AU626857B2 - Oxygen probe assembly - Google Patents

Description

OPI DATE 16/10/89 WC AOJP DATE 09/11/89 APPLN. I D 3413541 89

PCT

PCT NUMBER PCT/AU89/00113 INTERNATIONAL APPLICATION PUBLISHED UNDER THE PATENT COOPERATION TREATY (PCT) (51) International Patent Classification 4 1) 16 2 Ppl iation Number: WO 89/ 092 G01N 27/56, 27/58, 27/46 Al (4)I analo c' (43)1 I ic D 5 ttober 1989 (05. (21) International Application Number: PCT/AU89/001 13 (22) Interaational Filing Date: (31) Priority Application Number:, (32) Priority Date: (33) Priority Country: 21 March 1989 (2 1.03,89) PI 7373 22. March 1988 (22.03.88) (74) Agents: CORBETT, Terence, G. et al.; Davies Collison, I. Little Collins Street, Melbourne, VIC 30001

(AU).

(81) Designated States: AT (European patent), AU, BE (Eu- -i ropean paten CH (European patent), DE (Euro- I pean patent), FR (European patent), GB (European patent), IT (European patent), JP, LU (European patent), NL (European patent), SE (European patent), us.

Published With international search report, (71) Applicant (for all designated States exvcept US): COM- MONWEALTH SCIENTIFIC AND INDUSTRIAL RESEARCH ORGANIZATION [AU/AU]; Limestone Avenue, Campbell,, ACT 2-601 (AU), (72) Inventor; and Inventor/Applicant (for US only) ,BANNISTER, Michael, John [AU/AUI; 7 Montclair Avenue, Glen Waverley, VIC 3150 (AU), (54)Tltlell OXYGEN PROBE ASSEMI3LY 07?) Abstract I V 41 An oxypnr probe assenmbly comprising an oxypo en asor and means 3) for directdng a test gas and reference $as to the appropriate electrodes 1) of the snsor,'aaid assemibly having aisocated ther~ewith catalyst maeans (13, 6) or* ranged so that the test gas, thu r~eference gas, or both said goses are separately~ contacted with the catalyst means before ap* proaching the working iturrace(s) or the electrodci, whereby combustibles in the gases are oxidized; and/or means whereby the test gas, the reference gas, or both of 3ald gases are conveycd to the said working surface(s) by witv of a path (10, 4) which Is of sufficient length to allow the said gas or gases to eottain thermodyniaic equilibrium at the piobe temperature before coming Into conitact with the said working surface~s).

,WO 89/09398 PCT/AU89/00113 -1- OXYGEN PROBE ASSEMBLY This invention is concerned with oxygen probe 23 assemblies which embody solid electrolyte sensors and are used to measure the oxygen potential of gases or molten metals. The invention is particularly concerned with probe assemblies capable of operating accurately at 27 temperatures as low as 300C.

29 Measurement of the oxygen potential of gases and molten metals using solid electrolyte sensors is well 31 documented, For example, a sensor designed primarily for determinations in molten copper is described in Australian 33 patent No. 466,251 and in the corresponding patents, U.S.

No, 4,046,661, United Kingdom No, 1,347,937, Canadian No.

952,983, Belgian No. 782,180, Japanese No.80 17340 and West German Offenlegungsschrift No. 22 18227.0, 37 Modifications, particularly to the electrolyte, to improve the sensor for measurements in gases are described in 39 Australian patent No. 513,552 and its equivalents U.S' No. 4,193,857, United Kingdom No. 1,575,766, Canadian SUBSTITUTE SHEET WO 89/09398 PCT/AU89/00113 -2- 1 No, 1,112,438, West German Offenlegungsschrift document No. 27 54522.8 and Japanese patent application 3 No. 146208/77.

