ΞXTENOEO RANGE AIR FUEL RATIO SENSOR
Technical Field
This invention relates to determining the composition of a gaseous atmosphere.
Background Art
It is known to use high temperature oxygen sensors in the determination of a stoichiometric air fuel mixture in the exhaust gases of automobile internal combus¬ tion engines. The stoichiometric mixture is one in which the mass of air present contains just enough oxygen to react with the mass of hydrocarbons present so that there is the minimum amount of both oxygen and hydrocarbons remaining. The air fuel ratio (A/F - mass of air/mass of fuel) at the stoichiometric point is approximately 14.6. If, for example, an engine were running lean of stoichio- metry (A/F > 14.6), there would be an excess of air in the "charge" burned in the cylinder of an internal combustion engine and the exhaust gas would contain a substantial oxygen partial pressure. If rich operation were occurring (A/F < 14.6), the exhaust gas would contain unreacted or partially reacted hydrocarbons and very low oxygen partial pressure.
In particular, the equilibrium oxygen partial pressure in the exhaust gas can change by a great amount (as much as 20 orders of magnitude) as one moves from lean to rich operation. This large change forms the basis for detecting the stoichiometeric air fuel ratio with an exhaust gas oxygen sensor. The electrical output of such a sensor can then be fed back to an electrically
controllable carburetor or fuel injection system for maintaining engine operation at the stoichiometric point. Depending on engine type, operation at this point frequently offers a reasonable compromise for minimizing regulated exhaust gas emissions and maximizing engine performance. -___~
There are known high temperature oxygen sensors utilizing a single oxygen electrochemical concentration cell (usually made from zirconium oxide) and requiring the use of a reference atmosphere (usually air) which are suit¬ able for determining the stoichiometric air fuel ratio in a high temperature automotive environment. These devices give an output (EMF) proportional to the natural logarithm of the oxygen partial pressure. Despite their low sensi- tivity to oxygen partial pressure, the large change in oxygen partial pressure at the stoichiometric point allows their useful implementation. U.S. Patents 3,948,081; 3,738,341; 4,112,893; 4,210,509; and 4,107,019 relate to oxygen sensors of this type. O.S. Patent Nos. 3,907,657 to Heijne and
3,514,377 to Spacil et al relate to the measurement of oxygen (O2) concentrations using solid electrochemical devices. For applications at elevated temperatures ( > 500° C), for example, as might be encountered in the exhaust gases of furnaces or automobiles, the active material in these devices may be ceramic zirconium dioxide adapted for the conduction of oxygen ions. Electro¬ chemical cells made from this material are suitable at elevated temperatures for oxygen sensing and pumping applications.
The mode of operation of the Heijne device can be described as an oxygen counting mode in which oxygen partial pressure is determined on a sampling basis. A
OMP -fa TVIF
constant current is applied to an electrochemical cell which forms part of the enclosure of a volume for a period of time tp, for the purpose of electrochemically pumping out most of the oxygen from that volume. The ambient atmosphere is established within the volume prior to the pump out by means of a leak. An additional electro¬ chemical cell, which serves as a sensor of the reduced oxygen partial pressure within the volume and which also constitutes a portion of the enclosure, provides a signal indicating when oxygen has been sufficiently depleted from the volume (see Fig. 4 of Heijne). Knowing the tempera¬ ture, enclosed volume, pump out current and time allows one to calculate the number of oxygen molecules within the enclosure from the ideal gas law. The number of oxygen molecules is in turn proportional to the desired oxygen partial pressure. If a constant pump current is used, the pump out time tp is proportional to the oxygen partial pressure. If a constant current is not used, then the integral of the pump out current over the pump out time is proportional to the oxygen partial pressure.
The Heijne device can provide an output which is linearly proportional to the oxygen partial pressure. This is superior, for example, to single oxygen concentra¬ tion cells used as sensors which give an output (EMF) proportional to the natural logarithm of the oxygen partial pressure In (P0 ).
