GB2037432A - Oxygen sensor - Google Patents

Abstract

In order to obviate deterioration of solid electrolyte exhaust gas oxygen sensors due to poisoning of the electrodes by materials in the exhaust gas the electrodes 4 are coated by a metal oxide 5 having a perovskite crystal structure and having the formula A(B1-yB<1>Y)O3 wherein A represents one or more metal cations having ionic radii between 0.8 and 1.65 ANGSTROM , B represents cations of one or more platinum group metals, B<1> represents cations of one or more non-platinum metals having ionic radii between 0.4 and 1.4 ANGSTROM and y has a value of 0 to 0.99 <IMAGE>

Description

SPECIFICATION Oxygen sensor This invention relates to a sensor for determining changes in the oxygen concentration in hot gases, particularly exhaust gases from the combustion of hydrocarbons.

The need to minimize the discharge of noxious gases to the atmosphere has brought about the use of oxygen concentration sensors in exhaust gas systems. Such a device is a special type of voltaic cell which is inserted in the exhaust gas stream where the oxygen concentration of the stream is measured in terms of the electrical voltage generated by the sensor. By connecting the sensor to an electrical control system, it is possible to adjust the air/fuel ratio in response to changes in oxygen concentration in the exhaust gas, thereby controlling combustion and the content of the exhaust gases emitted to the atmosphere.Oxygen concentration sensors are now being used in some gasoline internal combustion engines, and it is expected that this use will increase as the automobile industry converts to three-way emission control catalysts which require sensitive adjustment of the air/fuel ratio. A discussion of the use of oxygen sensors in automobile engines is found in "Automotive Engineering", Vol. 85, No. 2 pp.

45-50, February 1 977.

Unlike conventional voltaic cells, such as storage batteries, which employ a liquid electrolyte, oxygen sensors use a solid electrolyte. The solid electrolyte in an oxygen sensor is usually a ceramic tube, composed of zirconia, closed only at one end, thus leaving the inside of the tube accessible to the atmosphere. The outside and the inside of the tube are coated with porous platinum and serve as electrodes separated by the wall of the tube, i.e., the solid electrolyte. The closed end of the tube can be sealed into an exhaust gas system while the open end remains outside where it is exposed to the ambient atmosphere. The difference in oxygen concentration in the two atmospheres generates a voltage across the electrodes and an electrical signal can be sent to the electronic control system which maintains the proper air/fuel ratio for clean combustion.

Since oxygen sensors must be maintained in contact with the hot exhaust gas stream, th y are subject to conditions where deterioration canbe severe. A part of the deterioration can be attributed to attrition by the stream of hot gases. Another possible cause is poisoning of the electrodes and the solid electrolyte by combustion by-products. For example, when an oxygen sensor is used in an automobile engine, there can be poisoning from sulfur, halogen, phosphorous or lead compounds present in the fuel. This is even true with lead-free gasoline which can contain sulfur from the original petroleum source, as well as lead and phosphorus as a result of contamination from pipelines, storage tanks, transport tanks and the like, which have previously handled leaded gasoline.Present U.S. regulations permit up to 3.05 gm of lead and up to 0.005 gm of phosphorus per gallon in unleaded gasoline.

Several improvements have been suggested in the art of oxygen sensors. U.S. Patent 3,935,089 discloses coating the electrode that is exposed to the hot gases with a porous inorganic material such as aluminum oxide impregnated with a catalyst material, such as a platinum or palladium, to oxidize the oxidizable components of the gas stream. Also, there is U.S. Patent 4,021,326 which discloses coating the outside electrode with a layer of magnesium spinel, kaolin or talc. This coating makes it possible to have larger pores or fissures in the electrode without decreasing the sensitivity of the device. In another embodiment of this patent, the magnesium spinel is further coated with a metallic layer of nickel, or a chrome-nickel alloy or silver or gold which acts as a getter to tie up poisons that might affect the operation of the sensor.

The present invention pertains to oxygen sensors in which the surfaces exposed to the exhaust gases are coated with an adherent, porous coating of a metal oxide with a perovskite crystal structure and having the chemical formula A(BjvB'v)O3 wherein A represents one or more metal cations having ionic radii between about 0.8 and 1.65 Angstrom, B represents cations of one or more platinum group metals, B' represents cations of one or more non-platinum metals having ionic radii between about 0.4 and 1.4 Angstrom, andy has a value of from 0 to about 0.99. It has been found that sensors having this perovskite coating are more durable in deleterious atmospheres, such as the exhaust from leaded gasoline where sensors have a much shorter useful life than when used to monitor the exhaust of unleaded fuel.

In the perovskite used in the present invention, preferred compositions are those in which the A site of the chemical formula is occupied by a combination of lanthanum and strontium cations: the B site is occupied by either platinum cations or a combination of platinum and ruthenium cations; the B' site is occupied by cobalt cations. Also, it is preferred to have y in the range of about 0.80 to 0.99, thus reducing the amount of the more expensive platinum group metal used. Specific compositions containing these metals which are preferred are La0.6Sr0Co0.8Pt0.2O3 LaO6sro4coo94ptoo3Ruoeo3 03.

