ELECTROCHEMICAL GAS SENSOR, FILTER AND METHODS

Abstract

The invention relates to an electrochemical gas sensing apparatus for sensing one or more analytes, such as NO.sub.2 and/or O.sub.3, in a sample gas and a method of using same. The apparatus uses Mn.sub.2O.sub.3 as a filter for ozone. The Mn.sub.2O.sub.3 may take the form of a powder which may be unmixed, mixed with various PTFE particles sizes, formed into a solid layer deposited onto a membrane and/or pretreated with NO.sub.2.

Claims

1. Electrochemical gas sensing apparatus for sensing at least one gaseous analyte and comprising a housing having an inlet, a sensing electrode and an ozone filter interposed between the sensing electrode and the inlet, wherein the ozone filter comprises Mn.sub.2O.sub.3.

2. Electrochemical gas sensing apparatus according to claim 1, wherein the Mn.sub.2O.sub.3 is Mn.sub.2O.sub.3 powder.

3. Electrochemical gas sensing apparatus according to claim 1 wherein the Mn.sub.2O.sub.3 is NO.sub.2 treated.

4. Electrochemical gas sensing apparatus according to claim 3, wherein the Mn.sub.2O.sub.3 is Mn.sub.2O.sub.3 powder which has at least 10.sup.12 molecules of NO.sub.2 adsorbed thereto per cm.sup.2 of surface.

5. Electrochemical gas sensing apparatus according to claim 2, wherein the ozone filter comprises Mn.sub.2O.sub.3 powder mixed with binder.

6. Electrochemical gas sensing apparatus according to claim 5, wherein the ozone filter comprises binder particles coated with the Mn.sub.2O.sub.3 powder.

7. Electrochemical gas sensing apparatus according to claim 1, wherein the ozone filter comprises a solid microporous layer which comprises Mn.sub.2O.sub.3.

8. Electrochemical gas sensing apparatus according to claim 1 for sensing NO.sub.2 and/or ozone.

9. Electrochemical gas sensing apparatus according to claim 1, further comprising a second sensing electrode, wherein the second working electrode is in gaseous communication with a gas sample without an intervening ozone filter.

10. Electrochemical gas sensing apparatus according to claim 9, wherein the at least one said analyte comprises ozone.

11. A method of forming an electrochemical gas sensing apparatus for sensing at least one gaseous analyte, the method comprising providing a housing having an inlet and a sensing electrode and providing an ozone filter interposed between the sensing electrode and the inlet, wherein the ozone filter comprises Mn.sub.2O.sub.3.

12. A method of forming an electrochemical gas sensing apparatus according to claim 11, wherein the ozone filter comprises Mn.sub.2O.sub.3 powder.

13. A method of forming an electrochemical gas sensing apparatus according to claim 12, comprising the step of treating the Mn.sub.2O.sub.3 powder with NO.sub.2.

14. A method of forming an electrochemical gas sensing apparatus according to claim 11, wherein the Mn.sub.2O.sub.3 is Mn.sub.2O.sub.3 powder is mixed with binder particles.

15. A method according to claim 14 wherein the Mn.sub.2O.sub.3 powder coats the binder particles.

16. A method of forming an electrochemical gas sensing apparatus according to claim 11, wherein the filter comprises a solid microporous layer of Mn.sub.2O.sub.3.

17. A method of forming an electrochemical gas sensing apparatus according to claim 16, wherein the method comprises depositing a layer of Mn.sub.2O.sub.3 on a gas porous membrane.

18. A method of forming an electrochemical gas sensing apparatus according to claim 11, comprising providing a second sensing electrode in gaseous communication with a gas sample without an intervening ozone filter.

19. A method of sensing NO.sub.2 and/or ozone comprising forming an electrochemical gas sensing apparatus by the method of claim 18, and exposing the gas sensing apparatus to a gas sample, wherein the difference in the signals from the sensing electrodes is representative of ozone in the gas sample, and thereby determining from the signals from the first and second sensing electrodes the concentration of NO.sub.2 and/or ozone in the gas sample.

20. A method of sensing NO.sub.2 and/or ozone comprising forming an electrochemical gas sensing apparatus by the method of claim 18, and exposing the gas sensing apparatus to a gas sample, wherein the difference in the signals from the sensing electrodes is representative of ozone in the gas sample, and thereby determining from the signals from the first and second sensing electrodes the concentration of NO.sub.2 and/or ozone in the gas sample.

21. A method of sensing NO.sub.2 and/or ozone comprising forming electrochemical gas sensing apparatus according to claim 1 and bringing the inlet into gaseous communication with a gas sample.

22. A method of sensing NO.sub.2 and/or ozone comprising forming electrochemical gas sensing apparatus by the method of claim 11 and bringing the inlet into gaseous communication with a gas sample.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0063] FIG. 1 relates to a gas sensing apparatus of the invention comprising a single electromechanical amperometric gas sensor for detecting NO.sub.2, having an ozone filter, working electrode and its associated counter electrode, reference electrode, electrolyte and additional working electrode;

[0064] FIGS. 2A and 2B relates to a gas sensing apparatus of the invention according to an embodiment comprising two individual electrochemical amperometric gas sensors one (shown in FIG. 2A) housing an ozone filter, the first working electrode and its associated counter electrode, reference electrode, electrolyte and additional working electrode, the other (shown in FIG. 2B) housing another (the second) working electrode and its associated counter electrode, reference electrode, electrolyte and additional working electrode;

[0065] FIG. 2C is a replacement for the gas sensor of FIG. 2B employed (along with the sensor of FIG. 2A) in embodiments of the invention in which a solid Mn.sub.2O.sub.3 layer is used as filter;

[0066] FIG. 3 is a schematic cross-sectional view of a gas sensing apparatus of the invention according to an embodiment wherein first and second working electrodes are combined into the same housing and share a common counter electrode, reference electrode and electrolyte;

[0067] FIG. 4 is a diagram of potentiostatic circuitry for a gas sensing apparatus in which first and second working electrodes share a common counter electrode, reference electrode and electrolyte as for example, in the embodiment of FIG. 1 or in which a working electrode and an additional working electrode share a common counter electrode, reference electrode and electrolyte as for example, in one of the individual gas sensors of the embodiment of FIG. 2;

