Gas sensors based on semiconducting polymers, such as polypyrrole, are well known. The conductivities of such polymers are perturbed by the adsorption of certain gaseous species onto the polymer, a phenomenon which is generally exploited by measuring the changes in dc resistance which accompany exposure of the sensor to a gas.

In general a given semiconducting polymer is sensitive towards a range of molecules; and this lack of selectivity is overcome by employing an array of different sensors (see, for example, K C Persaud and P Pelosi, "Sensor Arrays Using Conducting Polymers For An Artificial Nose", in Sensors and Sensory Systems for an Electronic Nose, Eds J W Gardner and P N Bartlett, Kluwer Academic Publishers, 1992).

In this instance, it is the intention that the pattern of sensor response across the array comprises a distinctive, molecule specific, "fingerprint". An alternative detection method, involving the measurement of changes in an ac impedance characteristic as a function of applied frequency, is described in UK Patent GB 2 205 553.

Semiconducting polymer conduction mechanisms, and, in particular, the manner in which such mechanisms are affected by the adsorption of gaseous species, are not well understood. Polymers are only conductive when the polymer backbone possesses a net charge, which is balanced by the presence of oppositely charged "dopants" or counterions within the polymer structure. In most cases, polymers are conductive in their "oxidised" state, wherein the polymer backbone is positively charged.

The ratio between the number of monomer units and the number of positive charges is variable, depending on the monomer and the counterion. For polypyrrole with BF4- as counterion, for example, there is on average a positive charge per four pyrrole monomer units.

Polarons (radical cations) and bipolarons (dications separated in space) have been implicated in semiconducting polymer electrical conduction mechanisms. In one experimental investigation, Scott et al made simulataneous measurements of conductivity and ESR signal for a polypyrrole film as a function of oxygen partial pressure ( J C Scott, P Pfluger, M T Kroundbi and G B Street; Phys. Rev. B., Condens. Matt., 28 (1983) 2140). An initial increase in both ESR signal and conductivity was observed, followed by a plateau in the conductivity and a reduction to zero in the ESR signal. The initial rise in ESR signal was interpreted as being due to the formation of polarons - these species possess both spin and charge. The polarons disappear at higher doping levels because they are unstable relative to bipolaron species. Since bipolarons have zero electron spin they do not give rise to an ESR signal. These conclusions have been confirmed by optical techniques ( J L Bredas, J C Scott, K Yakushi and G B Street; Phys. Rev. B.,30 (1984) 1023) and by further ESR investigations (F Genoud, M Guglielmi, M Nechtschein, E Genies and M Salmon; Phys. Rev. Lett., 55 (1985) 118). However, it is likely that the observed ESR signal is also affected, to some extent, by conformational changes in the polymer, such as changes in bond angle.

The present invention is based upon the observation that ESR signal derived from a semiconducting polymer can vary when the semiconducting polymer is exposed to a gas - presumably due to a modulation of the polaron and bipolaron populations - and provides a new method and apparatus for detecting gases.

Although Scott et al describes variations in ESR intensities and linewidths as a function of oxygen uptake, there is no suggestion that ESR type measurements might serve as the basis of a gas detection technique. Scott et al is concerned with the elucidation of fundamental conductivity mechanisms. Furthermore, and very importantly, in Scott et al the polymers are exposed to oxygen in order to oxidise the polymer, thereby converting from the insulating state to the conducting state. In the present invention the detected gases do not effect such a chemical reaction, and the polymer is present throughout in its conducting state.

It is understood that the term "gas" relates to all gas phase species, including vapours emanating from volatile liquids or solids subject to sublimation.

According to a first aspect of the invention there is provided a gas sensor comprising: at least one semiconducting polymer; magnetic field applying means for applying a magnetic field across said semiconducting polymer; microwave means for producing microwave radiation, said microwave radiation being directed to said semiconducting polymer; and detection means for detecting variations in an adsorption characteristic of the semi conducting polymer in the presence of a gas or a mixture of gases to be detected.

The microwave means may produce microwave radiation of constant or variable frequency.

The magnetic field applying means may apply a magnetic field of variable magnitude to said semiconducting polymer. One such approach - sweeping the magnetic field strength whilst employing a constant frequency microwave source - is analogous to conventional ESR spectroscopy.

Alternatively, the magnetic field applying means may apply a magnetic field of constant magnitude to said semiconducting polymer. If microwave radiation of constant frequency is employed, the strength of the applied field would, of course, be selected to bring a polymer electron spin flipping transition into resonance with the frequency of the microwave radiation.