Description of the Drawing Solid ion-conductive electrolyte 1 is a tapered ceramic member usually yttrium oxide-stabilized zirconium oxide with hollow core 2 open to the atmosphere at one end. The inner and outer surfaces of the solid electrolyte are coated with a porous layer 3 and 4 of platinum which serve as electrodes.

Electrical leads (not shown) are connected to the layers. The tapered end of the sensor is inserted through the wall of the conduit carrying the gas stream while the open-end remains outside the conduit in contact with the ambient atmosphere or a reference gas, thus creating an electrical potential across electrodes 3 and 4 when the sensor reaches operating temperature. This is usually about 3500C or higher. In the present invention, a coating 5 of a metal oxide containing a platinum group metal and having a perovskite crystal structure constitutes the outside layer of the sensor that is exposed to the exhaust gas atmosphere. In a more preferred embodiment, a magnesium spinel coating is applied before applying the metal oxide of this invention or the two coatings may be applied as a mixture.

The oxygen sensors which are coated with a perovskite metal oxide according to the present invention are available from a number of manufacturers. In these devices, the solid ion-conductive electrolyte is usually zirconium oxide stabilized with yttrium oxide or calcium oxide. However, it is also possible to use cerium oxide or thorium oxide stabilized with calcium oxide. The metal coating on the surface of the solid electrolyte must be sufficiently porous to allow exhaust gases to permeate to the face of the electrolyte. In addition to platinum, such metals as palladium or iridium may be used. Also, the electrode layer can comprise alloys of precious metals, such as platinum with aluminum, chromium or cobalt. Metal oxide systems such as a lanthanum-cobalt oxide are also known for this purpose.

Although the drawing shows the electrode coating covering substantially the entire inner and outer surface of the solid electrolyte, there are sensors on the market on which the electrode coating covers only a portion of the electrolyte, thus economizing on the use of platinum or other precious metal. These sensors can also be coated according to this invention.

In a preferred embodiment of this invention, the sensor has a porous, gas permeable coating of magnesium spinel interposed between the electrode layer and the perovskite metal oxide layer. Sensors having spinel coatings are disclosed in U.S. Patent 4,021,326, and the perovskite coatings of this invention are applied over the spinel to form the outside layer exposed to the atmosphere of exhaust gases.

One of the essential characteristics of the metal oxides used in the present invention is that they possess a perovskite crystal structure. The chemistry of perovskites and the methods for their preparation are described in the literature. For example, discussion of perovskites as well as preparative methods can be found in "Progress of Inorganic Chemistry", F. A. Cotton, Editor, Interscience, New York, 1 959, Vol. 1, pp. 496-520 and in "Structure, Properties, and Preparation of Perovskite-Type Compounds" by F. A. Galasso, Pergamon Press, 1 969.

The structure which is known to those skilled in the art as the ideal perovskite structure contains cations of appropriate relative size and coordination properties and has a cubic crystalline form wherein the corners of the unit cube are occupied by the larger A site cations, each coordinated with twelve oxygen atoms, and the center of the cube is occupied by the smaller B site cation, each coordinated with six oxygen atoms and the faces of the cube are occupied by oxygen atoms. Thus, the A site cations are normally somewhat larger than the B site cations. In the present compositions, the A site cations will have ionic radii from about 0.8 to 1.65 Angstroms while the B site cations will have ionic radii from about 0.4 to 1.4 Angstroms.Distortions and variations of the above-described cubic crystal structure are known among materials commonly considered to be perovskite or perovskite-like. The term "perovskite" as used herein includes these materials.

The particular metals present in the A site in the perovskite compositions used in this invention are less critical than the B site metals, the most important property of the A site metals being the radii of their cations. Ionic radii have been tabulated by Shannon and Prewitt Acta. Cryst: 326 1046 (1970); 825925 (1969), and their importance in perovskite crystal structures has been discussed by many authors, e.g., by Krebs in "Fundamentals of Inorganic Crystal Chemistry", McGraw Hill, London (1968).

The A site metals can be from the periodic table groups 1 A, 1 B, 2A, 2B, 3B, lanthanide rare earth metals (atomic number 57 through 71) and from the actinide rare earth metals (atomic number 89 through 104) as well as metals from aroups 4A and 5A.

A site metals having a valence of one which can be used are metals from groups 1 A and 1 B.

Preferably, they are cesium, rubidium, potassium, sodium, or silver and, more preferably, potassium or sodium. As for A site metals having a valence of two, those from groups 2A, 2B, 4A, and the lanthanide rare earth (RE) metals which form oxides of the type RE(II)O are useful. Preferably, they are barium, strontium, calcium, or lead and, more preferably, strontium or barium.

Likewise, the Type A metals having valence of three can be used. They are from groups 3B and 5A and the lanthanide and actinide rare earth metals. Preferably, they are lanthanum or a mixture of the lanthanide rare earth metals (e.g., a mixture containing about one-half cerium, one-third lanthanum, one-sixth neodymium, and smaller amounts of the remaining metals of atomic number 58 through 71 or a similar mixture from which a major part of the cerium has been removed).

The A site of the perovskite compositions useful in the present invention can be occupied by one or more cations of the above-mentioned A site metals. The total number of A site cations is essentially equal to the total number of B site cations although A- and B-site vacancies producing a slight imbalance are well known in the art and such compositions are intended to be included in the description of the invention. It is also necessary that the combined charge of the cations is substantially equal to the charge on the oxygen atoms.