[0068] FIG. 5 is an illustration of the experimental setup used for the treatment of the Mn.sub.2O.sub.3 filter powder;

[0069] FIG. 6A shows the current response to 0.5 ppm O.sub.3 of (a) unfiltered sensors according and (b) sensors in which the filter is 500 mg solid untreated Mn.sub.2O.sub.3; FIG. 6B shows the current response to 2 ppm NO.sub.2 of (a) a sensor in which the filter is 500 mg solid untreated Mn.sub.2O.sub.3 and (b) a sensor in which the filter is 500 mg of solid Mn.sub.2O.sub.3 treated with NO.sub.2 as described above; and FIG. 6C shows the current response to 0.5 ppm O.sub.3 of (a) unfiltered sensors and (b) sensors in which the filter is 500 mg solid Mn.sub.2O.sub.3 treated with NO.sub.2 as described above.

[0070] FIG. 7A illustrates the output current, over time, where the unfiltered sensor (trace (a)) and filtered sensor (trace (b)) are exposed in turn to zero air, then 1 ppm NO.sub.2, then zero air, then 1 ppm O.sub.3, then zero air, then a mixture of 1 ppm NO.sub.2 and 1 ppm O.sub.3; FIG. 7B shows the calculated O.sub.3 concentration (trace (a)) and NO.sub.2 concentration (trace (b)) which the traces of FIG. 7A represents;

[0071] FIG. 8 is a plot of the variation with time in cross-sensitivity to NO (as a fraction of the sensitivity to NO.sub.2) of (a) a sensor according to FIG. 2A(a) in which the filter is formed from 500 mg of solid Mn.sub.2O.sub.3, and (b) in which the filter is formed from 450 mg of solid MnO.sub.2;

[0072] FIG. 9 shows the output current with time in the presence of zero air, then 0.5 ppm O.sub.3, then zero air of sensors according to FIG. 2A (filtered) in which the filters were formed by 25 mg of powdered Mn.sub.2O.sub.3 mixed with 225 mg of PTFE particles having a size range of 710 μm to 1500 μm (i.e. 10% by weight of Mn.sub.2O.sub.3 in binder).

[0073] FIG. 10A shows the current response with time to 1 ppm of O.sub.3 of (a) a sensor according to FIG. 2A (filtered) in which the ozone filter comprised 500 mg of powdered Mn.sub.2O.sub.3, (not mixed with binder) and (b) a sensor according to FIG. 2B (unfiltered), and in FIG. 10B (a) a sensor according to FIG. 2A (filtered) in which the ozone filter comprised 25 mg of powdered Mn.sub.2O.sub.3 mixed with 225 mg of PTFE particles having a size range of 710 μm to 1500 μm (i.e. 10% by weight of Mn.sub.2O.sub.3 in binder) and (b) a sensor according to FIG. 2B (unfiltered).

[0074] FIGS. 11A-11B shows the current response to a range of concentrations of NO.sub.2 of sensors according to FIG. 2A (filtered) with filters comprising (A) 500 mg of powdered Mn.sub.2O.sub.3 and (B) 25 mg of powdered Mn.sub.2O.sub.3 mixed with 225 mg of PTFE particles having a size range of 710 μm to 1500 μm.

[0075] FIG. 12 shows the current response to 2 ppm NO.sub.2 of sensors according to FIG. 2A (filtered) with filters comprising, in FIG. 12A, 500 mg of powdered Mn.sub.2O.sub.3 (a) 30 days after manufacture and (b) 173 days after manufacture and, in FIG. 12B, 25 mg of powdered Mn.sub.2O.sub.3 mixed with 225 mg of PTFE particles having a size range of 710 μm to 1500 μm (i.e. 10% by weight of in Mn.sub.2O.sub.3 in binder), (a) 19 days after manufacture and (b) 187 days after manufacture.

[0076] FIG. 13 is a photograph, to the scale shown, of filter material according to the invention, in the form of PTFE particles coated with Mn.sub.2O.sub.3 crystals.

[0077] FIG. 14(A) shows the current response with time to air and then to 0.5 ppm O.sub.3 of (a) an SO.sub.2 sensor without an ozone filter, and (b) an SO.sub.2 sensor with an ozone filter comprising 10% by weight of Mn.sub.2O.sub.3 powder in binder (after NO.sub.2 treatment).

[0078] FIG. 14(B) shows the current response with time to air and then to 2 ppm SO.sub.2 of (a) an SO2 sensor without an ozone filter, and (b) an SO2 sensor with an ozone filter comprising 10% by weight of Mn2O3 powder in binder (after NO2 treatment).

[0079] FIG. 15A shows the current response with to 2 ppm NO.sub.2 and FIG. 15B shows the current response to 500 ppb O.sub.3 of sensors (a) according to FIG. 2B (unfiltered) and (b) sensors according to FIG. 2A in which the filters were formed by 40 mg of powdered Mn.sub.2O.sub.3 mixed with 220 mg of 500 μm PTFE particles (i.e. 16% by weight of Mn.sub.2O.sub.3 in binder).

[0080] FIG. 16A shows the current response to 2 ppm NO.sub.2 and FIG. 16B shows the current response to 500 ppb O.sub.3 of sensors (a) according to FIG. 2B (unfiltered) and (b) sensors according to FIG. 2C in which the filters were formed by 25 mg of powdered Mn.sub.2O.sub.3 deposited onto a membrane (15 mg.Math.cm.sup.−2).

[0081] FIG. 17 shows the current response with time to a range of concentrations of NO.sub.2 of sensors (a) according to FIG. 2B (unfiltered), (b) according to FIG. 2A with a filter comprising 40 mg of powdered Mn.sub.2O.sub.3 mixed with 220 mg of 500 μm PTFE particles and (c) according to FIG. 2C with 25 mg of powdered Mn.sub.2O.sub.3 deposited onto a membrane (15 mg.Math.cm.sup.−2).