The detection means may detect variations in the attenuation of the microwave radiation by said semiconducting polymer.

The variations in the absorption characteristic of the semiconducting polymer may be measured with respect to a reference characteristic. This reference characteristic may be an absorption characteristic of the semiconducting polymer in the presence of a reference gas.

The microwave means may comprise a klystron, a backward wave generator, or a Gunn diode.

According to a second aspect of the invention there is provided an array of gas sensors, each according to the first aspect of the invention, the gas sensors comprising different semiconducting polymer or polymers. Microwave radiation from a single microwave means may be directed to all of the gas sensors in the array.

According to a third apsect of the invention there is provided a method for detecting a gas or a mixture of gases comprising: applying a magnetic field across at least one semiconducting polymer; directing microwave radiation to said semiconducting polymer; exposing said semiconducting polymer to the gas or gas mixture; and detecting variations in an absorption characteristic of the semiconducting polymer in the pressence of the gas or mixture of gases.

Microwave radiation of constant frequency may be directed to said semiconducting polymer. The magnitude of the magnetic field may be varied or held constant.

Variations in the attenuation of the microwave radiation by said semiconducting polymer may be detected.

Variations in the absorption characteristic may be measured with respect to the absorption characteristic measured in the presence of a reference gas.

A gas sensor and methods for detecting a gas or a mixture of gases in accordance with the invention will now be described with reference to the following drawings, in which:- Figure 1 is a schematic diagram of a gas sensor; Figure 2 is a first ESR spectrum obtained with poly-N-methyl pyrrole in an atmosphere of air; Figure 3 is a first ESR spectrum obtained in an atmosphere of methanol and air; Figure 4 is a second ESR spectrum obtained in an atmosphere of air; and Figure 5 is a second ESR spectrum obtained in an atmosphere of methanol and air.

Figure 1 shows a schematic diagram of a gas sensor according to the invention comprising: at least one semiconducting polymer 10; magnetic field applying means 12 for applying a magnetic field across said semiconducting polymer 10; microwave means 14 for producing microwave radiation, said microwave radiation being directed to said semiconducting polymer 10; and detecting means 16 for detecting variations in an absorption characteristic of the semiconducting polymer in the presence of a gas or a mixture of gases to be detected.

In this embodiment the gas sensor is a commercial X band ESR spectrometer Varian E104 having semiconducting polymer 10 in its ESR cavity 18. The magnetic field applying means 12 is an electromagnet, producing a homogenous magnetic field, the field being linearly variable around ca. 0.34 T. The microwave means 14 is a klystron, producing monochromatic microwave radiation of frequency ca. 9 GHz.

This microwave radiation is directed to the ESR cavity 18 by an appropriate waveguide 20. The detecting means 16 is a standard crystal detector.

In keeping with standard ESR methodology, the klystron 12 produces microwave radiation of constant frequency, and the electromagnet 14 applies a magnetic field of variable magnitude to the semiconducting polymer 10. Modulation of the magnetic field is performed and the output of the crystal 16 is detected with a phase sensitive amplifier 22. The spectrum thus obtained is displayed by display means 24, which in this instance is a chart recorder but could comprise a computer, which might also perform some or all of the control/analysis functions described above.

The semiconducting polymer 10 may be present in any suitable form, e.g.

as a powder or a film, and may be polycrystalline or amorphous. For the present purposes 2 mg of poly-N-methyl pyrrole powder was placed in an ESR tube which was located in the ESR cavity 18.

ESR spectra were obtained with the poly-N-methyl pyrrole powder exposed to atmospheric air and to methanol vapour. The spectra were obtained at room temperature with a microwave frequency of 9.491 GHz; the magnetic field was swept between 0.32 and 0.36 T. The microwave power was 100 mW, and the magnetic field was modulated at 10 KHz with a modulation amplitude of 10 pT. Figure 2 shows the spectrum obtained in atmospheric air and Figure 3 shows the effect of injecting 1 ml of methanol vapour into the ESR tube. The methanol vapour was then removed and a second spectrum obtained in the presence of atmospheric air - this shown in Figure 4.

The poly-N-methyl pyrrole powder was then exposed to a second atmosphere of methanol vapour, by injecting 2 ml methanol vapour into the ESR tube, and an ESR spectrum (Figure 5) obtained.