The perovskite compositions useful in the present invention are particularly characterized by the presence of at least one platinum group metal in the B site of the perovskite crystal structure. By a platinum group metal is meant a metal in the second and third long periods in group 8 of the periodic table which include ruthenium, osmium, rhodium, iridium, palladium and platinum. From about 1 to 100 weight % of the B site can be occupied by one or more platinum group metals. However, since the platinum group metals are expensive, a non-platinum group metal can be used in conjunction with the platinum group metal. It is possible to use as little as about 1% of the platinum group metal and to have up to about 99 weight % of the B sites occupied by a non-platinum group metal.In order to make a distinction between the platinum group metal and non-platinum group metal at the 3 site of the crystal structure, the symbol B' is used in the empirical formula to denote non-platinum group metals which are present in place of the platinum group metal. Usually from about 1 to 50 weight %, and more preferably, from about 1 to 20 weight 50 of the B sites are occupied by a platinum group metal, the remainder being a non-platinum group metal represented by the symbol B'. Stated in terms of the empirical formula A(B1vB'v)O3, the lower limit for y is preferably no less than about 0.50 and most preferably, no less than about 0.80.While benefit is realized by the inclusion of smaller amounts of the platinum group metal, the additonal benefit achieved when more than 20 weight % of the sites are occupied by the platinum group metals may not justify the added expenditure.

The preferred platinum group metal is platinum which can be used alone or in combination with one or more of the other platinum group metals such as, for example, the combination of platinum with ruthenium, the combination of platinum with palladium, the combination of platinum with rhodium, the combination of platinum with palladium and ruthenium, the combination of platinum with palladium and rhodium and the like. A preferred combination is the use of platinum with ruthenium.

In type B sites in the perovskite crystal structures, palladium is typically diva lent, rhodium is typically trivalent, ruthenium, iridium and platinum are typically tetravalent, and osmium may be present in valences of either four, five, six or seven.

As previously mentioned, non-platinum group metals can act as a substitute for some of the platinum group metals in the perovskite crystal structure. The non-platinum group metals should have an ionic radii between 0.4 and 1.4 Angstroms, and they can each be present in an amount and with a valence consistent with the presence of the platinum group metal ion and with the perovskite crystal structure. These B' metal ions can thus have valencies of one to seven and can be from metals of the periodic table groups 1 A, 1 B, 2A, 2B, 3A, 3B, 4A, 4B, 5A, 5B, 6B, 7B and 8 or from the lanthanide and actinide rare earth metals.

The B' metals having a valence of one can be from groups 1 A and 1 B. Preferably, they are sodium, silver, or copper. Those having a valence of two can be from groups 1 B, 2A, 2B, 3B, 6B, 7B, and 8.

Preferably, they are magnesium, calcium, strontium, chromium, manganese, iron, cobalt, nickel, or copper. The B' metals having a valence of three can be from groups 3A,3B,4B,5A,5B, 6B, 7B, and 8 and the lanthanide and actinide rare earth metals. Preferably, they are lanthanum, a lanthanide are earth metal, aluminum, titanium, vanadium, chromium, manganese, iron, cobalt, or nickel. The B' metals having a valence of four can be from groups 4A, 4B, 5B, 6B, 7B, and 8. Preferably, they are titanium, vanadium, chromium, manganese, iron, cobalt, nickel, or rhenium. The B' metals having a valence of five can be from groups 5A, 5B, 6B and 7B. Preferably, they are selected from antimony, niobium, tantalum, vanadium and rhenium.The B' metals having a valence of six are preferably tungsten and molybdenum, and the non-platinum metal having a valence of seven is preferably rhenium.

Certain useful compositions of this invention contain B' metals having a single fixed valence, i.e., metals whose ions are commonly known to exist in several valences but are usually found in the perovskite crystal structures only in one particular valence state. Such compounds have a major proportion (e.g., at least about 50 weight % and preferably 75 weight % or more) of B' metals which are known in perovskite crystal structures primarily or only in one valence. The metals of this group are: Valence 1: lithium, sodium, silver; Valence 2: magnesium, calcium, strontium, barium, zinc, cadmium; Valence 3: aluminum, gallium, indium, thallium, lanthanum, yttrium, and neodymium; Valence 4: zirconium, hafnium, thorium, germanium, tin; Valence 5: antimony, tantalum; Valence 6: tungsten.

Preferably, the nonplatinum metals of this class are sodium, magnesium, calcium, strontium, aluminum, tin, or antimony. These relatively abundant metals can be present in the compounds of this embodiment in major proportions with relatively small reductions in the catalytic activity contributed to these compounds by other less readily available metals and therefore represent relatively inexpensive diluents in such compounds. More preferably, the compounds contain a valence three metal and especially aluminum as the principal nonplatinum metal. Aluminum is not only an inexpensive diluent but also imparts to perovskite crystal structures a high degree of thermal stability and durability in catalytic applications.