[0082] FIG. 18A is a plot of the variation with time in cross-sensitivity to O.sub.3 (as a fraction of the sensitivity to NO.sub.2) of a sensor according to FIG. 2A (filtered) containing 500 mg of powdered Mn.sub.2O.sub.3.

[0083] FIG. 18B is a similar plot to FIG. 18A of (a) a sensor according to FIG. 2A with a filter comprising 40 mg of powdered Mn.sub.2O.sub.3 mixed with 220 mg of 500 μm PTFE particles, (b) a sensor according to FIG. 2A with 45 mg of powdered Mn.sub.2O.sub.3 mixed with 235 mg of 100 μm PTFE particles and (c) a sensor according to FIG. 2C with 100 mg of powdered Mn.sub.2O.sub.3 deposited onto a membrane (30 mg.Math.cm.sup.−2).

DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS

[0084] FIG. 1 is a schematic cross-sectional view of gas sensing apparatus according to the invention in the form of a gas sensor (2) having a housing (16), a (single) sensing working electrode (11), a counter electrode (13), a reference electrode (14), an additional working electrode (35) for correcting for baseline drift, and a body or reservoir of electrolyte, mainly held in wetting filters (15). The housing has an inlet (17) through which the working electrode (11) is placed in communication with a sample gas (e.g. the atmosphere), through an ozone filter (19).

[0085] A series of circular separator discs, wicks or wetting filters (15) separate the working electrode from the counter electrode and the reference electrode. The circular separator discs, wicks or wetting filters (15) are made of a hydrophilic, non-conductive material permeable to the electrolyte which functions to transport electrolyte by capillary action. Typically the material is glass fibre. The circular separator discs, wicks or wetting filters serve to ensure that each of the electrodes is in contact with the electrolyte.

[0086] The working electrode (11) is a carbon electrode and typically comprise a catalytic layer of carbon mixed with polytetrafluoroethylene (PTFE) binder which is then bonded to a gas porous, but hydrophobic, PTFE support to allow gas support to the catalyst, i.e. the electrode active material, but avoid electrolyte leakage or flooding of the electrode. The carbon electrodes can be manufactured using common conventional technologies such as pressing, screen printing, inkjetting and spraying a carbon slurry onto a porous membrane. Here the working electrode catalyst will typically have a diameter of 14 mm. A mixture of carbon and microparticulate polytetrafluoroethylene (PTFE) is sintered and is preferably prepared by pressing the resulting mixture onto a support in the form of a gas porous membrane, such as PTFE sheet. Where carbon is pressed at the normal pressure used in the field, i.e. around 200 kg/cm.sup.2, the amount of catalyst is preferably between 5 and 30 mg per cm.sup.2 of electrode surface area. Preferably the binder is a Fluon matrix (Fluon is a Trade Mark) of around 0.002 ml per cm.sup.2. Other electrodes used in the gas sensing apparatus of the invention, such as the Platinum black electrodes can be prepared in a similar way.

[0087] An O ring is located at the top of the porous membrane (21) and acts to seal the sensor and to aid compressing the stack of components when the sensor is sealed. Also present are a number of platinum strips that serve to connect each electrode to one of the terminal pins (24) provided at the base of the sensor. Closing housing is a dust filter (25) to prevent dust and other foreign objects from entering.

[0088] The ozone filter (19) is formed of MnO.sub.2 powder which has been treated with NO.sub.2 as described below and the sensor is useful to sense NO.sub.2 in the presence of gases such as SO.sub.2, CO, NO, NH.sub.3, H.sub.2 and O.sub.3, which were it not for the choice of electrode and the presence of the ozone filter, would otherwise interfere with the measurement.

[0089] FIGS. 2A and 2B are together a schematic cross-sectional diagram of gas sensing apparatus according to the invention in the form of two individual gas sensors (2) (shown in FIG. 2A) and (4) (shown in FIG. 2B), each having its own housing (16), working electrode, counter electrode (13), reference electrode (14), additional electrode (35) and electrolyte held mainly in wetting filters (15). Together, these form a gas sensing apparatus (30) according to the invention. These gas sensors have similar stacked structures.

[0090] The first gas sensor (2) has an ozone filter (19) in internal chamber (23) above the working electrode (11A), functioning as the first working electrode. Again, the ozone filter is prepared by the method set out below. The second gas sensor (4) has a working electrode (11B), which is the same as the working electrode (11A) of the first gas sensor (4) but functions as the second working electrode, and has no ozone filter in the internal chamber (23). The two working electrodes (11A), (11B) correspond to those described above in relation to FIG. 1.

[0091] The embodiment of FIGS. 2A and 2B is especially useful for detecting NO.sub.2 in the presence of ozone, or for detecting ozone, or both. This is because ozone is removed from the sample gas which penetrates the inlet of the first gas sensor (20), before NO.sub.2 is sensed by the working electrode, and so the first gas sensor gives a measurement of the concentration of NO.sub.2 only, but ozone is not filtered from the gas received through the inlet of the second gas sensor (22) and so the second gas sensor gives a signal which is indicate of the concentration of both NO.sub.2 and ozone. The two signals can therefore be processed to independently determine the concentration of NO.sub.2 and ozone and therefore to measure either NO.sub.2, or ozone, or both.

[0092] In some of the embodiments discussed below, the sensor of FIG. 2A is replaced with the sensor of FIG. 2C. The sensor of FIG. 2C corresponds to the sensor of FIG. 2A except that there is no filter powder received in the chamber (23) and instead there is an additional microporous PTFE membrane (25) having a microporous solid filter layer (27) comprising Mn.sub.2O.sub.3 formed thereon between the inlet (17) and the first working electrode (11A). The filter layer can be formed on the inward or outward surface of the membrane (25). In some embodiments, there is a Mn2O3 powder present in the chamber (23) in addition to the solid filter layer (27).