A potential disadvantage of the Heijne device is response time. For this measurement procedure, the leak connecting the ambient to the enclosed volume must be small so that during the pump out of oxygen, no signifi¬ cant amount of oxygen leaks into the volume to cause an
OMPI .,„ V/IPO
error in the count of molecules (i.e., to erroneously increase tp ). However, if the leak is made small, it may take a long time, tv, for the ambient to reestablish itself with the volume after a pump out. If the changes in the oxygen partial pressure in the ambient occur rapidly with respect to this refill time, the device would not be able to follow these changes in repetitive operation.
In the case of the teachings of U.S. Patent No. 3,698,384 to Jones, the purpose is to measure oxygen partial pressure in a feedgas. This is done by measuring an electrochemical cell pumping current while holding the sensor cell voltage a constant. However, the flow rate of the feedgas must be kept constant. If the flow rate should attempt to vary, there is a relatively elaborate flow control circuit to keep the flow rate a constant. This scheme, which also employes a reference atmosphere, is relatively unsuitable for application in an auto exhaust where the exhaust flow rate would change substantially with RPM.
Figs. 1 and 2 of the drawings illustrate a known oxygen pumping sensor in which ionically conducting zir¬ conium dioxide material 1 with thin platinum electrodes 2 and 3 form an electrochemical cell which with additional ceramic structure 4 defines an enclosed volume 6. The ambient atmosphere can establish itself within the volume by means of a leak opening 5. A battery 7 is attached to the electrodes by means of lead wires 8 and 8*. A volt¬ meter 10 and ammeter 9 are provided to determine the voltage drop across the pump cell and the current flowing through it. Although similar to structure to Fig. 5 of U.S. Patent No. 3,907,657, the operation is different. Here one applies a pump voltage V to remove oxygen from an
OMPI fa. IPO
enclosed volume 6 until the pump current saturates. The saturated current is proportional to oxygen partial pressure or concentration.
This is a steady-state device. When steady state is reached, the flow of oxygen through leak opening 5 equals the pump current times a proportionality constant. The current saturates because the leak aperture in combina¬ tion with the platinum electrode 2, the cathode, will only allow a limited (saturated) amount of oxygen to enter and be electroche ically pumped from the volume per unit time. To the extent that the saturated current value depends on the properties of the electrode 2, the device calibration may be subject to drift as these properties may change during the sintering and wear of this thin layer. For some engines it is useful to operate lean of the stoichiometric A/F ratio for the purpose of reducing fuel consumption. Oxygen partial pressure varies in a systematic way in the lean region and this can form the basis for determining lean A/F. However, the variation in oxygen partial pressure in the appropriate lean A/F region is not large (in comparison to the changes occurring near stoichiometry) , so that suitable oxygen sensors with sensitivities greater than the natural logarithm of oxygen partial pressure are desirable for accurate measurement in the desired A/F range.
Oxygen partial pressure sensors for engines operating lean of stoichiometry are taught in U.S. Patent Nos. 4,272,331 and 4,272,330 to R. E. Hetrick and U.S. Patent No. 4,272,329 to R. E. Hetrick et al. The sensors (shown as prior art in Fig. 3 of the drawings) are placed entirely in the exhaust gas stream and include two oxygen
OMP V/IP
ion conducting electrochemical cells 11 and 12, a pump cell and a sensor cell, which in part provide the enclosing structure of a nearly enclosed volume 13. A portion of the remaining structure can be a hollow ceramic tube 14. The cells can be attached to the end faces of the tube by ceramic glue 16. A small aperture 17 in the enclosing structure allows the exhaust gases, containing oxygen in a percentage to be determined, to leak into the volume. Lead wires 18 are affixed to electrodes 15 attached to each side of electrochemical cells 11 and 12.