Another embodiment of this invention comprises compositions wherein a major proportion (e.g., at least 50 weight % and preferably more than 75 weight %) of the B' metals exhibit a variable valence, that is, are known in a first valence in one perovskite compound and in a second valence in a second perovskite compound. Such metals known in perovskite crystal structures in two valences differing in increments of one or two valence units are: Valences 1 and 2: copper; Valences 2 and 3: scandium, samarium, ytterbium; Valences 2 and 4: lead; Valences 2, 3, and 4: chromium, manganese, iron, cobalt, nickel, and cerium; Valences 3 and 4: titanium, praseodymium; Valences 3, 4, and 5: vanadium; Valences 3 and 5: bismuth, niobium; Valences 4 and 6: molybdenum; Valences 4, 5 and 6: rhenium and uranium.

The compositions of this embodiment can contain one and preferably contain two or more such variable-valence nonplatinum metals, particularly those transition metals which have atomic numbers between 22 and 29 inclusive, that is, titanium, vanadium, chromium, manganese, iron, cobalt, nickel and copper. Particularly preferred are manganese, iron, cobalt and nickel. These metals are readily available and are capable of existing in perovskite crystal structures in two or three valences differing by one valence unit increments. In one preferred embodiment of this invention, the nonplatinum Type B' metals are such metals each in a single valence.

Those compositions in which at least one B' metal is present in two valences constitute another preferred embodiment of this invention. Such metal oxides have increased activity as catalysts over similar compounds in which each of the component metals is present in only a single valence, possibly because of the enhanced electron mobility through t,heir crystal structures resulting from the presence of a variable-valence metal when at least 5 weight % of the B' sites are occupied by a variable-valence metal in a first valence and at least about 5 weight % of the B' sites are occupied by the same metal in a second valence. The valences preferably differ by one unit but will differ by two units with some metals, such as lead and niobium.

The present compositions which contain a single A site metal and at least one metal ion in the B' site which can be in perovskite crystal structures in two or more valences permits easy adjustment of the valence balance of the compound. The amounts of differing valence forms of a compound can be adjusted so that the total valence charge of the metals equals the total valence charge of the oxygen present. Examples of such balanced compounds that can be used in this invention include [La(lll)j [Ti(lV)02Cr(ll)O rCr(lll)O*Pd(11)01]03 tLa(lll)j (cutll)03co(111)04co(lV) osV(Iv)02pt(lv)o5]o3 [Sr(l I)] [Co (I aCo(lV)06Nb(V)O aRu(lV)02]03 Similarly, variable valence metals permit the formation of the perovskite crystal structure when deficiencies of up to about 25 weight % of a metal or oxygen might prevent the precise ABO3 stoichiometric ratio.

The useful compositions of this invention can be prepared by heating mixtures of metal oxides, hydroxides, metals, and/or metal salts for sufficient times at temperatures which permit spontaneous formation of the compounds. The mixture of materials which are heated are preferably finely subdivided and intimately mixed before heating and are thoroughly ground and mixed by any conventional techniques. Several times during the heating period, the mixture is cooled, reground and remixed since the compounds are in many instances formed by atomic diffusion, without melting of any of the starting or potential intermediate materials, and are subject to coating of unreacted particles by reaction products.The heating times and temperatures required for the formation of significant amounts of these catalytic compounds depend upon the particular compositions being formed, the required times usually being shorter at higher temperatures. Temperatures above about 8000C are usually suitable for the formation of these compounds but temperatures above about 9000C are usually preferred with firing times of hours to days with occasional intermediate grinding and mixing and temperatures of 1,000 to 1 ,5000C can be used.

In forming the compositions used in this invention, stoichiometric mixtures of starting materials are preferably heated in air or other oxygen-containing gas mixture.

The starting materials used in preparing the compositions used in this invention by anhydrous processes can be any salts which are converted to oxides by prolonged heating in oxidizing atmospheres at the temperatures at which these compositions are formed. For example, they can be carbonates, salts of carboxylic acids (e.g., acetates, oxalates, tartrates etc.), salts of the acids of sulfur (e.g., sulfides, sulfites, sulfates, etc.), halogen acid salts which are converted to oxides without volatilization (e.g., ruthenium chloride, strontium chlorate, barium perchlorate), salts of the acids of nitrogen (e.g., nitrates, nitrites, etc.). Preferably, they are carbonates, nitrates or sulfates.The presence of small amounts of the salts of other such acids in a mixture which is predominantly oxides or carbonates is usually not significantly deleterious since such salts are converted into oxides during heating to prepare these catalytic compositions.

As mentioned, the platinum group metal containing perovskite compositions useful in the present invention can be represented by the formula A[BI~yB',]03 wherein 3 represents one or more cations of the platinum group metals and y has a value of O to 0.99. Perovskite compositions containing only platinum group metal cations in B sites (i.e., y=O) are known and include inter alia, BaRu03, SrRuO3, CaRuO3, SrRu,,lr,,O,, Bairn3, SrlrO3, CaIrO3, InRhO3, La RhO3, PrRhO3, and the like. However, as also mentioned, preferred perovskite compositions contain from about 1 to 20% of platinum group metals as B site metals. Such perovskite compositions and their preparations are more fully described by Lauder in, U.S. Patents 3,897,367: 4,049,583; 4,110,258; and 4,110,254.