The solid electrolyte oxygen sensor uses the fact that when a solid membrane of a material with good oxygen 7 ion conductivity and negligible electronic conductivity, termed a solid electrolyte, is held with its opposite 9 faces in contact with materials having different oxygen potentials, an emf is established across the membrane. If 11 one of the oxygen-containing materials is the gas or molten metal under investigation and the other is a 13 reference material of known oxygen potential, then the emf is given by the Nernst relationship: RT 2 (reference material) 17 E n (1) 19 nF p 02 (test material) 19 21 where: R the gas constant, 23 T the absolute temperature, n 4 (the number of electrons transferred per oxygen molecule), F the Faraday Constant, and 27 P 02 the oxygen partial pressure.

29 This emf is measured using electrodes, reversible to the 02/02- redox equilibrium, placed in electrical contact 31 with the opposing faces of the solid electrolyte membrane.

33 Australian patent no.466,251 describes various geometrically distinct forms of solid electrolyte oxygen sensor. The most commonly-used form is that of a tube; either open-ended or closed at one end, made entirely from 37 the solid electrolyte. Other designs use the solid

L

WO 89/09398 PCT/AU89/00113 -3- 1 electrolyte as a disc or pellet, sealed in one end of a metal or ceramic supporting tube. In all cases the 3 reference environment, which is generally air, is maintained on one side of the tube (commonly on the inside) and the test environment is exposed to the other side of the tube.

7 Many solid electrolyte materials are known to be 9 suitable for use in oxygen sensors. Examples include zirconia or hafnia, both fully stabilized or partially 11 stabilized by doping with calcia, magnesia, yttria, scandia or one or a number of rare earth oxides and 13 thoria, also doped with calcia, yttria or a suitable rare earth oxide. Australian patent 513,552 discloses the addition of alumina to these solid electrolyte materials to produce a composite solid electrolyte which is 17 particularly suitable for sealing into the end of an alumina tube, thereby making a rugged and leak-tight 19 sensor useful for demanding industrial applications.

Australian patent application No.47828/78 and the 21 corresponding patents or applications U.S. Patent No.

4,240,891, United Kingdom application No, 79 19671, 23 Canadian application No. 329,100, Japanese application No.

69529/79 and West German application No. 29 22947.8 disclose the use of magnesium aluminate spinel as an alternative to alumina, either for the supporting tube or 27 as the inert diluent in the composite solid electrolyte material.

29 The electrodes on solid electrolyte oxygen sensors 31 generally consist of porous coatings of noble metals such as platinum, gold, palladium or silver, or alloys of these 33 elements. For measurements in gases using a gaseous reference an electrode is required on each surface of the WO_ 9/0939 P 13 SWO 89/09398 PCT/AU89/00113 -4- 1 solid electrolyte; for measurements in molten metals an electrode is required only on the reference side of the 3 solid electrolyte, and then only if a gaseous reference is used. If a solid reference, a metal/metal oxide mixture, is used there is no need for a separate noble metal electrode; the solid reference mixture serves as the 7 electrode.

9 The electrodes participate in the exchange reaction between gaseous oxygen molecules and oxygen ions in the 11 solid electrolyte by donating or accepting electrons. The overall equilibrium at each electrode is represented by 13 the equation: 02 (gas) 4e (electrode) 2 02- (electrolyte) (2) 17 The electrodes may also help to catalyse this reaction. For example platinum, the most commonly used 19 electrode material on solid electrolyte oxygen sensors, shows a high catalytic activity for this oxygen exchange 21 reaction at elevated temperatures.

23 In the past, solid electrolyte oxygen sensors with noble metal electrodes have generally been used at temperatures above 600-700oC. As operating temperature is reduced below 600 0 C, the following characteristics become 27 increasingly evident: 29 The total impedance rapidly increases, reaching megohm levels.

31 The response time increases from a fraction of a 33 second to many minutes.

W089/09398 P 8 13 I I1 89/09398 ]PCT/A U89/001 13 1 An emf appears when none is predicted (zero error).