The Hetrick and Hetrick et al patents describe various external circuitry which can be coupled to the sensors to permit operation in modes including an oscil¬ latory mode, a transient mode, and a steady-state mode. When operated in one of these modes, this device can be of great advantage in lean operation compared to the single-cell sensor since it affords a linear or greater sensitivity to oxygen concentration. Further, the various modes offer other advantageous features such as low temperature sensitivity and, in one case, independence from variations in absolute pressure.
In these modes, oxygen is electrochemically pumped into or out of the enclosed volume at a rate given by the pump-cell current, Ip. Simultaneously, oxygen diffuses into or out of the volume by means of the leak aperture. The oxygen fluxes due to leakage and pumping alter the oxygen partial pressure within the volume relative to the ambient so that an EMF (*■*■ Vs) develops across the sensor cell. The ambient oxygen partial pressure, which in turn is proportional to the A/F ratio, is dependent upon the relationship between I_ and Vs.
Further, in addition to the previously discussed stoichiometric and lean air/fuel operation, there are occasions where engine operation rich of stoichiometry is
OMPI
desired. In this region, the amount of partially reacted hydrocarbons (HC) such as carbon monoxide and hydrogen which increase with decreasing A/F can serve as a measure of A/F. Using oxygen pumping cells one can determine the A/F by measuring the rate or amount of oxygen which must be delivered to cause a measurable reaction with the partially reacted HC.
Thus, U.S. Patents 4,224,113 and 4,169,440 describe single-cell structures which combine electrochemical pumping of oxygen in zirconium oxide devices with the measurement of the current through, and potential difference across, that device to provide a measure of both lean and rich A/F values. However, such single-cell devices may be subject to significant loss of calibration (drifting) or deterioration with extended use as would be required in automotive applications. The potential drop across the pump cell, which for these devices is a critical parameter in establishing A/F, can be significantly affected by the quality of the cell electrodes. This arises because more or less potential difference may be required to assure that oxygen is passed through a thick or thin electrode at the necessary rate. Such electrode polarization phenomena are common. Thus, this electrode contribution to the potential difference may vary with time as the electrode sinters or otherwise deteriorates under high temperature usage. Further, the oh ic contribution to the potential difference across the cell will vary exponentially with temperature requiring tight temperature control causing possible penalties in cost and performance. An advantage of two-cell structures such as those described by the Hetrick and Hetrick et al. patents is that, the pump-cell potential difference is not a critical parameter thereby lessening the effects of electrode deterioration and temperature.
OMPI fa
Thus it can be appreciated that different sensor structures and different external circuitry are especially advantageous for A/F measurements in particular limited A/F regions. Current sensors which apply to a broader range may not possess the desirable features associated with the devices covering a more limited range. In any case, it would be desirable to have an exhaust gas oxygen sensor which could indicate engine A/F over an extended range of rich and lean A/F values, including stoichio- metry, which incorporate the most useful properties. These are some of the problems this invention overcomes.
DISCLOSURE OF THE INVENTION
In accordance with an embodiment of this inven¬ tion, an exhaust gas oxygen sensor for generating a signal indicative of A/F of an internal combustion engine genera¬ ting exhaust gases includes a first and a second electro¬ chemical cell. The two electrochemical cells are spaced from one another and define between them a partially enclosed volume. The volume is in communication with the exhaust gases through an opening. A first side of each of the first and second electrochemical cells is exposed to the partially enclosed volume. A second side of the first electrochemical cell is exposed to the exhaust gases. A second side of the second electrochemical cell is exposed to a reference atmosphere. In some cases, it may be desirable that the reference atmosphere is the ambient atmosphere. In other cases, it may be desirable that the reference atmosphere is an enclosed volume of a metal, metal-oxide compound.
__OMPI T -
A device in accordance with an embodiment of this invention can be used with different measurement tech¬ niques to determine exhaust gas A/F over a wide range of values including those richer than, leaner than, and near the stoichiometric air fuel value. Hence, the device has a "universal" air fuel sensing characteristic. Further, the cell structure allows the use of measurement techniques which are particularly advantageous in each of the three ranges.