It has been previously mentioned that the electrode coating must be sufficiently porous to allow the exhaust gases to permeate to the surface of the solid electrolyte. This also is true for the perovskite metal oxide coating used in the present invention since the exhaust gases must first permeate this coating in order to ultimately reach the electrode-electrolyte interface. Another requirement for the coating of the present invention is that it should be an adherent coating that will not flake off the sensor when subjected to the atmosphere of the hot exhaust gases. The procedures described hereinafter for applying the coating will enable those skilled in the art to obtain adherent porous coatings on the sensor.In this connection, careful consideration should be given to the selection of particles of proper size and the use of a crystal growth inhibitor such as Y2O3, Thou, ZrO2, MgAI20 or Al203 in the range of about 10 weight % to 50 weight %, basd on the total weight of the coating. From a preferred standpoint, about 20 weight % of inhibitor is used, and preferred inhibitors are ThO2 and ZrO2. These inhibitors often impart better stability and durability to the perovskite coatings. The use of a crystal growth inhibitor is recommended when applying the coating by plasma spraying. Various known methods can be used to apply the coating.For example, the sensor can be dipped into or brushed or sprayed with, a slurry of perovskite composition of suitable subdivision such as those having particle size in the range of 40 to 200 micron diameter. The article thus coated is then sintered in the temperature range of from 8000C to 1 3000 C. The perovskite coating can also be applied to the oxygen sensor element by plasma spraying using perovskite powders. The use of the perovskite compositions having the particle size in the range of from about 40 to 200 microns will assure coatings having satisfactory porosity such that the diffusion of gases in the exhaust stream is not impeded. The perovskite coating thickness can be from about 1 to 500 microns, preferably 10 to 200 microns.When a crystal growth inhibitor is used, the mixture of perovskite and the growth inhibitor is usually viscous milled to insure substantially uniform particle size and proper mixing.

To prepare the coating composition, the perovskite composition together with a crystal growth inhibitor, if used, can be viscous milled in an alumina ceramic cylindrical container together with about an equal weight of alumina granules of about 20 to 60 mesh and about 4050% by weight of water.

The milling is usually continued for about 4 to 6 hours using an aluminum disc rotated at a speed of about 1 500-1 700 rpm such that a doughnut-shaped viscous mixture is obtained. Water may be added to replace that lost by evaporation. After viscous milling is completed, water is added to the mixture and the total mixture is passed through a milk filter to separate the perovskite slurry from the alumina grinding media.

To apply the perovskite by wash-coating, the oxygen sensor element is dipped to the desired level into the aqueous suspension prepared as described above and then dried under vacuum. Dipping and drying are repeated as desired but usually two dippings are sufficient. After final drying, the sensor element is slowly heated to about 8000C to 1 3000 C, preferably 9000C for about 16 hours. Typically, the thickness of the perovskite coating is in the range of from about 1 to 20 microns.

In the plasma spraying procedure, the aqueous suspension of the perovskite composition obtained from viscous milling is dried under vacuum in a shallow pan. The resulting dry cake is heated for 8 to 100 hours at 900-1 4000C in a platinum dish. Analyses of thus heated perovskite composition by X ray should indicate that the perovskite crystal structure is retained. The cake is then crushed and sieved.

Sieve fractions 80/324 (44-1 80 microns) are used for plasma spraying. Typically, perovskite composition powder is plasma sprayed onto an air-cooled, rotating oxygen sensor element using argon gas for the generation of the plasma jet at a power load of about 8 to 1 8 Kilowatts. The resulting coating is typically 25-200 microns thick, porous and glassy-looking.

The sensors of the present invention operate efficiently in engines using non-leaded fuel. However, the significance of this invention is most apparent in engines operated on leaded fuel where the exhaust gases have a tendency to adversely affect the performance of the sensor within a short period of time.

The following specific Examples are intended to illustrate the invention. The oxygen sensors used in each of the Examples were commercially available magnesium spinel coated sensors manufactured by the Robert Bosch Corporation (Mfgr's Part No. 0258 001 001). The same model was used in all of the examples.

The gasoline used in all of the examples was Indolene 98, a test gasoline which meets the specifications of the U.S. Environmental Protection Agency for gasoline used in exhaust and evaporative emission testing. These specifications are published in the Federal Register, Vol. 41, No. 177, page 38682. September 10, 1976.

The sensors were coated by both plasma spraying and wash-coating. The sensors of Examples 1, 2,4 and 5 were coated by plasma spray, and the sensor of Example 3 was wash-coated. The procedure used to coat the oxygen sensors by plasma spraying was as follows: A mixture of 100 g La0.0Sr0Co0.8Pt0.2O3 and 25 g ThO2 was viscous milled by placing the mixture in a 200 ml alumina ceramic cylindrical container together with 120 g of 20/60 mesh alumina granules and 180 ml deionized water. The disc used for milling the above mixture was made of aluminum and was rotated at 1 500--1700 rpm such that doughnut-shaped viscous mixture of the ingredients was obtained within the alumina container.After milling for 6 hours, 200 ml water was added and the mixture passed through a milk filter to separate the aqueous suspension of the perovskite composition from the alumina granules. The aqueous suspension of the perovskite composition was then dried under vacuum in a shallow pan. The resulting dry cake was heated in a platinum dish for 64 hours at about 1 3000C when ThO2 was present, and at about 1100C when ZrO2 was used. Analysis of the thus heated perovskite composition by X-ray indicated that the perovskite crystal structure was retained. The cake was then crushed and sieved. Sieve fractions 80-325 mesh (44-180 microns) were used for plasma spraying.

