figure 1.

The direct measurement from the procedure is the current flow as potential is applied.
The values of particular interest are peak values of current flow together with the applied potentials at which these peaks of current occur. However, it is also possible for the data obtained throughout a cyclic voltammetry experiment to be used as input to a computer program for modelling the chemical processes which occur.

Figure 2 shows the results of a laboratory experiment in which cyclic voltammetry was carried out at room temperature on pH 6.5 aqueous solutions containing 0.5 mM
t-butylferrocene sulfonate (t-BuFcSO3) containing increasing concentrations of sulfide.
For this laboratory experiment a glassy carbon electrode can be used as the working electrode with a standard calomel electrode as the reference electrode.

The left side of Figure 2 shows the voltammetric responses (sometimes termed voltammograms) obtained as potential applied to the working electrode relative to the reference electrode was scanned (i.e varied) linearly between 0.0 and +0.6 volt and back to O.Ovolt at a scan rate of 0.1 volt/sec. The response in the absence of sulfide (bottom curve) revealed an oxidative wave at +0.4volt, and a corresponding reduction wave at +0.3volt. This is consistent with the oxidation and reduction of the ferrocene/ferricenium redox couple. In the presence of sulfide the voltammograms at the left show an increase in the oxidation peak current along with a concomitant decrease in the reduction wave. The graph at the right shows the oxidation peak currents plotted against the concentrations of sulfide, with tielines to the corresponding voltammograms.
These changes in the presence of sulfide are attributed to at least some of the oxidised ferricenium species undergoing reduction by homogenous chemical reaction with sulfide in solution, instead of electrochemical reduction. Because this alternative reductive pathway is available, the oxidative current is increased and the reductive current is reduced.

The homogenous chemical reduction requires interaction between molecules of the oxidised mediator (ie the ferrocinium compound) and molecules of the analyte.
Consequently the rate of this reduction will depend on both the concentrations of the mediator compound and on the concentration of the sulfide which is the analyte. The scan rate, i.e. the rate at which applied electrical potential is varied, it is chosen to be sufficiently slow that mediator compound which has undergone electrochemical oxidation will have time to diffuse and collide with analyte, in order to be reduced by reaction in solution.

As mentioned previously, such cyclic voltammetry in the presence of a mediator can be used to observe the concentration of the sulfite analyte, implementing point (iii) of the invention as set out above. However, obtaining a meaningful result requires the concentration of the mediator compound present during the voltammetry experiment to be known. The present inventor has found that this becomes problematic at elevated temperatures.

Figure 3 shows the cyclic voltammetric responses of 0.5mM t-BuFcSO3 solutions (pH 7) obtained in the absence of sulfide over a period of 24 hours at 87 C and 133 C
respectively. In each case the voltammograms shown were measured at hourly intervals. At 87 C, while some shift in the peak potential can be seen, only a small decay (about 15 %) in the oxidative peak current was observed over the 24 hour period. This can be attributed to thermal degradation of the mediator compound. When the temperature was raised to 133 C, a much larger decrease (about 87 %) in the oxidative peak current was observed. This shows that the t-BuFcSO3- mediator compound decays at these temperatures - confirming a hypothesis that the decay process is thermally activated. The concentration of the t-BuFcSO3 mediator compound thus ceases to have a known value.

The decay of the mediator compound at various temperatures is also shown by Figure 4.
This plots the oxidative peak current against time for each of five temperatures. The values are all shown as the remaining percentage of an initial amount at the temperature concerned. It can be clearly seen that as the temperature is increased from 87 to 133 C, the loss of mediator compound, leading to a loss of analytical signal, becomes progressively quicker.

To overcome this problem, in accordance with this invention, the concentration of the mediator is observed with rapid change to the applied potential. In conventional cyclic voltammetry the variation of electrical potential is sufficiently slow that mediator compound which has undergone electrochemical oxidation will have time to diffuse and collide with analyte, so as to be reduced by the chemical reaction with the analyte. In contrast with this practice of cyclic voltammetry with a slow variation of electrical potential, when a rapidly changing potential is applied as required for point (iv) of the 5 invention, there is insufficient time for diffusion of the mediator compound and so little or no chemical reduction by the sulfide analyte can take place. This chemical reaction cannot now compete effectively with the electrochemical reduction, or to put it another way the electrochemical reduction outruns the chemical reaction. As will now be shown, the consequence is that the electrochemical process becomes independent of the 10 presence of analyte.

There are several possibilities for providing the rapid change of potential required for the calibrating part (iv) of the invention. One possibility is to use cyclic voltammetry with a fast scan rate so that the variation of the applied potential becomes rapid.
Calculation, carried out by use of a simulator program, has indicated that a scan rate of 15 volt/sec would be suitable. Other possibilities are to adopt one of the various forms of pulse voltammetry. There are three main pulse techniques available, normal pulse voltammetry (NPV), differential pulse voltammetry (DPV) and square wave voltammetry (SWV). All three entail changing the applied potential in steps and making a brief measurement of current flow between changes. The following description of these techniques will assume for the purposes of explanation that the applied potential is positive and causes electrochemical oxidation. However, it should be appreciated that these techniques could also be carried out with negative potential causing electrochemical reduction.