BRIEF DESCRIPTION OF THE DRAWINGS
Figs. 1, 2 and 3 show the construction of prior art electrochemical oxygen pumping devices;
Fig. 4 is a schematic cross section of a sensor in accordance with a first embodiment of this invention;
Fig. 5 is a graphic representation of the sensor cell voltage, VB, versus an air fuel ratio, A/F, for the sensor shown in Fig. 4;
Fig. 6 is a schematic drawing of a sensor cell voltage, VA, versus pump cell current, Ip, at various rich air fuel values for a sensor in accordance with Fig. 4;
Fig. 7 is a graphic representation of the pump cell current, Ip, required to hold the voltage of the sensor cell at a reference voltage for various air fuel ratios, A/F, in accordance with the sensor of Fig. 4;
Fig. 8 is a schematic diagram of a sensor device, similar to that of Fig. 4, and external circuitry in accordance with an embodiment of this invention for measuring A/F; and Fig. 9 is a schematic cross section of a sensor in accordance with the second embodiment of this invention.
OMPI -^ IPO
BEST MODE FOR CARRYING OUT THE INVENTION
Referring to Fig. 4, an air fuel (A/F) sensor 110 includes an electrochemical cell 111 including a disk-like electrolyte 112 of a solid ionic conductor of oxygen such as Y2°3 3oped Zrθ2- Cell 111 also includes two thin porous catalytic platinum electrodes 113 with attached lead wires 114. Similarly, an electrochemical cell 121 includes an electrolyte 122, electrodes 123 and leads 124. Electrochemical cell 111 is separated from electrochemical cell 121 by a thin, hollow spacer 125 so that an enclosed volume v is defined. Cell 111 has a small hole or leak aperture 126 in it so that an ambient atmosphere, the exhaust gas, can establish itself within the volume v.
Electrochemical cell 121 has a thimble-like tubular shape closed at one end thereby defining a reference volume and exposing one side of cell 121 to a reference atmosphere. In particular, a flat disc-shaped electrolyte 122 has a tubular structure 131 attached to it to form the thimble-like shape. As a result, one side of the sensor is exposed to the exhaust gas and one side is exposed to the reference atmosphere. Alternatively, the electrolyte itself may have a thimble-like shape. In a similar way, cell 112 and spacer 125 might be made from a single piece of material or fabricated from two separate components as shown. A sensor supporting structure 128 provides a seal between exhaust and reference atmospheres and structural support and protection as well as allowing for attachment to the exhaust pipe wall 127. Openings 130 in a sensor support structure cover 228 allow easy access of the exhaust gas to sensor 110. Lead wires 114 and 124 are passed through a support structure 128 for attachment
to external circuitry. A heater 129 is provided to keep A/F sensor 110 within a desired operating temperature range.
Referring to Fig. 9, another embodiment in accor- 5. dance with this invention replaces the air reference on one side of cell 121 of Fig. 4 with a metal, metal-oxide mixture. Air fuel sensor 140 of Fig. 9 has an electro¬ chemical cell 141 with an electrolyte 142 and electrodes 143 attached to lead wires 144. Sensor 140 also has a 0 second electrochemical cell 145 with an electrolyte 146 coupled to electrodes 147 which are connected to lead wires 148. A spacer 149 separates cell 141 from cell 145. An aperture 150 in cell 141 provides access from an exhaust atmosphere into the enclosed volume of sensor 140. 5 A generally cup-shaped retaining structure 151 retains a metal metal-oxide mixture 152 adjacent to one side of electrochemical cell 145. Air fuel sensor 140 is posi¬ tioned completely within the exhaust gas stream and can be mounted on a support structure 153 which is mounted in an exhaust pipe wall 154. Use of air fuel sensor 140 pro¬ vides for fabrication simplicity and attendant reduced cost since no seal for sensor 140 is required between the exhaust and exterior atmosphere and the entire device can be contained within the exhaust gas. Referring to the operation of the device of Fig.
