Electrochemical NO2 Sensor Using a NiFe1.9Al0.1O4 Oxide Spinel

lean-burn and GDI (gasoline direct injection) technology has been put forward as an alterative to achieve maximum power and reduce fuel consumption...
2 downloads 0 Views 531KB Size
Download PDF
Anal. Chem. 2007, 79, 3561-3567

Electrochemical NO2 Sensor Using a NiFe1.9Al0.1O4 Oxide Spinel Electrode Weizhen Xiong and Girish M. Kale*

Institute for Materials Research, University of Leeds, Leeds LS2 9JT, United Kingdom

A novel solid-state electrochemical sensor using (Sc2O3)0.08(ZrO2)0.92 (ScSZ) electrolyte solid and a NiFe1.9Al0.1O4 oxide spinel electrode was tested for the detection of NO2 at temperatures greater than 700 °C for automobile applications. The sensor was found to respond rapidly, reproducibly, and selectively to NO2 at 703 and 740 °C. The response time of the sensor was approximately 8 s, and the recovery time was 10 s at both 703 and 741 °C. The response of the sensor was highly reproducible to the change in concentration of NO2 and also showed negligible cross-sensitivity to potentially interfering gases such as O2, CO, and CH4 in the gas stream. The detrimental effects such as global warming and health hazards of gaseous environmental pollutants that are released into the environment by different types of fossil fuel combustion processes in automobiles and power generation plants urges scientific researchers to develop new technologies to cope with this problem. In automobiles, three-way catalysts (TWC) combined with oxygen sensors promote deep oxidation of hydrocarbon and CO and simultaneous reduction of NOx (x ) 1, 2) to N2 before the exhaust gases are released into the atmosphere. Recently, the lean-burn and GDI (gasoline direct injection) technology has been put forward as an alterative to achieve maximum power and reduce fuel consumption. However, this leads to an increase in NOx emission. Some approaches, including the installation of a NOx storage catalyst and on-board NOx sensors, have been currently explored to compensate for the low NOx removal ability of the TWC in the lean-burn situation.1 In this situation, two hightemperature NOx sensors, one before and the other after the NOx storage catalyst, are required. The installation of NOx sensors serving the function of both a monitoring and a controlling device is likely to indicate the timing for the regeneration of the NOx storage catalyst. Various types of ceramic gas sensors have been researched so far, e.g., resistive-type sensors using oxide semiconductors such as SnO2 or electrochemical sensors incorporating ceramic solid electrolytes and a variety of binary and ternary oxides including spinels as sensing electrodes.2-14 The harsh environment and * To whom correspondence should be addressed. E-mail: [email protected]. (1) Miura, N.; Nakatou, M.; Zhuiykov, S. Sens. Actuators, B 2003, 93, 221228. (2) Miura, N.; Wang, J.; Nakatou, M.; Elumalai, P.; Hasei, M. J. Solid State Electrochem. Lett. 2005, 8, H9-H11. (3) Cantalini, C. J. Eur. Ceram. Soc. 2004, 24, 1421-1424. 10.1021/ac061944v CCC: $37.00 Published on Web 04/10/2007