Figure 5A shows the waveform used for normal pulse voltammetry.. There is a baseline potential applied to the reference electrode and at intervals the potential is stepped up to a higher value, held at that value for a short duration and then returned to the baseline. The steps in potential are increased progressively. The pulses, i.e.
the steps up to a higher potential, are of short duration separated by much longer intervals at the baseline potential. The measurements of current flow are taken towards the end of a pulse as indicted by the points tin Figure 5A. The data obtained consists of a succession of values of applied potential paired with the measured current at that potential. This data can be plotted in similar manner to the oxidative portion of a voltammogram obtained in cyclic voltammetry with continuously varying potential. However, because the pulses are short each measurement of current is made before the electrochemical process becomes influenced by any homogenous chemical reduction. As shown schematically in Figure 513, the plot typically shows a rise in current as the pulses reach the potential required to effect oxidation.

Figure 6A shows the waveform used for differential pulse voltammetry. Again the applied potential is stepped up in short pulses separated by much longer intervals, but this is superimposed on a baseline which is of a staircase form so that the baseline potential rises after each pulse. Current flow is measured just before each pulse begins, as indicated by the points a in Figure 6A and also measured just before each pulse ends, as indicated by the points b in Figure 6A. The difference between the current measured at each point b and that measured at an adjacent point a is determined and this difference between the two measured values is recorded as the output from the procedure.

Figure 6B shows schematically this difference in current plotted against the applied potential. When the baseline potential is well below that required for oxidation, the step in potential has little effect on the current flow and the difference in current before and during a pulse of raised potential is negligible, as seen at the left of Figure 6B. When baseline potential is well above the value required for oxidation the electrochemical process is proceeding at a rate limited by diffusion and the increase in potential during a pulse does not increase the rate of reaction very much, so that the difference in current during a pulse of higher potential is fairly small as seen at the right hand side of Figure 6B. However, in between these extremes, the increase in potential during a pulse significantly increases the current and the plot of current difference rises to a peak at a potential associated with the electrochemical oxidation, as seen in Figure 6B.

Figure 7A shows the waveform used in square wave voltammetry which is the preferred technique for the characterizing part (iv) of this invention. A square wave is combined with a staircase baseline, so that the potentials applied at the peaks and troughs of the square wave increase progressively. Current flow is measured close to the end of each peak and trough of the square wave as indicated by the points f and r in Figure 7A. The amplitude of the square wave is chosen to be sufficiently large that, in at least part of the scan, the potential at the peak of the square wave causes oxidation and the measured current is a forward current while the potential at the trough of the square wave reverses the oxidation and the measured current is a reverse current. Both the peaks and troughs of the square wave are pulses of short duration (usually the troughs and peaks are of equal duration) and so, as with the other forms of pulse voltammetry each measurement of current is made before the electrochemical process becomes influenced by any homogenous chemical reduction.

Three outputs can be obtained from a square wave voltammetry experiment. One is the applied potentials and associated current flow at the points f, i.e at the peaks of the square wave, plotted as the'curve W(f) in Figure 7B. Another is the applied potentials and associated current flow at the points r, i.e at the troughs of the square wave, plotted as the curve W(r). The third is the applied potentials and the associated current difference. This is plotted as the curve OW in Figure 7B and with differential pulse voltammetry this rises to a peak at a potential associated with the electrochemical oxidation.

One advantage of square wave voltammetry is that the peak in the curve OW is larger than the peaks in the plots of forward and reverse current W(f) and W(r) so that the technique has good sensitivity. A second advantage is that the technique can be carried out with a more rapid scan rate than other forms of pulse voltammetry so that measurements can be made quickly.

Figure 8 shows current difference (i.e. OW) curves obtained by square wave voltammetry applied to pH 7 solutions containing 0.5mM t-BuFCSO3- and increasing sulfide concentrations at room temperature. The sulfide concentrations ranged from 0 to 1mM by 0.2mM steps. It can be seen that the curves are virtually superimposed thus showing that with this square wave voltammetry the electrochemical oxidation and reduction of the t-BuFcSO3- had been successfully isolated from its reduction by sulfide.
This was further confirmed by the data shown in Figure 9. Square wave voltammetry was applied to pH 7 solutions at room temperature containing varying concentrations of t-BuFcSO3- and varying concentrations of sulfide.. In Figure 9A the oxidative peak current (ie the maximum of the W(f) curve) is plotted as a function of sulfide concentration for various concentrations of t-BuFcSO3 and it can be seen that the recorded values are effectively independent of the sulfide concentration. In Figure 9B the same data is presented as a plot of the oxidative peak currents against t-BuFcS03 concentration. The points lay on a straight line, showing that the measured peak currents were proportional to the concentration of the t-BuFcSO3 mediator compound.