4, air fuel sensor 110 can be used with two different measurement techniques to determine exhaust gas air fuel ratio over a wide range of values including those richer than, leaner than and near the stoichiometric air fuel value. Hence, the device can be considered to have "universal" sensing characteristics. First, a steady-state oxygen-pumping mode is used for an extended
range of rich and lean air fuel ratio values. Second, the previously described single electrochemical cell technique is used near stoichiometry. The structure of the device of Fig. 4 permits use of multiple measurement techniques so that the functional advantages of each technique can be realized in a particular air fuel ratio region of applica¬ tion. The use of the single electrochemical cell for measuring A/F near stoichiometry is well known and is taught in U.S. Patent 3,948,081 to Wessel et al. The use of pumping techniques for lean A/F sensing using a two cell structure are taught by U.S. Patent 4,272,329 to Hetrick et al and U.S. Patents 4,272,330 and 4,272,331 to Hetrick. The disclosures of these patents are hereby incorporated by reference. The use of two-cell pumping techniques for rich A/F sensing are further discussed in a copending application filed on even date herewith entitled "Steady-State Method of Determining Rich Air/Fuel Ratios" by Hetrick et al.
When air fuel sensor 110 of Fig. 4 is used in connection with internal combustion engine operation at stoichiometric and near lean operation, such as air fuel ratios in the range of about 14.6 to about 17, leads 114 to cell 111 are disconnected and air fuel sensor 110 operates as a single electrochemical cell sensor previously described in connection with sensing stoichiometric air fuel ratios. The -equilibrium oxygen partial pressure for the exhaust gas, PEX, is established at the catalytic electrode 123 of cell 121 within volume v. In combination with the oxygen partial pressure in the air reference. PREP being equal to 0.2 atmospheres, an EMF, VB, is generated across cell 121 given by the Nernst equation:
VB - (RT/4F) In (PREF/PEX) U>
OMPI
where R is the gas constant, T is the absolute temperature and F is the Faraday constant. As P χ and the corres¬ ponding air fuel ratio decrease, the cell EMF increases as shown in Fig. 5. The strong variation of exhaust gas oxygen and, correspondingly, the cell EMF in this region makes this a relatively simple and desirable technique for the stoichiometric and near lean regions of air fuel ratio operation. At larger or smaller air fuel values, however, the variation of the cell EMF with air fuel becomes too small for a desirably effective sensor operation.
Advantageous modes of operation for lean air fuel ratios greater than about 15.5 are the steady-state, oscillatory or transient operating modes described by U.S. Patents 4,272,329; 4,272,330 and 4,272,331. In these modes oxygen is pumped into or out of the enclosed volume v by a pump cell, e.g. cell 121, while changes in the EMF induced on the other "sensor" cell, e.g. cell 111, are monitored. Due to the change in oxygen pressure within v from the combined effects of oxygen pumping and oxygen diffusion through leak aperture 126, systematic relation¬ ships occur between the pump-cell current Ip and the "sensor" cell EMF which provide a basis for oxygen sensing with high sensitivity in the lean region.