3 The emf is not ideal it does not follow the Nernst equation) even after correction for zero error.

7 The emf may depend on the gas flow rate.

9 The emf under fuel-rich conditions gas mixtures with high CO/CO 2 ratios) is very greatly in 11 error.

13 The increased impedance may be at least partly counteracted by controlling the electrolyte composition, in particular its impurity content, to maximise its ionic conductivity, and by increasing the ratio electrode 17 area/solid electrolyte membrane thickness of the sensor.

Response time and emf accuracy at low temperatures may be 19 enhanced by controlling the physical characteristics of the electrodes, i.e. porosity, particle size, layer 21 thickness, etc. In general this means ensuring a very porous, extremely fine-grained texture in the electrodes, 23 Noble metal electrodes, however, undergo morphology changes due to sintering and grain growth when exposed to higher temperatures, to the detriment of their subsequent low temperature performance.

27 Xn International Patent Application No.

29 PCT/AU84/00013 (WO 84/03149) we described electrode materials for solid electrolyte oxygen sensors which 31 enable such sensors to generate ideal Nernstian) emfs under oxygen-excess gaseous conditions at 33 temperatures substantially below those at which conventional noble metal electrodes begin to show non-ideal behaviour, and as low as 300OC.

|t WO 89/09398 PCT/AU89/00113 -6- 1 These electrode materials are solid solutions in 3 uranium dioxide of one or more other metal oxides with oxygen-metal atom ratios less than or equal to two, provided that at least one of said other metal oxides has an oxygen/metal atom ratio of less than two. Typical 7 oxides useful as solutes in these uranium dioxide-based solid solution electrode materials are scandia (Sc20 3 and 9 yttria (Y 2 0 3 other oxides which dissolve in uranium dioxide, such as calcia (CaO) and the rare earth oxides, 11 are also suitable, 13 It is an object of the present invention to provide means whereby the performance of oxygen sensors especially at lower temperatures can be enhanced.

17 A further object is to provide improvements in oxygen probe assemblies which have the effect of reducing 19 the importance of the electrode material, enabling sensors with conventional electrodes to give quite good behaviour 21 at low temperatures, The improvements also seek to provide better low temperature performance with assemblies 23 incorporating the advanced electrode materials described in our earlier application (WO 84/03149).

d The present invention arises from our observation 27 that emf errors in oxygen sensors operating at 300-5000C result predominantly from the presence of low levels of 29 combustible species in the gases exposed to the sensor.

It is well known that carbon monoxide, which displaces 31 oxygen from adsorption sites on the electrode at such temperato;es, is a source of emf error, We have also 33 found that other species such as methane and propane have a qualitatively similar effect, Low levels of unburnt

L.

I ~1 WO 89/09398 PCT/AU89/00113 -7- 1 combustibles are common in boiler flue gases, one of the important applications for oxygen sensors. We have also 3 found that most sources of reference air for the sensors also contain combustibles which cause errors at the reference electrode.

7 To prevent these combustibles from influencing the sensor emf, it is necessary to ensure that they are 9 removed before they reach the electrodes on the sensor.

If the sensor is operated at high temperatures, e.g. 700°C 11 or above, such removal occurs by reaction of the combustibles with oxygen in the reference air or with the 13 oxygen which is normally in excess in the boiler flue gases. When the sensor is operated at lower temperatures the oxidation process does not bring about complete removal of combustibles in the case of conventional probe 17 assemblies, 19 One approach to this problem is to oxidise the combustibles over a suitable catalyst external to the 21 oxygen probe and we have found that errors in emf are reduced by passing the gases over a heated platinum 23 catalyst before they reach the inner and outer electrodes of the sensor.

Such an arrangement, however, would be generally 27 unsuitable for commercial operations and accordingly we have developed a modified form of oxygen probe which a 29 enables a suitable catalyst to be incorporated therein.

31 Thus, in accordance with one aspect of the present invention, there is provided an oxygen probe assembly 33 comprising an oxygen sensor and means for directing a test gas and reference gas to the appropriate elewtrodes of the I gas WO 89/09398 P /A<O1 13' WO: 89/09398 PCT/AAU89/001 13 -8- 1 sensor, said assembly having associated therewith catalyst means arranged so that the test gas, the reference gas, or 3 both gases are separately contacted with the catalyst means before approaching the working surface(s) of the electrodes, whereby combustibles in the gases are oxidized.