[0093] FIG. 3 is a schematic cross-sectional view of a further embodiment of gas sensing apparatus (6) in which the first (11A) and second (12A) working electrodes share a common counter electrode (13), a common reference electrode (14) and a common body or reservoir of electrolyte, again mainly held in wetting filters (15). In this case, the housing (16) has two inlets (17) and (18), which place the second working electrode (11) and the O.sub.3 filter (19) in direct communication with the sample gas (e.g. the atmosphere) in parallel. Inlets (17) and (18) can be capillary inlets, i.e. inlets which are sized so that they control the rate of sample gas reaching the electrodes so that the gas sensing apparatus is diffusion limited, A central portion (20) divides a cavity defined by the housing (16) and a porous membrane (21) into first (22) and second (23) internal chambers. The first and second working electrodes are located in the same horizontal plane and are situated underneath the porous membrane (21). The second working electrode (11B) is situated underneath the first internal chamber (22) and the first working electrode (12) is situated underneath the second internal chamber (23). The ozone filter (19) is located in the second internal chamber (23), adjacent/on top of part of the porous membrane (21) covering the first working electrode (12). The ozone filter (19) covers the surface of the first working electrode (12) that would, absent the filter, be exposed to the sample gas.

[0094] Example circuitry for the gas sensing apparatus of the invention in the embodiment of FIG. 3 is shown in FIG. 4, where WE1 is the second working electrode (11B) and WE2 is the first working electrode (11A). This circuitry could also be used for either of the individual sensors making up the gas sensing apparatus of the invention in the embodiment of FIG. 2, where, for example, WE1 is the second working electrode and WE2 is the first additional working electrode, or WE1 is the first working electrode and WE2 is the second additional working electrode. It could also be used for a simple individual sensor according to FIG. 2A (or 2B) but lacking an additional working electrode in which case WE1 is the working electrode and is the only one recorded.

Forming the Ozone Filter

[0095] In order to form the ozone filter, Mn.sub.2O.sub.3 filter powder (99% purity) is treated by passing a flow of NO.sub.2 gas therethrough. With reference to FIG. 5, the Mn.sub.2O.sub.3 filter powder is inserted into a gas washing bottle (or Drechsel's bottle) which is then connected through the inlet to a source of NO.sub.2 gas. The Mn.sub.2O.sub.3 may first be also be mixed with PTFE powder, acting as binder.

[0096] In an illustrative example, 20 g of Mn.sub.2O.sub.3 or 30 g or 50 g of Mn.sub.2O.sub.3/PTFE filter powder was treated in a 250 ml gas washing bottle. The NO.sub.2 gas was passed through the filter powder at a concentration of 100 ppm and a flow rate of 0.5 l.Math.min.sup.−1 for 2 hours.

[0097] By monitoring the concentration of NO.sub.2 gas leaving the bottle, we therefore calculated the amount of NO.sub.2 which was adsorbed before the Mn.sub.2O.sub.3 was saturated. This experiment was repeated. It was found that 1.78×10.sup.−8 moles of NO.sub.2 were adsorbed per gram of Mn.sub.2O.sub.3. Using the value of surface area of the MnO.sub.3 powder obtained from BET analysis of 2.42 m.sup.2/g we estimated that 2.41×10.sup.12 molecules of NO.sub.2 were adsorbed per cm.sup.2 of Mn.sub.2O.sub.3 surface. This is consistent with the values repeated in the paper J. Phys. Chem. C 2014, 118, 23011-23021 where the number of NO.sub.2 molecules adsorbed on TiO.sub.2 are in the order of 10.sup.13 molecules/cm.sup.2.

[0098] The resulting material is advantageous in that it filters ozone with minimal degradation of the NO.sub.2 signal compared to untreated Mn.sub.2O.sub.3. The layer remains gas permeable. Furthermore, it filters ozone efficiently, enabling a relatively low proportion of Mn.sub.2O.sub.3 mixed with PTFE to function effectively, reducing cost. An image of the resulting Mn.sub.2O.sub.3 coated PTFE particles is shown in FIG. 13.

[0099] A mixture of Mn.sub.2O.sub.3 and PTFE particles having a size of having a size range of 710 μm to 1500 μm was prepared by manually mixing Mn.sub.2O.sub.3 powder with PTFE in a glass container until all Mn.sub.2O.sub.3 powder coats the PTFE. In some of the examples below, Mn.sub.2O.sub.3 and PTFE were mixed in a weight ratio of 1:10. The resultant mixture is then sieved through a 710 micron sieve stack using a motorised mechanical sieve for 1 hour. The portion that did not fall through the 710 micron sieve is then collected while the remaining material that did fall through the 710 micron sieve is discarded.

[0100] In some examples below, Mn.sub.2O.sub.3 and PTFE were also mixed in a weight ratio of 1.6:8.4 (16% by weight). In those examples, the resultant mixture was then sieved through a 500 micron sieve stack using a motorised mechanical sieve for 1 hour. The portion that did not fall through the 500 micron sieve was then collected while the remaining material that did fall through the 500 micron sieve was discarded.

[0101] In the examples where the Mn.sub.2O.sub.3 was mixed with 100 micron PTFE in a weight ratio of 1.6:8.4 (16% by weight), resultant mixture was then thoroughly shaken using a motorised mechanical sieve for 1 hour.

Filter in the Form of a Solid Layer of Powdered Mn.sub.2O.sub.3:

[0102] For examples 13 to 15, solid filter layers of powdered Mn.sub.2O.sub.3 were formed with a diameter of 14 mm and 19 mm. A mixture of powdered Mn.sub.2O.sub.3 and microparticulate polytetrafluoroethylene (PTFE) was sintered and the resulting mixture was pressed onto a support PTFE sheet, acting as a gas porous membrane. Mn.sub.2O.sub.3 was pressed at the normal pressure used in the field, i.e. around 400-600 kg/cm.sup.2, and the amount of Mn.sub.2O.sub.3 was in the range 15 to 30 mg per cm.sup.2 of surface area. The Mn.sub.2O.sub.3 was mixed with a Fluon matrix (Fluon is a Trade Mark) of around 0.0065 ml per cm.sup.2, which contains PTFE which acted as binder. In these examples Fluon with a PTFE particle size in the range of 200-300 microns diameter was used. Mn.sub.2O.sub.3 was 74% by weight of the solid microporous layer, with the balance being PTFE.