In particular, in the steady-state mode of operation, external circuitry causes a current Ip to be passed through the pump cell 121 to withdraw just enough oxygen from v so that a constant sensor cell EMF, termed Vs, is established. As the percentage of oxygen increases, so does the required pump current thereby providing a measure of the oxygen percentage and corres¬ ponding A/F. In particular, one finds that
τ9 " σ pEX (1 " e - s Vo) (2)
where V0 ■ RK/4F and σ is a constant of proportionality defining the rate at which oxygen can diffuse into v through the leak aperture. For example, σ is proportional to the oxygen diffusion coefficient and the area of the leak aperture. Thus, by always passing just enough Ip to keep Vs a constant, Ip *-. PEχ α» A/F thereby allowing for high sensitivity A/F sensing. For automotive applications it is also advantageous that this technique has weak temperature and absolute pressure sensitivities. The cited patents describe device operation for both sensor and pump cells completely immersed in the exhaust gas. Sensor 110 has an analogous mode of opera¬ tion even though the exterior electrode 123 of pump electrochemical cell 121 is exposed to a reference atmosphere with high oxygen concentration as shown in Fig. 4. The reason is that the effect of the reference atmosphere is to add a small increment to the total potential difference across pump cell 121. However, only the current, Ip, through the pump cell and not the potential drop across the cell is important for device operation. Accordingly, all lean operating modes described in these patents can be accomplished with the present structure where electrochemical cell 121 is used as the pump cell and electrochemical cell 111 is used as the sensor cell.
During operation with air fuel ratios rich of stoichiometry (A/F < 14.7), the concentrations of partially reacted HC increase with decreasing air fuel ratios thereby providing a measure of the air fuel ratio. In a manner analogous to that used for lean operation, a method to determine rich air fuel ratios with air fuel sensor 110 includes causing oxygen to be pumped into v from the reference atmosphere at a rate given by Ip.
Simultaneously, the oxygen partial pressure within v is decreased by oxygen diffusion through leak aperature 126 and chemical reaction of interior catalytic electrodes 123 and 113 with the partially reacted HC which continuously diffuses into volume v through leak aperture 126.
As pump cell current Ip increases, the equilibrium oxygen partial pressure within volume v increases causing an EMF to be induced across electrochemical cell 111. The magnitude of this EMF, termed VA, is again given by Equation 1 where PREF S replaced by Pv which represents the near equilibrium oxygen partial pressure within volume v resulting from the reaction of pumped oxygen and partially reacted HC. Since pv > PEX n this case, the sign of the EMF will be opposite that induced by pumping action during lean air fuel ratio measurement.
Figure 6 shows a plot of induced EMF, VA, versus pump current, Ip, at different rich air fuel ratio values.
The EMF is low for small pump currrents and increases with Ip« For lower air fuel ratio an ever increasing amount of oxygen must be pumped into volume v to accomplish a significant reaction with the HC. In particular, the value of Ip required to cause the EMF on electrochemical cell 111 to reach an arbitrary reference value VA(REF) (maintained in the external circuitry) will increase systematically with decreasing (i.e. richer) air fuel ratio as indicated in Fig. 7. Such a calibration curve provides the basis for measuring rich air fuel ratios.
The choice of VA(REF) would be influenced by a number of design considerations, but could for simplicity be chosen equal to, but opposite in sign, to the reference voltage used in the steady-state mode for lean sensing operation.
The magnitude of Ip will be an increasing function of cell volume and leak aperture size.
Measurement of A/F and subsequent feedback control of engine A/F could be achieved in a manner analogous to that employed for lean operation. A circuit similar to the one shown schematically in Fig. 8 would be attached to both cells. In Fig. 8, the supporting structure is not shown for clarity. Resistors Rl, R2 and capacitor C control the gain and frequency response of amplifier A so that A will always generate enough pump current Ip to maintain the EMF across cell 111 at a constant value equal to VA(REF). A resistor R3 is included in the pump cell circuit so that Ip can be determined by measuring the voltage across R3 with voltmeter V. Using the calibration curves of Fig. 6, the air fuel ratio would be determined. Using standard electronic circuitry this current can be compared to the value of Ip required for a desired air fuel ratio. If the current is too high or low, intake fuel could be increased or decreased, respectively, thereby accomplishing feedback control. Also shown is a temperature sensor 140, which in combination with the voltage drop across R3, form the inputs to correction circuitry 141, to adjust Ip to a temperature compensated value if necessary.