7 Our investigations have also shown that the use of a catalyst may not be essential if the reference and/or test 9 gases are held at the probe temperature for sufficient time to allow the gases to attain thermodynamic 11 equilibrium at that temperature. This can be achieved by allowing the gases to diffuse to the electrode along a 13 path of sufficient length for equilibrium to be attained.

Such a path may take the form of a narrow labyrinth or annulus through which the gases must diffuse before reaching the electrode. We have termed such a labyrinth 17 or annulus a "diffusion space", 19 Preferably the probe assembly includes both the catalyst and the diffusion space. Preferably also, the 21 catalyst material is coated on at least part of the wall(s) defining the diffusion space(s), 23 The above and other aspects of the invention are further described and illustrated by the following description, which relates to one preferred (and 27 non-limiting) embodiment of the invention. Reference will be made to the accompanying drawings, in which: 29 Figure 1 is a diagrammatic cross-section of a probe 31 constructed in accordance with the invention, 33 Figure 2 is a modified construction of the probe of Figure 1; WO 89/09398 PCT/AU89/00113 -9- 1 Figure 3 is a graph showing the effects of 3 combustible contaminants on sensor emf; and Figure 4 is a graph showing how such emf errors are eliminated by passing the combustible-containing 7 gases over an oxidation catalyst.

9 In the modified oxygen probe assembly of Figure 1, the oxygen sensor comprises a solid electrolyte disk (1A) 11 which is mounted at the end of a ceramic tube 5 (the "sensor tube"). The electrolyte disk IA has porous 13 electrode 1 on its inner surface. Catalytic oxidation of combustibles is achieved within the assembly. Instead of impingeing directly on the inner electrode 1, reference air is introduced via an inner tube 2 and emerges fromr one 17 or more apertures 3 located typically 1-2cm from the electrode 1. Impurities in the air can only affect the 19 sensor emf by diffusing from aperture(s) 3 to inner electrode 1 through the annular space 4 between tubes 2 21 and 5. In this region, the outer wall of tube 2 and, optionally, the inner wall of sensor tube 5 have a coating 23 6 of a suitable catalyst e.g. porous platinum, Combustibles in the air are removed by oxidation at the catalyst surface before they diffuse to the electrode 1.

27 The same principle is used to clean the test gases before they reach the outer electrode 7, which is in 29 contact with the electrolyte disc lA, The gases approach the electrode through one or more holes 8 in metal sheath 31 9 which protects the sensor 5 and provides the electrical connection to the outer electrode 7, which is supported on 33 the base 12 of sheath 9. To reach the electrode they must then diffuse along the Onnular space 10 between sensor 897039g--'-- PCT/AU89/00113 1 tube 5 and a sleeve 11 which forms part of the base 12 of the sheath 9. The outer wall of sensor tube 5 and, 3 optionally, the inner wall of sleeve 11 are coated with catalyst 13, for example porous platinum. Combustibles in the external gas are removed by oxidation at the catalyst surface as they diffuse towards outer electrode 7 along 7 annular space 10. The best low temperature behaviour is obtained if catalyst layers 6 and 13 are both present; 9 however, the presence of diffusion spaces 4 and 10 without catalyst also gives significant improvements in 11 performance.

13 Typical dimensions for the elements described above are as follows; Inner tube 3.5-4 mm o/d 17 Sensor tube 8 mm o/d; 5 mm i/d Sleeve (11) 9-10 mm i/d; length 25 mm 19 Inner Annulus 1-1.5 mm wide; Outer Annulus (10) 1-2 mm wide; depth 25 mm.

2)1 These dimensions are given by way of example only,, 23 In the probe of Figure 2, the integral sleeve 11 is replaced by a separately formed and attached sleeve 11A and the catalyst 13A is coated on the inner wall of the 27 sleeve.