EXPERIMENTAL SECTION

[0103] The sensors used in the following experiments were tested on standard potentiostatic circuit boards (FIG. 4). Generally, the sensors were stabilised for a minimum of 2 days before being tested. All the experiments involving gas tests were controlled using computer controlled valves and digital mass flow controllers. Sensor output data collection is also made using a computer. The NO.sub.2 gas tests were made using a certified 100 ppm bottle (Air Products, UK) and filtered air, except that a 10 ppm bottle (Air Products, UK) was used for Experiment 2 and the linearity test of Experiment 7. The ozone was obtained using a calibrated ozone generator equipped with an internal analyser (Thermo Scientific, Model 49i-PS), except for Experiment 2 and the filter capacity test of Experiment 5 where the ozone was obtained using a calibrated ozone generator (Ultra-violet Products Ltd, SOG-1, Cambridge, UK). During the laboratory tests the sensors were exposed to a gas flow of 0.5 l.Math.min.sup.−1.—

[0104] The following materials were used in the filter powder. These materials were used in each following experiment unless indicated to the contrary.

[0105] Manganese (III) oxide (Mn.sub.2O.sub.3), 99% purity, approx. 325 mesh (44 micrometers) powder, from Sigma Aldrich, product number 377457. BET analysis on the sample gives a surface area of 2.242 m.sup.2/g.

[0106] Manganese (IV) oxide (MnO.sub.2), 99.9% purity, approx. 325 mesh (44 micrometers) powder, from Alpha Aesar, product number 42250. BET analysis on the sample gives a surface area of 2.08 m.sup.2/g.

[0107] PTFE binder (Fluon PTFE G307 (median particle size 500 to 1500 microns) (Fluon is a trade mark). The powder was sieved to collect only particles with a size of at least 710 microns. Also: Fluon PTFE G201 (median particle size 500 microns) and Fluon PTFE G204 (median particle size 100 microns).

[0108] The working electrodes were made from carbon graphite (particle size <20 μm, Aldrich, product code 282863).

[0109] Calibration of the sensors according to the invention was carried out as follows. The sensor output is given in nA. Amperometric gas sensors have a linear output with target gas analyte concentration. This makes possible to use a simple calibration procedure where the relation between the sensor output and the gas concentration is determined by exposing the sensor to a known concentration of gas analyte. For this application the sensitivity to both the first and the first working electrodes is determined for each of the target gases, NO.sub.2 and O.sub.3. The sensor output can then be used to calculate the concentration of NO.sub.2 and O.sub.3.

[0110] If the following parameters are defined as follows: [0111] i.sub.1 is the current observed on the second working electrode. [0112] I.sub.2 is the current observed on the first working electrode. [0113] S.sub.1(NO2) is the second working electrode sensitivity to NO.sub.2. [0114] S.sub.1(O3) is the second working electrode sensitivity to O.sub.3. [0115] S.sub.2(NO2) is the first working electrode sensitivity to NO.sub.2. [0116] S.sub.2(O3) is the first working electrode sensitivity to O.sub.3. [0117] C.sub.(NO2) is the NO.sub.2 analyte concentration to determine. [0118] C.sub.(O3) is the O.sub.3 analyte concentration to determine.

[0119] Then, by definition and taking into account the linear relationship between the sensor output and the analyte concentration, the following can be written:

i.sub.1=S.sub.1(NO2).Math.C.sub.(NO2)+S.sub.1(O3).Math.C.sub.(O3)

i.sub.2=S.sub.2(NO2).Math.C.sub.(NO2)+S.sub.2(O3).Math.C.sub.(O3)

[0120] The ozone filter on top of the first working electrode means that S.sub.2(O3)=0. So C.sub.(NO2) can be calculated using the simple relation:

C.sub.(NO2)=i.sub.2/S.sub.2(NO2)

[0121] The NO.sub.2 analyte concentration being now known it is then possible to calculate the O.sub.3 analyte concentration using the second working electrode output:

C.sub.(O3)=(i.sub.1−S.sub.1(NO2).Math.C.sub.(NO2))/S.sub.1(O3)

or C.sub.(O3)=(i.sub.1−S.sub.1(NO2).Math.(i.sub.2/S.sub.2(NO2)))/S.sub.1(O3)

Experiment 1

[0122] In a first example, sensors according to FIG. 2A (filtered) were formed with Mn.sub.2O.sub.3, 99% purity (without PTFE) with and without the nitrogen dioxide pretreatment step described above. The response of these sensors and sensors according to FIG. 2B (i.e. unfiltered) to ozone and NO.sub.2 was compared. The unfiltered sensor is commercially available under the trade name OX-A421, manufactured and sold by Alphasense Limited of Great Notley, United Kingdom.

[0123] For these and subsequent experiments the first working electrode, the second working electrode, the additional electrode and the reference electrode are made of carbon graphite and the counter electrode is made of platinum black.

[0124] FIG. 6A shows the current response to 0.5 ppm O.sub.3 of (a) unfiltered sensors and (b) sensors in which the filter is 500 mg of untreated Mn.sub.2O.sub.3.

[0125] FIG. 6B shows the current response to 2 ppm NO.sub.2 of (a) a sensor in which the filter is 500 mg of untreated Mn.sub.2O.sub.3 and (b) a sensor in which the filter is 500 mg of Mn.sub.2O.sub.3 treated with NO.sub.2 as described above.

[0126] FIG. 6C shows the current response to 0.5 ppm O.sub.3 of (a) unfiltered sensors and (b) sensors in which the filter is 500 mg of Mn.sub.2O.sub.3 treated with NO.sub.2 as described above.

[0127] It is apparent from the results that the untreated Mn.sub.2O.sub.3 filter efficiently removes O.sub.3. No signal is observed in the presence of 500 ppb of O.sub.3 (FIG. 6A, curve (b)) but a sensor which differs only by the omission of the filter gives a clearly defined current response. However, it can be seen that the untreated filters added to the sensors to remove O.sub.3 do affect the NO.sub.2 signal. It is apparent from FIG. 6B, curve (a) that the resulting signal is not adequate for reliable sensing of NO.sub.2. However, we have found that a well-defined signal for NO.sub.2 can be obtained with the NO.sub.2 pretreatment step set out above. It can be seen from FIG. 6C, curve (b), that the treated Mn.sub.2O.sub.3 remains effective for the filtering of NO.sub.2.