In summary, operation rich or lean of stoichiometry would be accomplished by pumping oxygen into or out of the enclosed volume until a predetermined VA (REF) appropriate for rich or lean conditions is achieved. With rich and lean calibration curves electronically avail¬ able as in an onboard computer, the measured and desired values of pump current are compared and a feedback or error signal, sent to an electrically controlled carburetor or fuel injection system, accomplishes feedback control.
Because of the highly exothermic nature of the HC-oxygen reaction, very small amounts of pumped oxygen can cause wide variations in VA at or near stoichiometry. Accordingly, the most appropriate technique in this region utilizes the conventional single electrochemical cell approach with a reference electrode at atmospheric oxygen partial pressure. Feedback control is achieved by comparing the output of the cell with that voltage corresponding to the desired air fuel ratio which is a known value and can be made electronically available in computer memory.
As a result, a single unit, sensor 110, provides high sensitivity to air fuel ratio both over an extended range of lean and rich conditions using a pumping mode of operation and near stoichiometry using a single electrochemical cell.
Alternatively, it may be advantageous to use cell 111 as the pump, removing oxygen from v and returning it to the exhaust, and cell 121 as the "sensor" in lean operation. This is possible with only a small modification to the operating results. As an example, one finds in the steady state mode that
τp * σ (PEX " PREF e-Vs/ 0 ) (3>
Thus, by adjusting lp to keep Vs always fixed (PREF is assumed to be constant) at an arbitrary value, Ip is still proportional to PEX although offset by a constant amount from the value found in Equation (2). A judicious choice of Vs will still allow convenient lean operation with high sensitivity.
OMPI
The advantage of this reversal of pump and sensor cells would be to eliminate current flow in cell 121 which may also be used as the sensor cell in subsequent stoichiometric operation. It is known that if current flow is too large in oxygen ion conductors, electrolyte or electrode deterioration can occur. This in turn could cause false or spurious EMF's to develop under open circuit conditions so that subsequent operation as a sensor cell would be compromised. In this case, however, : the fact that the sensor cell electrode is not immersed in the exhaust results in the air fuel ratio calibration curve which has some small sensitivity to absolute exhaust pressure.
In the embodiment shown in Fig. 9, the air reference is replaced by an alternate reference having metal-metal oxide mixtures 152 (e.g. Ni/Ni02, Cu/Cu02). The two-cell structure is similar to that shown in Fig. 4 except that the metal-metal oxide mixture is retained adjacent to the cell 145 reference electrode 147 by a retaining structure 151. This embodiment is appropriate for lean and stoichiometric operation where cells 145 and 141 act as sensor and pump cells, respectively. Since the effective oxygen partial pressure at a typical metal-metal oxide reference electrode, P-4/M0 (REF), is much less than - ?E under lean conditions, a substantial EMF (e.g. 200-500 mV) will appear across the sensor cell 145 at Ip » 0. As oxygen is pumped from volume v9 by pump cell 141, this EMF will be reduced. Choosing an appropriate EMF in this reduced range as a reference value, analysis analogous to that used in U.S. Patent 4,272,329 shows that the pump current required to keep the reference voltage constant is proportional to the percentage of oxygen in the exhaust gas thereby serving as a sensor of lean air fuel ratio as
OMPI
■ WIrO
in the previously discussed cases. For .near stoichio¬ metric operation the pump cell 141 is disconnected and the open circuit EMF of sensor cell 145 is monitored. As for other single-cell sensors, passage of the exhaust gas through stoichiometry is attended by a large variation in the cell EMF which is adequate to determine air fuel ratios in a narrow range. A pumping mode for rich air fuel ratio detection requires that sufficient oxygen be available. Various modifications and variations will no doubt occur to those skilled in the various arts to which this invention pertains. For example, the electrochemical cell shape may vary from that disclosed herein. These and all other variations which basically rely on the teachings through which this disclosure has advanced the art are properly considered within the scope of this invention.