29 The following examples further illustrate the invention.

WO 89/09398 PCT/AU89/00113 -11- Example 1 The following tests were carried out to demonstrate that combustible contaminants cause errors in the signals developed by soci 1 electrolyte oxygen sensors at low temperatures, and that these errors can he eliminated by catalytic oxidation of the contaminants.

A ceramic oxygen sensor comprising a zirconia-scandia-alumina disc welded to the end of an alumina tube, carrying porous electrodes which were a mixture of platinum metal and a solid solution of urania and scandia (Sce23), was held in a furnace with its outer electrode exposed to flowing oxygen and its inner electrode exposed to flowing air. A system of flow meters and flow control valves was used to introduce controlled quantities of various gases (carbon monoxide, methane, natural gas and liquified petroleum gas) into the air stream.

Two experiments were performed with the sensor temperature at 427QC.

a) The concentration of comhuStthle* contaminant was varied and the effect on the sensor signal was recorded.

h) For a fixed level of each combustible contaminant, the effect was observed of passing the air/comhustible mixture over an oxidation catalyst, whose temperature was regulated, before sending It to to the sensor.* All the combustible species influenced the emf developed hy the sensor (Figure The changes in signal, which depended on the nature of the contaminant and the concentration introdued, were considerahly greater than would he expected simply on the basis of their influence on the residual oxygen concentration in the air stream, assuming complete combustion. The results suggest that the WO 89/09398 PCT/AU89/00113 12combustit-e molecules successfully displaced oxygen from adsorption sites on the sensor electrode, creating a surface condition typical of that expected. for low oxygen partial pressure, conditions.

These emf errors were eliminated by passing the comhustihle/air mixture over a heated oxidation catalyst (in this case, an automotive exhaust catalyst comprising platinuri-grqup metals dispersed on a ceramic substrate) before it reached the sensor (Figure The catalyst temperature required to nuecmlt elimnaio o teerror depended on the notrtre of the comhust hle con taminant, being lowest r4. 10000C for CO and higher (400-50 0

'C)

for, propane and natural gas.. The mechanism of, removal is presumed to inov ,cmution of the contaminants to form miolecular species (C0 2 H' 0) which do not influence the emf developed by the seris0r.

Example The following tests were performed to demonstrate that, it is possihle to greatly reduce the inflioence of combustihie c.ontaminants or, the emf of an oxygen sensor operated at low temperatures by providing a stagnant, diffusion zone and an in~ternal oxidation catalyst reqion in the, manner ouItl ined in Figure 1.

Two, sensors of the type. described in Example 1 were used in these experiments. Roth carried electrode of porous platinum (Platinum Paste NO6V Engelhard Industries Pty, Ltdh) for sensor N'o.1 the electrodes Were haked on at 60O"C* and for sensor No,? the electrodes were fired at 1Qono~c. The hfohor firing. tomperature yields An electrode with, coarser WO 89/09398 PCT/AU89/00113 13microstructure and lower electrochemical activity.

The sensors were held simultaneously in a furnace with their outer electrodes exposed to flowing air. A cylindrical plug of automotive oxidation catalyst was located in the hot zone of the furnace to remove combustible impurities from the air. Contact to the outer electrodes was made using platinum wires.

Two different internal contact arrangements were alternated between the two sensors. The first (contact A) comprised a twin-bore alumina tube with a platinum wire in one bore, the wire ending in a flat spiral which made contact to the inner electrode of the sensor. The two bores were also used to carry a test gas which impinged directly on the inner elestrode. The second contact differed from contact A in two important respects. Openings were cut intersecting each bore about from the contact end of the alumina tube, so that the test gas emerged via the openings and did not impinge directly on the inner electrode, Additionally, the wall of the twin-bore tube from the gas exit openings to the contact end was coated with porous platinum (Platinum Paste No.6(Oi, rngelhard Industries Pty. Ltd.) baked on at 600QC, The platinum coating served as an oxidation catalyst to remove combustihle 5 contaminants from the test gas as it diffused towards te inner electrode from the gas exit openings.