Experiment 2

[0128] In this example, sensing apparatus comprised a sensor according to FIG. 2A in which the filter (59) in the chamber (52) was formed with 500 mg Mn.sub.2O.sub.3, (without PTFE), and a sensor according to FIG. 2B (unfiltered).

[0129] For this and all subsequent experiments, the Mn.sub.2O.sub.3 had been treated with NO.sub.2 gas described above.

[0130] FIG. 7A illustrates the output current, over time, where the unfiltered sensor (trace (a)) and filtered sensor (trace (b)) are exposed in turn to zero air (zero air is air that has been filtered to remove most gases (NO.sub.2, NO, CO, SO.sub.2, O.sub.3, etc.) and is used as a calibrating gas for zero concentration), then 1 ppm NO.sub.2, then zero air, then 1 ppm O.sub.3 and then zero air and then a mixture of 1 ppm NO.sub.2 and 1 ppm O.sub.3. FIG. 7B shows the calculated O.sub.3 concentration (trace (a)) and NO.sub.2 concentration (trace (b)) which these traces represent.

[0131] This figure shows that with a mixture of NO.sub.2 and O.sub.3, the filtered sensor senses only NO.sub.2, whereas the unfiltered sensor detects both gases and, in the presence of both gases simultaneously, the output of the unfiltered sensor corresponds to the sum of the output expected for each of the gases. Accordingly, Mn.sub.2O.sub.3 is suitable as a filter material to remove ozone without affecting the signal due to NO.sub.2.

Experiment 3

[0132] Sensors according to FIG. 2A (filtered) were formed with filters comprising (a) 500 mg of Mn.sub.2O.sub.3, and (b) 450 mg of MnO.sub.2 were exposed to 2 ppm NO.sub.2 for 10 minutes and then to 2 ppm nitrogen monoxide NO for 10 minutes. FIG. 8 is a plot of the variation with time in cross-sensitivity of the sensors (a) and (b) to NO, i.e. the current response to NO as a fraction of the current response to a corresponding concentration (in this case ppm) of NO.sub.2.

[0133] This shows that the cross-sensitivity to NO is systematically lower with Mn.sub.2O.sub.3 than MnO.sub.2, and that this improvement persists.

Experiment 4

[0134] Sensors according to FIG. 2A (filtered) were formed with ozone filters comprising mg of powdered Mn.sub.2O.sub.3 mixed with 225 mg of PTFE particles having a size range of 710 μm to 1500 μm (i.e. 10% by weight of in Mn.sub.2O.sub.3 binder), and exposed to zero air, then 0.5 ppm O.sub.3, then zero air. FIG. 9 shows the output current with time and demonstrates that ozone is efficiently filtered out by Mn.sub.2O.sub.3 mixed with PTFE binder (10% by weight of Mn.sub.2O.sub.3 in PTFE binder particles).

Experiment 5

[0135] The ozone filtering capacity of the sensors was further tested and FIG. 10 shows the current response to 1 ppm of O.sub.3 (2 ppm of O.sub.3 is relatively high in comparison to levels typically measured in environmental monitoring) of in FIG. 10A (a) a sensor according to FIG. 2A (filtered) in which the ozone filter comprised 500 mg of powdered Mn.sub.2O.sub.3 (not mixed with binder) and (b) a sensor according to FIG. 2B (unfiltered), and in FIG. 10B (a) a sensor according to FIG. 2A (filtered) in which the ozone filter comprised 25 mg of powdered Mn.sub.2O.sub.3 mixed with 225 mg of PTFE particles having a size range of 710 μm to 1500 μm (i.e. 10% by weight of in Mn.sub.2O.sub.3 binder) and (b) a sensor according to FIG. 2B (unfiltered). In the case of FIG. 10A the experiment was continued for 14 days and in the case of FIG. 10B for 10 hours.

[0136] These results again demonstrate that Mn.sub.2O.sub.3 powder, whether unmixed or mixed with mg of PTFE particles having a size range of 710 μm to 1500 μm forms an efficient ozone filter.

Experiment 6

[0137] Experiments were carried out to assess the cross sensitivity of the sensors to common interferents. Table 1, below, shows the cross sensitivity of sensors formed with (a) mg of powdered Mn.sub.2O.sub.3 and (b) 25 mg of powdered Mn.sub.2O.sub.3 mixed with 225 mg of PTFE particles having a size range of 710 μm to 1500 μm (i.e. 10% by weight of Mn.sub.2O.sub.3 in binder) to specific gases, relative to NO.sub.2, at the concentrations stated. Two sensors of each type were formed and tested.

TABLE-US-00001 TABLE 1 500 mg Mn.sub.2O.sub.3 25 mg Mn.sub.2O.sub.3 GAS ppm Sensor 1 Sensor 2 Sensor 1 Sensor 2 SO2 5 1.53% 0.81% 0.86% 0.79% CO 5 1.25% 0.85% 1.41% 1.39% H2 100 0.24% 0.04% 0.06% 0.00% CO2 50000 0.00% 0.00% 0.00% 0.00% NH3 20 0.42% 0.18% 0.09% 0.06%

[0138] Further experiments demonstrated that the quality of ozone filtering was unsatisfactory if the Mn.sub.2O.sub.3 powder was reduced to 5% or less by mass of the combined mixture of Mn.sub.2O.sub.3 and PTFE binder particles. Mixture of Mn.sub.2O.sub.3 powder with at least the same mass, and ideally more, of PTFE particles leads to the Mn.sub.2O.sub.3 coating the particles rather than forming solid masses.

[0139] For unmixed Mn.sub.2O.sub.3 powder (as specified in the Experimental Section above) we found that the ratio of the active surface area of filter material (by BET analysis) to the cross-sectional area of the filter (perpendicular to the gas path through the filter) was about 0.85 m.sup.2 per cm.sup.2. For Mn.sub.2O.sub.3 mixed with PTFE binder to 10% by mass of the combined mixture of Mn.sub.2O.sub.3 and binder, this ratio was 0.025 m.sup.2 per cm.sup.2 and for Mn.sub.2O.sub.3 mixed with PTFE binder to 8% by mass of the combined mixture it was 0.02 m.sup.2 per cm.sup.2.