The intensities of the observed first derivative lineshapes are shown in Table 1. Clearly, there is a substantial reduction in ESR signal in the presence of methanol, presumably because methanol adsorption leads a decrease in the number of polarons present in the polymer with a concomitant increase in the number of spinless bipolarons. It is interesting to note that the resistance of poly-N-methyl pyrrole increases on exposure to methanol vapour; these two observations might suggest at first sight that the polarons are the main current carriers. However, the true picture is likely to be rather more complicated, and at present the conductivity mechanism or mechanisms of semiconducting organic polymers are not well understood. For the present purposes, it is quite sufficient to note that the presence of a gas - methanol - may be recognised by detecting a variation in ESR line intensities.

It should be noted that no change can be discerned in ESR line position or linewidth in Figures 2 to 5. However, this does not exclude the possibility of detecting perturbations in either of these absorption characteristics with other semiconducting polymers.

It is also interesting to note that the second measurement in air, after exposure of poly-N-methyl pyrrole to methanol, is substantially identical to the first measurement in air. This is not altogether surprising, since it is known that semiconducting polymers display rapid and reversible adsorption kinetics, but it is reassuring that the charge distribution in the bulk polymer, and hence the ESR signal, appears to return to its original state.

The concentration of methanol employed is rather high (1-2 ml corresponds to ca. 2.6 - 5.2 x 10-3 M methanol vapour in the ESR tube). The similarities in peak-to- peak response between the two methanol measurements, made at methanol concentrations differing by a factor of two, suggest that saturation is occuring. This conclusion is convincingly supported by measurements of area. By analogy with measurements of semi conducting polymer conductivities, it might be expected that approximately linear responses would be obtained at lower vapour concentrations, although this is yet to be confirmed.

Variations in ESR linestrength have also been observed with pyridine and 2-methoxy ethanol vapour. It is to be expected that a single semiconducting polymer will be sensitive towards a range of gases.

Gas sensors of the present invention are desirably rather smaller than a conventional ESR spectrometer. This can be accomplished using standard knowledge in the art and standard components such as Gunn diodes, klystrons and electromagnets.

The precise configuration adopted is dependent upon the application desired. For example, the combination of applied magnetic field and the microwave frequency is selected so as to satisfy the well known resonance criterion: hv=gpBB where h is Planck's constant, v is the microwave frequency, g is the g factor, pB is the Bohr magneton and B the magnetic field. Device sensitivity should be enhanced at higher magnetic fields (and correspondingly higher microwave frequencies) because the separation between spin energy levels is increased, and therefore the population of the upper state by thermal excitation is diminished. However, such high magnetic fields may not be commensurate with the production of a cheap and compact gas sensor, and therefore a trade-off may be necessary. The optimum trade-off is likely to be application specific. In principle, of course, the magnitude of the magnetic field may be such that the resonant frequency lies outside of the microwave region. Although such is not practical at the present time, it is to be understood that the use of very large magnetic fields is within the scope of the invention.

The variations in absorption characteritic are detected with respect to a reference spectrum, which in this case is the value obtained in atmospheric air. Such an approach is susceptible to variations in ambient humidity, and an improvement would be to pass a pure, inert reference gas such as nitrogen over the semiconducting polymer.

A gas sampling system, such as one commonly adopted with conductimetric semiconducting polymer arrays, would be desirable. In such systems, the reference gas and the gas to be analysed are slowly pumped across the sensor or sensors.

Although in the example described above the magnetic field is swept, it may be possible to detect linestrength variations by employing a magnetic field of constant strength, the combination of magnetic field and microwave frequency being selected, of course, to probe within the ESR lineshape. If phase sensitive detection is employed, the optimum resonance lies at magnetic fields slightly removed from the line centre. If phase sensitive detection is not employed, then the magnetic field is advantageously at the absorption maximum.

The present invention also encompasses arrays of gas sensors, the gas sensor employing different semiconducting polymers in conjunction with ESR detection of the general type described above. Each semiconducting polymer will respond differently to a gas, and therefore the pattern of response across the array would represent a distinctive molecular "fingerprint". It is also possible, by use of an appropriate plurality of waveguides, to direct microwave radiation from a single microwave source to an array of gas sensors.

Another possible embodiment involves the use of gas sensors or gas sensor arrays of the present invention in combination with other kinds of gas sensors or gas sensor arrays.

Table 1. ESR Intensities Measurement Peak to Peak Intensity Integrated Areab/ Variationa/Arb. Units Arb. Units Air 89 290 MeOH (lml) 59 182 Air 93 315 MeOH (2ml) 56 184 a Measured difference between the two first derivative lineshape peaks b Estimated area under line shape, measured with respect to extreme field baseline and ignoring the sign of the line intensity