Four different test gases were used inside the sensors, i.e.

oxygen, air, 1.5% oxygen in nitrogen and 0.36% or 0.114% oxygen in nitrogen. Tests were carried out between 300" and 6001 0 Every posstble cmhfrnatfon of temperature, sensor, internal contact and test gas was investigated. Each test condition was monitored continuously and held for 24 hrs. to ensure that equilbhrium had been attained.

WO SWO 89/09398 PCI i AU 91IUU 113 14- Results are presented in Tahle 1, in which the observed sensor emfs are compared against the theoretical values calculated using the Nernst equation. In general at 40n"C and below contact R gave errors which were considerably less than contact A. Except for the results with sensor 1 and 0.365 0 in N 2 which suggest that the gas flow rate in that experiment was not high enough to prevent contamination of the test gas by atmospheric oxygen, contact P gave emfs accurate to ImV down to 400 0 C and in most cases down to 3500C. On the other hand contact A began to show significant errors at 4nO°C, and very large errors indeed at 350* and 300°.C The signs of the errors (inner electrode too negative) were consistent with the effect of combustible contaminants reported in Example 1. The tendency for the error to become worse with a decrease in the oxygen concentration of the test gas is as expected if the error is related to the presence of an oxidisable contaminant, Thus these results are best explained by trace levels of combustibles in the signals of both sensors at low temperatures. The experiment shows that the provision of an internal catalyst and the modification in gas path geometry incorporated in contact R suhstantially reduced the errors and lowered the minimum operating temperature of each sensor.

Table 1- Accuracy of' Sensor Emfs 0 00 ~'0

C

"0 ~0 00 Temperature Contact A Contact IB (00C Sensor 102 in gas t(ohs)-F(thor)(Wi) Sensor %O in gas FCnhs.) -F(theor) (W) 604 509 402 355 29Q 604 507 400 350 600 504 404 10a 100 100A 100 1N0 20.95 20.95 2m,95 20.95 '0.1 -1.1 05.2 -32.0, -0.1I 20.9q5 20.95 20.95 20.95 '20.1q5 ion 100 in 100 100 100 i00 100 0.2 0.2 3. 1 0.3I 0. 1 -0.3 -0.4 1. -3.6 Temperatuire Contact A Contact R 01 0 0) Sensor %0 in gas E(obs)-F(theor)(mV) Sensor %02 in gas E (oh s.)E(t heo r) (0n) 0 348 2 20.95 -16.4 1100.

297 ?09 -59J)0 1 100 A.3 601 2 100.0 1 0.5 513 2 100 120.195 0.4 401 2 100 -1.912.5-.

349 2 100 -R.7 120. 95 -0.3 3 10 2 100 -2 4 9 12 n .Q 5 3 604 2 1.50 1.9 .3 508a 2 1.50 n, 0.6 10.36 4.3 403 21.50 1 0.36 3.5 r 7 0 3 6 0.

-17 00 00 Temperature Contact A Contact R tac Sensor %02 in gas E(ohs)-E(tbeor) (m) Sensor %2in gas E(ohs.)-E(theor) (mV) 347 301 601 507 403 353 Z91, 608 510 408 350 1.50 1.501 0.36 0.36 0.365 0.36 114 0).114 0.114 0.114 -72. A -117.4 0.7 0 .8 -15.6 -66.4 -138.6 0.2 0.3 -12.7 -97.2 0.36 0.36 1.50 1 .50 1.50 1.50 1.50 1.50 1.50 1 .50 1.50 q -25.1 0.7 0.6 0).2 1 -36.6 0.1 '0.1 0.? 0.0 18- 00 Temperature Contact A Contact R \0 0 0) Sensor %02 in, gas I(obs)-E(theor)(mV) Sensor %02 in gas F(obs.)-E(theor)(mV) 1 0.114 -165.0 2 1.50 -1.6 607 1 1.50 1.1 2 0.114 509 1 1.50 1L1 2 0.114 409 1 1.50 -2.8 2 n.114 0.0 338 1 1.50 -68.0 2 0.114 00 301 1 1.50 -121.1 2 0.114 Theoretical values of E caljuiated, using the equation: E(theorAmV) 4.960xlfF- 2 (T+272.2)1ogln (x120.95) where x is the 7 02 in. the test gas T is the experimental temperature 00