The perovskite composition powder was plasma sprayed onto an air-cooled, rotating oxygen sensor using argon gas for the generation of plasma jet at a power load of 14 kilowatts (400 A, 35 V). A similar procedure was used to coat the sensors that were tested for comparison purposes.

As previously mentioned, the sensor of Example 3 was wash-coated, and the procedure used was as follows: A perovskite composition, La06Sr0Co0g4Pt003Ru003O3, 33 g was mixed with 50 9 of 20/60 mesh alumina granules and 50 ml deionized water and viscous milled for 5 hours as described above.

After adding 50 ml of water, the mixture was passed through a milk filter to separate the aqueous suspension of the perovskite composition from the alumina granules. The oxygen sensor element was dipped into the aqueous suspension thus obtained and dried under vacuum. Dipping and drying under vacuum was repeated. The dried sensor element was then slowly heated to 9000C and kept at 9000C for 16 hours.

EXAMPLES 1--4 For the test, two sensors are mounted opposite to each other in an insulated stainless steel chamber bolted to the exhaust port of a Kohler Model K-9 1 single cylinder gasoline engine (8.86 cubic inch displacement, nominally 4HP) fitted with an electronic spark ignition system. The engine was operated at about 3,000 RPM at an A/F ratio of about 13.9 under which conditions the exhaust gas temperature was in the range of 6900 to 7500C.

The engine was operated on Indolene 98 containing a commercial antiknock compound at a loading of 2 g Pb/gallon.

The oxygen and carbon monoxide content (vol. percent) in the exhaust gases was determined chromatographically after condensing out most of the water via an ice-bath cooled trap and removing the entrained particulate matter by passage through a small pore filter. The date obtained was used to calculate the volume percent of excess CO and/or the volume percent of excess 02 in the exhaust gases according to the following equations: Percent Excess CO = Measured CO-2X Measured O2 Percent Excess 02= Measured 02-0.5X Measured CO Initially and after each period of approximately 100 hours of steady-state operation under the above-described conditions, the A/F ratio was varied stepwise until the exhaust gas showed about 3% excess carbon monoxide.During this stepwise change in the A/F ratio, the sensor voltage was measured with Keithly Digital Multimeter, Model 160B and recorded with a Varian Associates strip recorder, Model G-1 4. The voltage values were then plotted vs. A/F ratio to obtain the step function performance of the sensor.

Table I reports the hours that the sensors operated before failure under the above-described conditions.

TABLE I.

OXYGEN SENSORS LABORATORY SINGLE CYLINDER ENGINE EVALUATIONS Engine: Kohler Model K-91 Fuel: Indolene 98 plus Motor Mix Antiknock at a level of 2 g of Pb per gallon Hours Stabi- of lizer Opera 20 wt. tion % of Before Coating Fail Sensor Coating Applied Applied ure* Commer cial Used as purchased 300 Compari son A Cobalt metal None 300 Compari son B LaAl03 ThO2 300 Compari son C La0Sr02CoO3 ThO2 300 Ex. 1 La0.8Sr0.4Co0.94Pt0.03Ru0.03O3 ThO2 600 Ex. 2 La0.8Sr0.4C0.8Pt0.2O3 ZrO2 1,100 Ex. 3 La0.6sro4coo94ptoo3Ruo o3o3 None 1,500 Ex. 4 La0.6Sr0.4CO0.8Pt0.2O3 ThO2 300 A drop in the peak voltage output of the oxygen sensor from the initial range of about 900 to 1000 millivolts to about 500 to 600 millivolts was considered to be a failure.

The above results show that the commercial oxygen sensor lost its sensitivity in about 300 hours when the engine is operated with leaded fuel. The commercial sensor coated with metallic cobalt by plasma spraying also failed in about 300 hours (Comparison A). Coatings of perovskite compositions outside the scope of this invention (Comparisons B and C) also had no effect upon the durability of the oxygen sensor. Improved sensors of the invention having coatings of perovskites containing platinum group metal have increased durability of from about 300 to 1,200 hours over the commercial sensors or commercial sensors coated with compositions outside of the present invention.

EXAMPLE 5 The performance of the oxygen sensor was determined on 1977 Model 240 Series Volvo automobiles equipped to meet air pollution standards for the state of California. The engines used a fuel injection system and an oxygen sensor feed back system designed to hold the A/F ratio slightly above stoichiometric in oxygen content. The stoichiometric ratio is the theoretical oxygen-to-fuel ratio necessary to completely oxidize the fuel to carbon dioxide and water. The Volvo was equipped with a three-way catalyst system and a constant A/F ratio slightly above stoichiometric is essential for proper operation of such a catalyst system which simultaneously oxidizes hydrocarbons and carbon monoxide and reduces oxides of nitrogen.