© 2007 American Chemical Society

limited selectivity to targeted gaseous species in the exhaust gas mixture makes the semiconducting gas sensors impossible to last for a longer exposure. Further, they exhibit slower response and are unable to detect the very low gas concentrations required by the automotive industries in order to satisfy environmental legislation. The potentiometric NOx sensors based on a mixedpotential principle seem to exhibit greater potential for practical application because of their higher sensitivity especially in the lower concentration range, less than 100 ppm, compared to that of amperometric gas sensors which need expensive and sophisticated control units to attain the linear characteristic of the sensor signal within the concentration range.15,16 In a mixed-potential NO2 gas sensor, the mixed potential at each electrode is established when the kinetics of the rates of NO2 reduction and oxygen ion oxidation are equal. The measured open-circuit electromotive force (EMF) is the difference in mixed potentials established at each electrode. Several researchers have reported the ability of a single NOx sensor to detect both NO and NO2 in oxygen-rich atmospheres in the temperature range of 400-700 °C employing single or complex metal oxides such as spinels as the sensing electrodes.6,13,14 In spark ignition engines, the temperature of the exhaust gas near the engine site is typically about 600-700 °C and is slightly lower for diesel engines. However, the temperature of the engine exhaust in a vehicle occasionally reaches above 700 °C during acceleration or during high-speed driving conditions.2 The operation of these sensors at temperatures higher than 700 °C seems to be difficult because of the drastic decrease in the NOx sensitivity with increasing temperature. So far, there is no report on reliable, reproducible, and fast-responding planar mixedpotential NOx sensors exhibiting equal and high sensitivity, to the (4) Mukundan, R.; Brosha, E. L.; Garzon, F. H. J. Electrochem. Soc. 2003, 150, H279-H284. (5) Li, X.; Xiong, W.; Kale, G. M. Electrochem. Solid-State Lett. 2005, 8, H27H30. (6) Xiong, W.; Kale, G. M. Electrochem. Solid-State Lett. 2003, 8, H49-H53. (7) Dutta, A.; Kaabbuathong, N.; Grilli, M. L.; Bartolomeo, E. D.; Traversa, E. J. Electrochem. Soc. 2003, 150, H33-H37. (8) Xiong, W.; Kale, G. M. Sens. Actuators, B 2006, 119, 409-414. (9) Xiong, W.; Kale, G. M. Int. J. Appl. Ceram. Technol. 2006, 3, 210-217. (10) Hibino, T.; Kuwahara, Y.; Wang, S.; Kakimoto, S.; Sano, M. Electrochem. Solid-State Lett. 1998, 1, 197-199. (11) Miura, N.; Raisen, T.; Lu, G.; Yamazoe, N. Sens. Actuators, B 1998, 47, 84-91. (12) Szabo, N. F.; Dutta, P. K. Solid State Ionics 2004, 171, 183-190. (13) Lu, G.; Miura, N.; Yamazoe, N. J. Mater. Chem. 1997, 7, 1445-1449. (14) Zhuiykov, S.; Muta, M.; Ono, T.; Hasei, M.; Yamazoe, N.; Miura, N. Electrochem. Solid-State Lett. 2001, 4, H19-H21. (15) Riegel, J.; Neumann, H.; Wiedenmann, H. M. Solid State Ionics 2002, 152153, 783-800. (16) Ramamoorthy, R.; Dutta, P. K.; Akbar, S. K. J. Mater. Sci. 2003, 38, 42714282.

Analytical Chemistry, Vol. 79, No. 10, May 15, 2007 3561

best of our knowledge, to selectively detect both NO and NO2 at temperatures above 700 °C incorporating oxide spinel solid solutions as the sensing electrode. In earlier investigations, the sensitivity to NO has always been found to be significantly lower than to NO2.13,14 In this work, seven different oxide spinel solid solutions and composite materials, NiFe0.5Cr1.5O4, NiFe0.75Cr1.25O4, NiFe1.25Cr0.75O4, NiFe1.75Cr0.25O4, NiFe1.9Al0.1O4, NiO + NiCr2O4, and CuO + NiCr2O4, have been employed as the sensing oxide electrode material and their sensitivity and selectivity for NO2 has been examined. The choice of the oxide spinels is dictated by our quest to understand the effect of the catalytic activity of transition metal cations and their distribution in crystallographically nonequivalent tetrahedral and octahedral sites of a cubic spinel lattice. Among these oxide spinel solid solution and composite oxide sensing electrode materials investigated in this study, NiFe1.9Al0.1O4 has been found to exhibit good sensing behavior and good selectivity to NO2 compared to CO and CH4 between 703 and 740 °C. EXPERIMENTAL SECTION Materials. Oxide spinel materials in the (NiO)(Fe2O3)x(Cr2O3)1-x (x ) 0, 0.25, 0.375, 0.625, 0.875) system and NiFe1.9Al0.1O4 were prepared according to the process described by Poomiapiradee et al.17 Oxide spinels of different compositions were prepared using conventional solid-state ceramic routes. The powders of NiO and Al2O3, Cr2O3, or Fe2O3, used as precursor materials for the synthesis of the desired oxide spinel products, were obtained from Aldrich Chemicals Co. Ltd. (Gillingham, U.K.). Powders of NiO + NiCr2O4 and CuO + NiCr2O4 mixed oxides were prepared by intimately mixing fine powders of NiO or CuO and NiCr2O4 using a pestle and mortar for at least 2 h. Fine powders of NiO and CuO were obtained from Aldrich Chemicals Co. Ltd. (Gillingham, U.K.), whereas the fine powder of NiCr2O4 was prepared following the procedure described above. High-purity fine powders of Sc2O3 and ZrO2 used for the synthesis of 8 mol % scandium oxide stabilized zirconia (ScSZ) solid electrolyte were obtained from Aldrich Chemicals Co. Ltd. (Gillingham, U.K.). Pt-wire and Pt-gauze were obtained from Goodfellow Ltd. (Cambridge, U.K.). Pt-ink (Engelhardt A4731) used for making electrodes and electrode contacts was obtained from Johnson Matthey Ltd. (Enfield, U.K.). High-purity certified gases such as 481 ppm NO2 in air, 510 ppm CO in Ar, and 515 ppm CH4 in air were obtained from BOC (Leeds, U.K.) or Air Products (Leeds, U.K.) as appropriate. Solid Electrolyte. ScSZ electrolyte was synthesized by a conventional solid-state ceramic route. The detailed process for the synthesis of ScSZ has been reported elsewhere.6 The pellets were sintered at 1700 °C for 15 h in order to obtain the dense ScSZ solid electrolyte pellets. The sintered pellets were found to have a relative density of more than 90%. Thin slices of approximately 600 µm were cut out of the dense pellets using an Accutom-2 (Struers, Glasgow, U.K.) for the fabrication of the electrolyte-supported NO2 sensor. Sensing Electrode. The Pt-ink was applied to one side of the 600 µm thick ScSZ disk, and the surface was heat treated at 1000 °C for 1 h to form a Pt electrode. The oxide paste was made by (17) Poomiapiradee, S.; Brydson, R. M. D.; Kale, G. M. In Ceramic Transactions; Kale, G. M., Akbar, S. A., Liu, M., Eds.; American Ceramic Society: Westerville, OH, 2001; Vol. 130, pp 79-89.