L-

WO 89/09398 PCT/AU89/00113 19- Example 3 Tests were carried out on a complete oxygen probe assembly similar to that shown in Figure 1. The sensor body and sensor electrodes were as described in Example 1, and the catalytic surfaces incorporated in the assembly comprised coatings of platinum paste (Nr 608?, Engelhard Industries Pty.Ltd.) applied to the outer wall of the sensor tube and the outer surface of the internal multibore contact tube,respectively,and baked on at 600 C.

The probe was mounted in a laboratory furnace and held at temperatures between 300 and 500°C. A gas mixing pump was used to supply various oxygen/nitrogen mixtures to the outer electrode, and also to add to the test gas mixtures 1 percent of a 95 volutme percent COi is volume percent CO gas mixture, thus introducing 495 ppm of CO into the gas stream. Air was supplied to the inner electrode via the multihore contact tube. The probe temperature and sensor emf were recorded continuously and each test condition was maintained for 24 hrs. to ensure that equilihrium had been reached, The purpose of the experiment was to establish the absolute accuracy of the probe over a range of oxygen concentrations (2 to percent by volume) typical of air-excess combustion conditions, and to test whether that accuracy was affected by the presence of a typical level of unhurnt combustihle (here, CO). After these tests the probe was heated to 900C in air and held for 4 hrs, then returned to the range 300° to 500 0 C and retested. The purpose of this test was to evaluate whether the accuracy of the probe was affected by exposure to a temperature well above its anticipated temperature of application.

p. WO89/09398 PCT/AU89/00113 Results are given in Table 2. The small negative emf error at 400°C and above (1 to 2mV) is almost certainly due to a slight temperature gradient across the sensor in the small laboratory furnace. This constitutes a "zero error" which may he suhtracted from all the results. With this correction, the observed emfs are accurate to 1 mV down to 350 C for oxygen concentrations from 2% to The accuracy was not influenced by the presence of 495 ppm of CO, or hy heating the probe at 900°C for 4 hrs. This example shows that a probe made according to the invention is capable of accurate performance down to relatively low temperatures, is tolerant of the presence of low levels of combustibles, and is able to withstand temperatures well above its normal operating temperature without degradation.

1 5 1 I I WO 89/09398 PCT/AU89/00113 Table 2. Accuracy of Probe Emfs.

Probe Condition Temperature %02 ppm CO EVobs.)-E(theor)(mV) 0 c) As assembled 476 20.95 0n1.

I,3 14 20.95 0 U354 Z0.95 0 +0.3 304 20,95 0 502 100, 0 1.

it399 100 0 +0.3 it 354 100 0 it 304 100 0-4 .6 2, 0 95 0 II503 ?.074 495 "1,4 400 2,095 0 -0,7 3911 2,n74 4 45 -0,7 It350 P.095 0 0.0 350 2,0n74 4951, 099 ?.095 G03.

U300 2.074 4q5 -3.3- U 501 49 0 i.4 $02 4,140 495 it398 4,19 0 If3R4,140 49$ it34A 4.1q 0-.

WO 89/09398 PCT/AU89/001 13 22-- 4, Probe Condition Temperature %n Ppm CO E(ohs )-F(theor)(mV) 0 r) As assembled 349 4.149 4.95 ?99 4.19 n) +1.9 984.149 495 +1.9 502 10.475 502 10.371 495 39q 10,475 fl0, 400 10.371 495 s 350 10.475 0 350 10.3171 495 -015 9Q1n.475 0-4.7 ?99q 10.3,71 49$ 3.