Figures 10 to 12 illustrate equipment used to perform the method of the invention below ground, within a wellbore. The tool 10 comprises an elongate substantially cylindrical body which is suspended on a wireline 14 in the wellbore 16, adjacent an earth formation 18 believed to contain recoverable hydrocarbons. The tool is provided with a radially projecting sampling probe 20. The sampling probe 20 is placed into firm contact with the formation 18 by hydraulically operated rams 22 projecting radially from the tool on the opposite side from the sampling probe 20 and is connected to a conduit 26 within the tool. A pump 28 within the tool 10 can be used to draw a sample of the hydrocarbons into the conduit 26. The pump 28 is controlled from the surface at the top of the wellbore via the wireline 14 and control circuitry (not shown) within the tool.
The conduit 26 leads through an electrochemical sensor 30 located close to the sampling probe 20.

The sensor 30 is shown rather schematically in cross section in Figure 11. It may be constructed as described in greater detail in W02004/063743 and/or W02005/066618.
The sensor 30 is generally cylindrical. A flowpath for the fluid whose sulfide content is to be determined extends through the sensor 30 and forms part of the conduit 26.
A gas permeable membrane 36 separates this flow path from an axial bore through the sensor, within which an electrode assembly 38 is located. This assembly 38 comprises an insulating body, having three electrodes on its face 40, namely a working electrode 42 made from boron-doped diamond, a reference electrode 44 in the form of a silver dot coated with silver chloride or silver iodide, and a counter electrode 46 comprising a printed platinum track. The electrodes 42, 44, 46 are connected via respective electrical conductors moulded into and extending axially through the body of the electrode assembly 38 to respective electrical leads 48, which connect the sensor 30 to control circuitry 32 within the tool. The space 50 between the face 40 of the electrode assembly and the membrane 36 is filled with a polar electrolyte which may be an aqueous solution in which the mediator compound t-BuFcSO3 is dissolved.

Once the tool is in place, fluid is drawn through the conduit 26 by the pump 28.
Hydrogen sulfide in the fluid can pass through the membrane 36 into the electrolyte in the space 50. After a time for equilibrium to be reached, the control unit 32 (possibly on command received via the wireline 14) applies varying potential to the electrodes and meters the current flowing. This is done as cyclic voltammetry with a scan rate which is slow enough to allow time for reaction between the mediator compound and the sulfide which has entered the electrolyte. The current flowing and the applied potential may be communicated to the surface in real time via the wireline 14 or may be recorded until the tool is retrieved to the surface. Separately, either before or after this voltammetric measurement, the control unit 32 applies a square wave to the electrodes to calibrate the sensor 30 by square wave voltammetry. The current flowing is recorded during the peaks and troughs of the square wave as described above with reference to Figure 7.
The recorded current and/or the values of current difference together with the associated values of applied potential may be communicated to the surface via the wireline 14 or recorded until the tool is retrieved.

Claims (14)

1. A method of analysing a fluid for an analyte, comprising i) providing a sensor comprising electrodes and a mediator compound capable of undergoing a cycle of oxidation and reduction in response to varying potential applied to the electrodes, where the compound is also able to undergo a portion of the cycle through reaction with the analyte;

ii) bringing the sensor and the fluid into contact sufficiently for the analyte to react with the mediator, iii) applying potential to the electrodes and observing current flow with sufficient time for reaction between the mediator compound and the analyte, for observing the concentration of the analyte, and iv) calibrating the sensor by applying a rapidly changing potential to the electrodes and observing current flow, the change of this potential and observation of current being sufficiently fast to allow and observe electrochemical oxidation and reduction of the mediator taking place independently of the concentration of analyte, thereby observing the concentration of the mediator compound.

2. A method according to claim 1 wherein the mediator compound is dissolved in a conductive electrolyte separated from a fluid by a membrane which is permeable to the analyte.

3. A method according to claim 1 wherein (at iii) the application of potential to the electrodes for observing the concentration of the analyte is carried out by applying varying potential.

4. A method according to claim 1 wherein (at iii) the application of potential to the electrodes for observing the concentration of the. analyte is carried out as cyclic voltammetry.

5. A method according to claim 1 wherein (at iv) calibrating the sensor is carried out as cyclic voltammetry with a sufficiently fast scan rate that oxidation and reduction of the mediator takes place independently of the concentration of analyte.

6. A method according to claim 1 wherein (at iv) the step of calibrating the sensor is carried out as pulse voltammetry.

7. A method according to claim 6 wherein (at iv) the step of calibrating the sensor is carried out as square wave voltammetry.

8. A method according to claim 1 wherein the analyte reduces the mediator compound.

9. A method according to claim 1 wherein the analyte is hydrogen sulfide and /or sulfide anion.

10. A method according to claim 1 wherein the analyte oxidizes the mediator compound.

11. A method according to claim 1 performed below ground to determine the concentration of analyte in a subterranean fluid.

12. A method according to claim 11 wherein the sensor is part of a tool lowered down a wellbore.

13. A method according to claim 11 wherein the subterranean fluid is fossil hydrocarbon.

14. A method according to claim 11 wherein the subterranean fluid is underground water or brine.