Experiment 7

[0140] FIG. 11 shows the current response to a range of concentrations of NO.sub.2 of sensors according to FIG. 2A (filtered) with filters comprising (A) 500 mg of powdered Mn.sub.2O.sub.3 and (B) 25 mg of powdered Mn.sub.2O.sub.3 mixed with 225 mg of PTFE particles having a size range of 710 μm to 1500 μm (i.e. 10% by weight of Mn.sub.2O.sub.3 in binder). The results show good sensitivity and linearity with NO.sub.2 concentration with both the mixed with binder and unmixed filters.

Experiment 8

[0141] FIG. 12 shows the current response to 2 ppm NO.sub.2 of sensors according to FIG. 2A (filtered) with filters comprising, in FIG. 12A, 500 mg of powdered Mn.sub.2O.sub.3 (a) 30 days after manufacture and (b) 173 days after manufacture and, in FIG. 12B, 25 mg of powdered Mn.sub.2O.sub.3 mixed with 225 mg of PTFE particles having a size range of 710 μm to 1500 μm (i.e. 10% by weight of Mn.sub.2O.sub.3 in binder), (a) 19 days after manufacture and (b) 187 days after manufacture.

[0142] The results demonstrate that sensors using NO.sub.2 treated Mn.sub.2O.sub.3 powder, with or without dilution, shows a good and reproducible response to NO.sub.2 6 months after NO.sub.2 treatment.

Experiment 9

[0143] Table 2 below shows the change with time (in days) of the cross sensitivity to 2 ppm NO (i.e. the ratio of the current response to 2 ppm NO to the current response to 2 ppm NO.sub.2) of sensors according to FIG. 2A (filtered) with filters comprising 25 mg of powdered MnO.sub.2 mixed with 225 mg of the PTFE particles having a size range of 710 μm to 1500 μm; 450 mg of powdered MnO.sub.2 (not mixed with binder), 25 mg of powdered Mn.sub.2O.sub.3 mixed with 225 mg of PTFE particles having a size range of 710 μm to 1500 μm (i.e. 10% by weight of Mn.sub.2O.sub.3 in binder), and 500 mg of Mn.sub.2O.sub.3 (i.e. not mixed with binder) (a) 19 days after manufacture and (b) 187 days after manufacture.

TABLE-US-00002 TABLE 2 Age 25 mg 450 mg 25 mg 500 mg (days) ppm MnO.sub.2 MnO.sub.2 Mn.sub.2O.sub.3 Mn.sub.2O.sub.3 10 2 1.6 9.7 1.3 1.7 110 2 5.7 13.9 2.2 4.5 210 2 9.2 13.3 4.4 5.7

[0144] This shows that the cross-sensitivity to NO is lower when the ozone filter comprises powdered Mn.sub.2O.sub.3 than when the ozone filter comprises powdered MnO.sub.2, and that the sensors in which Mn.sub.2O.sub.3 is mixed with PTFE give the best (lowest) cross-sensitivity to NO. Of particular benefit, the NO cross-sensitivity increases with time but this deterioration is reduced with powdered Mn.sub.2O.sub.3 in comparison to powdered MnO.sub.2 and is minimised when powdered Mn.sub.2O.sub.3 is mixed with PTFE.

Experiment 10

[0145] Experiments were carried out to assess the temperature sensitivity of sensors according to FIG. 2B (unfiltered) (columns 2 and 3 below), and according to FIG. 2A (filtered) in which the ozone filter was 10% by weight of Mn.sub.2O.sub.3 powder in binder (after NO.sub.2 treatment), 10% by weight of Mn.sub.2O.sub.3 powder in binder (without NO.sub.2 treatment) and NO.sub.2 treated Mn.sub.2O.sub.3 powder without binder (columns 4 through 6 below, respectively). The current response due to 2 ppm NO.sub.2 at different temperatures, as a percentage of the current response due to 2 ppm NO.sub.2 at 20° C. is shown in Table 3 below. The results show that, surprisingly, the ozone filter in which Mn.sub.2O.sub.3 powder was pretreated with NO.sub.2 and mixed with PTFE has a much better current response at low temperatures than unmixed Mn.sub.2O.sub.3 powder and accordingly has a substantially better operating temperature range.

TABLE-US-00003 TABLE 3 10 w/w % 10 w/w % 100 w/w % Temp. No No Mn.sub.2O.sub.3 NO.sub.2 Mn.sub.2O.sub.3 not Mn.sub.2O.sub.3 NO.sub.2 ° C. filter filter treated treated treated −30 70 68 70 39 −2 −20 75 73 80 56 6 −10 82 79 88 75 45 0 89 87 94 90 85 10 95 94 98 98 97 20 100 100 100 100 100 30 105 104 104 101 103 40 110 107 103 99 107 50 112 117 102 100 128

Experiment 11

[0146] In order to demonstrate that the ozone filter is useful with analytes other than NO.sub.2, we modified a commercial electrochemical sensor for SO.sub.2, having gold/ruthenium working and reference electrodes and a platinum black counter electrode (brand SO2-A4 available from Alphasense Limited, Great Notley, UK) by replacing an H.sub.2S filter with 25 mg of Mn.sub.2O.sub.3 powder mixed with 225 mg PTFE particles having a size range of 710 μm to 1500 μm, treated with NO.sub.2 as above.

[0147] FIG. 14(A) shows the current response with time to air and then to 0.5 ppm O.sub.3 of (a) an SO.sub.2 sensor without an ozone filter, and (b) an SO.sub.2 sensor with an ozone filter comprising 10% by weight of Mn.sub.2O.sub.3 powder in binder (after NO.sub.2 treatment).

[0148] FIG. 14(B) shows the current response with time to air and then to 2 ppm SO.sub.2 of (a) an SO.sub.2 sensor without an ozone filter, and (b) an SO.sub.2 sensor with an ozone filter comprising 10% by weight of Mn.sub.2O.sub.3 powder in binder (after NO.sub.2 treatment).