The testing of the sensors under steady-state conditions was tarried out with the vehicle running on a chassis dynamometer. The vehicle was operated at each of the several speeds for approximately 5 minutes with the speed being decreased to zero (idle) in- increments of 5 miles per hour (mph). The selection of several speeds was based on the assumption that each speed represents a unique combination of engine speed, load, exhaust flow, exhaust temperature and exhaust gas composition and that proper sensor operation at each of the speeds would mean satisfactory operation at varying vehicle speeds and load conditions encountered under actual driving conditions. This steady-state test therefore measures the ability of the sensor to supply proper signals to the air-fuel control devices over a broad range of engine operation.

Five sensors were tested. Three of the sensors without the coating of this invention were used as controls. The other two were coated with a perovskite compostion, La0.6Sr0Co0.8Pt0.2O3, said coating being stabilized by inclusion of 20% ThO2 based on total weight of coating applied. The coating procedure was by plasma spraying as heretofore described.

All tests were carried out in 1977 Model 240 Series Volvo automobiles. Three different cars were used, and since the sensors of the present invention were durable for much longer periods than the control sensors, there was an opportunity to use them on more than one vehicle. Table II shows the composition of the leaded fuels that were used, the laboratory identification number for the test vehicle, and the number of miles that the sensor was operated in that particular vehicle.

TABLE II.

Exposure Fuel Used: Indolene 98 plus Sensor Miles Vehicle lead content as indicated Commer cial-l 1000 VO-i 0.5 g Pb/gal as Motor Mix' Commer cial-2 1000 VO-1 0.5 g Pb/gal as Motor Mix Commer cial-3 5000 VO-i 0.5 g Pb/gal as Motor Mix Inven tion-l 5000 VO-i 0.5 g Pb/gal as Motor Mix 6000 VO-3 0.5 g Pb/gal as TEL2 + 1 theory of ethylene dichloride Inven tion-2 5000 VO-4 No lead 4700 VO4 0.5 0.5gPb/galasTEL 3000 VO-i 0.5 g Pb/gal as Motor Mix 3000 VO--4 0.5 g Pb/gal asTEL 'Antiknock composition containing, by weight, 61.48% tetraethyllead, 18.81% ethylene dichloride, 17.86% ethylene dibromide and 1.85% dye, solvent, etc.

2Tetraethyllead Oxygen content of the exhaust was determined-polarographically using Beckman Model ON-i 1 polarograph, and carbon monoxide content was determined by infra red spectroscopy using Beckman Model 315 NDIR. The excess oxygen content was calculated from this data according to the following equation.

Percent Excess 2 = Measured 02-0.5 X Measured CO At the start of the tests, the engine was adjusted to have a slight excess of oxygen in the exhaust in accordance with manufacturer's recommendations. As sensors are subjected to prolonged exposure to the hot exhaust gases, their voltage output drops, thus transmitting a weaker signal to the A/F control system. As a result, the A/F ratio goes to the rich side (less oxygen) and this is evidence in a drop in the excess oxygen content of the exhaust gases. Table Ill reports the excess oxygen content of the exhaust at the beginning of the tests and after the vehicle had operated the number of miles stated. Table IV expresses the data in Table Ill in terms of percent change in excess oxygen content.It will be seen that the sensors of this invention operated for a significantly longer period of time before there was a shift in excess oxygen content that would be evidence of significant deterioration of the sensors.

TABLE Ill.

Sensor Performance and Durability Vehicle: 1977 240 Series Volvo Vol. Percent Excess 02% vs. MPH Sensor 15 20 25 30 35 40 45 50 Commer- 0.15 0.11 0.13 0.23 0.19 0.10 0.25 0.26 cial-1 atO miles expo sure Commer- 0.04 0.06 0.03 0.08 0.05 -0.03 0.10 0.16 cial-1 at 1000 miles expo sure Commer- 0.25 0.25 0.24 0.35 0.29 0.22 0.34 0.29 cial-2 atO miles expo sure Commer- 0.12 0.06 0.08 0.09 0.03 -0.07 0.10 0.08 cial-2 at 1000 miles expo sure Commer- 0.29 0.21 0.27 0.40 0.37 0.32 0.38 0.39 cial-3 atO miles expo sure Commer- 0.08 0.11 0.08 0.22 0.11 -0.01 0.17 0.27 cial-3 at 1000 miles expo sure TABLE Ill (cont'd.) Sensor Performance and Durability Vehicle: 1977 240 Series Volvo Vol. Percent Excess O2% vs.MPH Sensor 15 20 25 30 35 40 45 50 Commer- 0.17 0.04 0.06 0.10 0.04 -0.09 0.08 0.13 cial-3 at 5000 miles exposure Inven- 0.15 0.14 0.16 0.32 0.24 0.17 0.29 0.33 tion-1 atO miles exposure Inven- 0.18 0.18 0.18 0.32 0.24 0.13 0.29 0.32 tion- 1 at 2000 miles exposure Inven- 0.18 0.10 0.16 0.19 0.20 0.10 0.24 0.33 tion- 1 at 5000 miles exposure Inven- 0.21 0.28 0.62 0.63 0.61 - 0.66 0.66 tion-1 at 1 1000* miles expo sure Inven- 0.27 0.17 0.17 0.25 0.24 0.14 0.31 0.32 tion-2 atO miles exposure Inven- 0.19 0.16 0.20 0.32 0.27 0.18 0.35 0.39 tion-2 at 4700 miles exposure TABLE Ill (cont'd).