3562

Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

mixing the fine oxide powder and cellulose in equal weight ratio with a further addition of several drops of turpentine oil. The sensing electrode paste was applied on the other surface of the ScSZ disk and was calcined at 800 °C for 3 h. A Pt-wire spot welded to Pt-gauze served as the electrical leads for the cathode and anode. Characterization. The NiFe0.5Cr1.5O4, NiFe0.75Cr1.25O4, NiFe1.25Cr0.75O4, NiFe1.75Cr0.25O4, NiFe1.9Al0.1O4, NiO + NiCr2O4, and CuO + NiCr2O4 oxide electrode materials were characterized by X-ray diffraction analysis using Cu KR1 radiation (λ ) 1.5406 Å). The experimental diffraction patterns were recorded at ambient temperature over a range of 10 e 2θ e 90 (step-scanned at 0.02° s-1). The microstructure of the oxide sensing electrode of the NO2 sensor was examined using a scanning electron microscope equipped with an energy-dispersive X-ray analysis facility (LEO 1530, Carl Zeiss SMT, Cambridge, U.K.). Sensing Measurements. The NO2 sensor tested in the present investigation can be schematically represented as

(-) m-NO2 (CO or CH4) + n-air, Pt-wire|Pt-gauze|Pt|ScSZ|NiFe1.9Al0.1O4|Pt-gauze| Pt-wire, m-NO2 (CO or CH4) + n-air (+) (I) where each vertical line indicates an interface between two different materials and m and n are the flow rates of test gas and air, respectively. The NO2 sensing experiments and tests for the examination of cross-sensitivity of the above sensor to CO and CH4 were carried out in a conventional gas flow apparatus equipped with a heating facility. The schematic diagram of the sensor testing apparatus, the corresponding sensor structure, and the detailed procedure for testing the gas sensor has been presented elsewhere.6 RESULTS AND DISCUSSION Sensing Mechanism. According to the literature,2,4-14 the measured open-circuit EMF of a mixed-potential sensor is essentially due to the difference in the electrokinetics of the redox equilibria at each electrode. In the presence of an air/NO2 mixture, each of the electrodes in the planar NO2 sensor will tend to reach a steady-state mixed potential determined by the kinetics of the electrochemical redox reactions, viz.,

2NO2(g) + 4e- f 2NO(g) + 2O2-

(1)

2O2- f O2(g) + 4e-

(2)

where O2- represents an oxide ion in the ScSZ lattice and the other terms have their usual meaning. Electrodes composed of materials with dissimilar catalytic properties will tend to reach different steady-state potentials. Hence, the measured sensor EMF is the difference between the two different steady-state potentials (mixed potentials) resulting from the dissimilar (e.g., Pt and NiFe1.9Al0.1O4) electrodes. At high overpotential, the current density-overpotential relationship for the electrode reaction can be approximated to be the Tafel behavior, and hence the EMF can be given by the expression

EMF ) C1 + C2 ln CNO2 + C3 ln CO2

(3)