[0149] The results demonstrate that the ozone filter can be used to remove ozone from a gas sample in an SO.sub.2 sensor. Again, O.sub.3 may be measured with a first sensor which has an electrode which is sensitive to SO.sub.2 and O.sub.3 and ozone filter according to the invention and a second sensor, also having an electrode which is sensitive to SO.sub.2 and O.sub.3, but no ozone filter, and comprising output signals.

Experiment 12

[0150] FIG. 15A shows the current response with to 2 ppm NO.sub.2 and FIG. 15B shows the current response with time to 500 ppb O.sub.3 of sensors (a) according to FIG. 2B (unfiltered) and (b) sensors according to FIG. 2A in which the filters were formed by 41.6 mg of powdered Mn.sub.2O.sub.3 mixed with 218.4 mg of 500 μm PTFE particles (i.e. 16% by weight of Mn.sub.2O.sub.3 in binder) and pretreated with NO.sub.2 as described above.

[0151] No signal is observed in the presence of 500 ppb of O.sub.3 (FIG. 15B, curve (b)) but a sensor which differs only by the omission of the filter gives a clearly defined current response (FIG. 15B, curve (a)). Thus we have found that a well-defined signal for NO.sub.2 can be obtained with the sensors in which the filters were formed by 40 mg of powdered Mn.sub.2O.sub.3 mixed with 220 mg of 500 μm PTFE particles.

Experiment 13

[0152] FIG. 16A shows the current response with time to 2 ppm NO.sub.2 and FIG. 16B shows the current response with time to 500 ppb O.sub.3 of sensors (a) according to FIG. 2B (unfiltered) and (b) sensors according to FIG. 2C in which the filters were formed by 25 mg of powdered Mn.sub.2O.sub.3 mixed with Fluon (Fluon is a trade mark) and deposited onto a membrane (15 mg.Math.cm.sup.−2).

[0153] Fluon comprises PTFE and a solvent and the resulting solid layer has about 26% PTFE by mass. Most of the Mn.sub.2O.sub.3 is therefore not associated with the PTFE particles and the purpose of the PTFE is to provide some porosity, as well as to make the material easier to handle in a manufacturing setting than unmixed Mn.sub.2O.sub.3.

[0154] No signal is observed in the presence of 500 ppb of O.sub.3 (FIG. 16B, curve (b)) but a sensor which differs only by the omission of the filter gives a clearly defined current response (FIG. 16B, curve (a)). We have therefore found that a well-defined signal for NO.sub.2 can be obtained with the sensors in which the filters were formed by a solid layer of 25 mg of powdered Mn.sub.2O.sub.3 deposited onto a membrane (15 mg.Math.cm.sup.−2).

Experiment 14

[0155] FIG. 17 shows the current response to a range of concentrations of NO.sub.2 of sensors (a) according to FIG. 2B (unfiltered), (b) according to FIG. 2A with a filter comprising 40 mg of powdered Mn.sub.2O.sub.3 mixed with 220 mg of 500 μm PTFE particles and (c) according to FIG. 2C with 25 mg of powdered Mn.sub.2O.sub.3 (again with about 26% PTFE by mass) and deposited onto a membrane (15 mg.Math.cm.sup.−2).

[0156] This demonstrates that powdered Mn.sub.2O.sub.3 pressed into a solid layer may form an effective ozone filter.

Experiment 15

[0157] Sensors according to FIG. 2A (filtered) were formed with filters comprising 500 mg of Mn.sub.2O.sub.3(FIG. 18A) and (a) 40 mg of powdered Mn.sub.2O.sub.3 mixed with 220 mg of 500 μm PTFE particles, (b) 45 mg of powdered Mn.sub.2O.sub.3 mixed with 235 mg of 100 μm PTFE particles and (c) 100 mg of powdered Mn.sub.2O.sub.3, mixed with Fluon (Fluon is a trade mark), and deposited onto a membrane (30 mg.Math.cm.sup.−2) (FIG. 18B). In this case, the mass fraction of the deposited layer formed by PTFE is 15%. In contrast to Experiment 13, the Mn.sub.2O.sub.3 was pretreated with NO.sub.2 by the process described above.

[0158] The sensors were exposed to 2 ppm NO.sub.2 for 10 minutes and then to 500 ppb ozone O.sub.3 for 5 minutes. FIG. 18 is a plot of the variation with time in cross-sensitivity of the sensors to O.sub.3, i.e. the current response to O.sub.3 as a fraction of the current response to a corresponding concentration (in this case 2 ppm) of NO.sub.2.

[0159] These results demonstrate that Mn.sub.2O.sub.3 powder, whether unmixed, mixed with various PTFE particles sizes or as a solid layer deposited onto a membrane, forms an efficient ozone filter.

[0160] Although in the examples shown which use a solid layer, the layer comprises some PTFE in addition to Mn.sub.2O.sub.3, it is provided simply to ensure that the solid layer is microporous and can therefore be penetrated by analyte gas. Alternatively, Mn.sub.2O.sub.3 may be deposited by other means to give a microporous structure, for example by depositing Mn.sub.2O.sub.3. microparticles without binder, or screen printing optionally with glass or other particles. The PTFE is not required.

[0161] Furthermore, in further example embodiments, Mn.sub.2O.sub.3 powder, whether unmixed, or mixed with binder as described above, may be used in combination with a microporous solid Mn.sub.2O.sub.3 layer, to give a more robust sensor, for example for use in atmospheres with a particularly high concentration of O.sub.3.

CONCLUSIONS

[0162] In the case of gas sensing apparatus for detecting NO.sub.2 and/or O.sub.3, the results demonstrate that Mn.sub.2O.sub.3 is useful as an ozone filter, whether as a powder or as a solid microporous layer. The cross sensitivity to NO can be reduced by mixing the Mn.sub.2O.sub.3 powder with binder, or by treating it with sufficient NO.sub.2. Although these mixing or treatment steps could compromise ozone filtering, we have found that they have a greater effect in reducing cross sensitivity to NO and that it is therefore possible to obtain an efficient ozone filter with low cross sensitivity to NO.