0 C 4 hrs, 400 20.95 0 351 2,9 302.095 q F450 .074 49,5 -0.9 399 2.014 495 351. ?.095 0 'I351 P.074 495 -Io7 II301 2.n7440 3 4.09 0 NYO 89/09398 PCT/ALJ89/O01 13 Probe Condition Temperature, %02 ppm CO E(obs.)-E(theor)(nV)

(OC)

900%C 4 hrs. 502 4.149 495 -0.q 400 4.19 a011 401 4,149 49 -1.2 351, 4.19 0 -11 351 4.14q 495 -1.1 301 4,19 0 -111 302 4.149 495 -1.6 Theoretical values of Ecalculated as described, below Table 1.

Claims (4)

1. An oxygen probe assembly comprising a solid electrolyte oxygen sensor and means for conveying a test gas and a reference gas to the appropriate electrodes of the sensor, characterised in that the means for conveying the test and gas 4 the reference gas. -m h-mf-sa t e to the appropriate electrode(s) include a pathway in the form of a labyrinth or annulus which extends to the electrode and through which the gas must diffuse before reaching the electrode, and wherein at least part of the said labyrinth or annulus consists of, comprises or is coated with an oxidation catalyst, whereby combustible substances in the gas(es) passing through said pathway(s) are removed by catalytic oxidation over said catalyst before the gas(es) reach the appropriate electrode(s), 15

2, An oxygen probe assembly as claimed In Claim 1, wherein the oxygen sensor comprises a solid electrolyte disk or pellet which is mounted at the end of a tube (the "sensor tube"), the electrolyte disk or pellt having or being in electrical contact with an electrode on each of its inner and outer faces; characterised in that at least the electrode end portion of the sensor tube is S* 20 surrounded by an outer tube which, together with the sensor tube, defines a first annular space surrounding the sensor, whereby one of test or reference gases can enter the outer tube and diffuse through the first annular space to the outer electrode surfacethe ouer su n the sensor tube houses an inner tube which, together with the sensor tube, defines a second annular space within the sensor tube; said inner tube being provided with gas Inlet and outlet means whereby the other of said gases can enter the second annular space and diffuse to the inner electrode surface, and at least one wall of the tubes which define the first and second annular space wholly or partly consists of, comprises or Is coated with the oxidation catalyst.

3. An oxygen probe assembly as claimed in Claim 1, wherein the oxygen sensor comprises a solid electrolyte disk or pellet which is mounted at the end w 1 U,- 25 of a u.abe (tt~e "sensor ub&e 9 the electrolyte disk or pellet having or being in electrical contact with an electrode on each of its inner and outer faces; characterised in that at least the electrode end Portion of the sensor tube is surrounded by, an outer tube which, together with the sensor tube, defines a first annular space surrounding the sensor, whereby the test gas can enter the outer tube and diffuse through the first annyir space to the outer electrode surface; and the sensor tube houses an inner tube which, together with the sensor tube, defines a. second annular space within t-he sensor tube; said inner if tube being provided with gas inlet and outlet means wiereby the reference gas cap.l enter the second annular space and difue to the Inner electrode surface, and at least one wall of the, tubes which, define the first and second annular space wholly or partly co'nsists of, comprises or is coatedwtthoiain a 0 acatalys t. 5

4, An oxygen probe assembly as claimed in. Claim 2 or Claim 3, 0 chracrersed inthat te senor Is housed within a tubular shethhvga closed end which carries an. inwardly-diroct-A cup or sleeve, the walls of which comprise the soid Outer tube, the sheath being, provided with one or more gas An oxygen probe. assembly as clalnied In Claim, 4, characterised In that 0 ~the cup or sleeve formfs an Intogral part of' the, closed end of the set 6, An oxygen: probl, iembly as Claimed In Claim 4, characterlscd In that 25thecupor leee i fomed separate.,y and attached to the closed ed or the 7, An oxyge-F probe assembly substantially ashrlneo described with reference to the. accompanying dlrawings. A T