Sensor Performance and Durability Vehicle: 1977 240 Series Volvo Vol. Percent Excess O2% vs. MPH Sensor 15 20 25 30 35 40 45 50 Inven- 0.18 0.17 0.21 0.29 0.27 0.18 0.33 0.37 tion-2 at 12700 miles expo sure Inven- 0.13 0.16 0.16 0.26 0.24 0.15 0.33 0.35 tion-2 at 15700 miles expo sure *Data believed to be unreliable since a slight air leak is suspected in the sampling line.

TABLE IV.

EXCESS 02% vs.

Average (All 15 20 25 30 35 40 45 50- Speeds) % % % % % % % % % Sensor CHG CHG CHG CHG CHG CHG CHG CHG CHG Commer- -73 -45 -77 -65 -74 -70 -62 -38 -63 cial-1 at 1000 miles expo sure Commer- -52 -76 -67 -74 -90 -132 -61 -72 -78 cial-2 at 1000 miles expo sure Commer- -71 -48 -70 -45 -70 -103 -53 -31 -61 cial-3 at 1000 miles expo sure TABLE IV CONTINUED.

EXCESS O2% vs. MPH Average (All 15 20 25 30 35 40 45 50 Speeds) % % % % % % % % % Sensor CHG CHG CHG CHG CHG CHG CHG CHG CHG Commer- -39 -81 -78 -75 -89 -128 -78 -67 -79 cial-3 at 5000 miles exposure Inven- +20 +28 -12 0 0 -24 0 -3 +4 tion-1 at 2000 miles exposure Inven- +20 -28 0 -41 -17 -41 -17 0 -16 tion-1 at 5000 miles exposure Inven- +39 +100 +287 +66 +112 0 +90 +67 +108 tion-1 at 11000* miles exposure Inven- -30 -6 +18 +28 +13 +29 +13 +22 +11 tion-2 at 4700 miles exposure Inven- -33 0 +24 +16 +13 +29 +7 +16 +9 tion-2 at 12700 miles exposure Inven- -52 -6 -6 +4 0 -7 +7 +10 -9 tion-2 at 15700 miles exposure *Data believed to be unreliable since å slight air leak is suspected in the sampling iine The data summarized in the above tables show that commercial sensors when exposed to exhaust gases of a vehicle operating with leaded fuels show considerable shift to richer fuel-air mixture as indicated by the decrease in excess oxygen in the exhaust gases. As indicated by the average overall shifts for all speeds in Table IV, the shifts are such that after 1 ,000 miles of exposure, each of the commercial sensors shifted such that only about 30% of the original excess oxygen is present. In contrast, the invention sensors 1 and 2 show little or no shift up to about 1 5,000 miles of exposure.

Claims (11)

1. In an oxygen concentration sensor comprising a solid electrolyte, a porous electrode layer on one side of said solid electrolyte to communicate with an oxygen-containing reference gas, a second porous electrode layer on the opposite side of said solid electrolyte to communicate with combustion exhaust gases, the improvement which comprises a gas-permeable adherent coating covering at least the surface of said sensor which may be exposed to said exhaust gases, said coating comprising a metal oxide of perovskite crystal structure and having the chemical formula A(B,~,B',)O, wherein A represents one or more metal cations having ionic radii between about 0.8 and 1.65 Angstroms, B represents cations of one or more platinum group metals, B' represents cations of one or more nonplatinum metals having ionic radii between about 0.4 and 1.4 Angstroms, and y has a value of from 0 to about 0.99.

2. The sensor of Claim 1 wherein a porous gas permeable coating of magnesium spinel is interposed between the electrode layer and the platinum metal perovskite layer.

3. The oxygen sensor of Claim 2 wherein the perovskite coating is stabilized with 20% by weight of a crystal growth inhibitor based on total Weight of the coating, said inhibitor being a metal oxide selected from the group consisting of ThO2 and ZrO2.

4. The sensor of Claim 1 wherein y has a value of from about 0.80 to about 0.99.

5. The sensor of Claim 2 wherein B is platinum.

6. The sensor of Claim 2 wherein B is platinum and B' is cobalt.

7. The sensor of Claim 2 wherein B is a combination of platinum and ruthenium and B' is cobalt.

8. The sensor of Claim 2 wherein said composition represented by A(B"~yBty)02 is La0.0Sr0COc.8Pt0.2O3 stabilized with 20% by weight of ThO2 or ZrO2 based on total weight of the coating.

9. The sensor of Claim 2 wherein said composition represented by A(BivB'v)O3 is Lao.oSroCOo.Pton3Ruo.o3O3 stabilized with 20% by weight of ThO2 or ZrO2 based on total weight of the coating.

10. A sensing element as claimed in Claim 1 and having the structure described with reference to, and as illustrated in, the accompanying drawing.

11. A sensing element as claimed in Claim 1 and substantially hereinbefore described in any one of the Examples.

11. A sensing element as claimed in Claim 1 and substantially hereinbefore described in any one of the Examples.

New claims or amendments to claims filed on 18th May 1 979.

Superseded claims 10,11.

New or amended claims:

10. A sensing element as claimed in Claim 1 and having the structure described with reference to, and as illustrated in, the accompanying drawing.