Figure 1. Sensing properties of the NO2 sensor using different oxides as the sensing electrode to 463 ( 5 ppm NO2 at 700 ( 7 °C.

where C1, C2, and C3 are constants. Equation 3 suggests that at a constant CO2 condition, the EMF varies linearly with the logarithm of CNO2. Further, it is clear from the literature13,14 that the sensor described in cell I will register a positive EMF value for NO2 and a negative value for NO in the test gas at a fixed temperature. Selection of Oxide Sensing Electrode. Seven different kinds of oxide compositions, NiFe0.5Cr1.5O4, NiFe0.75Cr1.25O4, NiFe1.25Cr0.75O4, NiFe1.75Cr0.25O4, NiFe1.9Al0.1O4, NiO + NiCr2O4, and CuO + NiCr2O4, were strategically used as sensing electrodes, and the EMF of the resulting devices was measured in the oxygen-rich gas mixture containing an NO2 concentration between 100 and 500 ppm over the temperature range between 700 and 740 °C. The measured EMF of the sensor, sensitivity, and 90% response time of the NO2 sensor incorporating different oxide electrodes at 700 ( 7 °C at a gas flow rate of 250 mL min-1 is shown in Figure 1 for 463 ( 5 ppm NO2. The sensitivity data shown in Figure 1 is defined as the value of the slope of the measured EMF of the NO2 sensor as a function of the variation of gas concentration assuming a Tafel-type relationship at high overpotential between the measured EMF and NO2 concentration given by eq 3. At temperatures in excess of 600 °C, the catalytic activity of the Pt electrode for the redox reaction of the minor component of the gas mixture is high and therefore its sensitivity is very low or even becomes negligible toward NO2, CO, or hydrocarbon.2,10,11 Hence, the EMF response of the planar device represented by cell I, investigated in this study, is likely to originate mainly from the mixed potential established at the oxide electrode at temperatures above 600 °C. As seen in Figure 1, the response of the sensor strongly depends on the oxide used as a sensing electrode. It can be seen from Figure 1 that NiFe1.9Al0.1O4 exhibits the highest measured EMF and sensitivity and fast response to NO2 at 700 ( 7 °C among the different oxides being used. Although the sensing properties of the mixed-potential NO2 sensor incorporating NiFe1.75Cr0.25O4 are comparable to that of the mixed-potential NO2

sensor using NiFe1.9Al0.1O4, the NO2 sensitivity of the mixedpotential NO2 sensor using NiFe1.75Cr0.25O4 was found to be significantly affected by the presence of oxygen.18 In addition to this, it is clear from Figure 1 that although the 90% response time of the NO2 sensor using NiFexCryO4 for 0.5 e x e 1.75 and 1.5 g y g 0.25 where x + y ) 2 remains almost constant between 5 and 7 s, the sensitivity and the magnitude of the EMF of the NO2 sensor progressively decreases with increasing substitution of Fe3+ by Cr3+ in the octahedral sites of the oxide spinel electrode. Since NiFe2O4 is a partially inverse spinel and NiCr2O4 is completely normal, the cation distribution of the oxide spinel electrode can be represented by the general chemical formula as (Nix+y2+ Fe1-x-y3+)tet[Ni1-x-y2+ Fe1+x-y3+ Cr2y3+]octO4. There is no report of such an observation in any of the earlier studies, and we strongly believe that the trend in the measured EMF and sensitivity of the NO2 sensor may be due to the higher catalytic activity of the trivalent chromium (Cr3+) cation in the octahedral site of the spinel lattice for the reduction of NO2 to NO according to reaction 1. We have also noticed that the EMF of the NO2 sensor employing certain oxide spinel electrodes is significantly dependent on the oxygen concentration of the gas mixture in the presence of NO2 suggesting that the kinetics of O2- oxidation at these oxide spinel electrodes according to reaction 2 is significantly slower than that at the Pt reference electrode. Since this is the first observation, we are in the process of conducting a systematic investigation of understanding more of this phenomenon. Figure 1 also indicates that when a transition metal cation such as Cr3+ is substituted by a nontransition metal cation such as Al3+, it is likely that the Al3+ will prefer to go to the tetrahedral site unlike Cr3+ due to the lower octahedral site preference energy of Al3+ (-78 kJ mol-1) compared to that of Cr3+ (-158 kJ mol-1),19 (18) Xiong, W. Development of Solid-State NO2 Sensor for Monitoring High Temperature Process Emission. Ph.D. Thesis, University of Leeds, 2006. (19) Kale, M. G. Thermodynamic Studies of Selected Ceramic Oxide Systems. Ph.D. Thesis, Indian Institute of Science, 1990.

Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

3563

Figure 2. (a) XRD pattern of the NiFe1.9Al0.1O4 oxide spinel solid solution electrode before and after the sensor testing and (b) an SEM image of the NiFe1.9Al0.1O4 oxide electrode after calcining at 800 °C for 3 h.

and this therefore perhaps significantly alters the catalytic activity for NO2 reduction and O2- oxidation at the sensing electrode relative to that of the electrode materials containing Cr3+ thus giving rise to the higher EMF, higher sensitivity, and lower 90% response time of the NO2 sensor. Since NiFe1.9Al0.1O4 appears to be the most promising candidate for the sensing electrode among the list of sensing electrodes used in this study, detailed sensing properties of only the NO2 sensor using NiFe1.9Al0.1O4 are presented below. XRD and SEM Characterization. The XRD pattern of the NiFe1.9Al0.1O4 oxide electrode before and after the sensor testing campaign is shown in Figure 2a. The diffraction pattern of the electrode material in two different conditions is compared with the standard diffraction pattern reported in the Joint Committee of Powder Diffraction Standards (JCPDS) compilation. The XRD pattern of the solid electrolyte confirmed that the ScSZ solid electrolyte crystallizes in a cubic solid solution formed by Sc2O3 and ZrO2 (JCPDS file no. 27-997) end-members after sintering at 1700 °C for 15 h. A few minor peaks of ScSZ including one at 3564 Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

around 50° (2θ) in the XRD pattern of the NiFe1.9Al0.1O4 electrode after the sensor testing, shown in Figure 2a, are due to the contamination of the NiFe1.9Al0.1O4 oxide electrode stripped from the surface of the solid electrolyte by the mechanical abrasion. The XRD analysis of NiFe1.9Al0.1O4 before and after the sensor testing indicated that the chemical composition of the NiFe1.9Al0.1O4 oxide electrode remained unchanged during the sensor testing campaign and unaffected by thermal cycling. The SEM image of the NiFe1.9Al0.1O4 electrode shown in Figure 2b indicates that the NiFe1.9Al0.1O4 electrode calcined at 800 °C for 3 h was highly porous and consisted of homogeneous particles between 1 and 5 µm. The porous structure of the oxide sensing electrode has been found to enhance the molecular gas diffusion and attainment of rapid equilibrium at the three-phase boundary (TPB) consisting of the electrode/electrolyte/gas phase, which eventually leads to the faster response and enhanced gas sensitivity.5,6,8 Gas Flow Rate Influence. The influence of the gas flow rate on the sensing performance of the NO2 sensor is shown in Figure

Figure 3. Influence of gas flow rate on the EMF of the NO2 sensor to 463 ppm NO2 at 658 ( 2 °C.

3. The EMF of the sensor to 463 ppm NO2 at 658 ( 2 °C increased continuously from 58.04 to 89.9 mV with an increase in the total gas flow rate from 100 to 250 mL min-1, whereas a further increase in flow rate from 250 to 300 mL min-1 caused a decrease in the sensor EMF. This is possibly due to the compromising effect of chemical reduction of NO2 before it reaches the TPB and diminished adsorption of NO2 at the triple point formed at the electrode-electrolyte interface. To obtain an optimum sensor signal, the flow rate of 250 mL min-1 was chosen to test the sensor performance in the subsequent investigations. NiFe1.9Al0.1O4 Sensing Performance. The response of the NO2 sensor using NiFe1.9Al0.1O4 as a function of increasing and decreasing NO2 concentration between approximately 100 and 500 ppm at 703 and 740 °C is illustrated in Figure 4. In each diagram of Figure 4, the top half corresponds to the actual concentration of NO2 in the gas mixture as determined by the electronic mass flow controllers and the lower half corresponds to the measured EMF of the sensor at a fixed temperature as a function of NO2 concentration and time. The response of the NO2 sensor using NiFe1.9Al0.1O4 shown in Figure 4, parts a and b, clearly indicates that the sensor responds to changes in concentration of NO2 very rapidly and reproducibly between 100 and 500 ppm NO2 at 703 and 740 °C. The response of the NO2 sensor using NiFe1.9Al0.1O4 was found to be stable within (0.6 mV at 703 °C and (0.9 mV at 740 °C, respectively, over the entire range of NO2 concentrations. Similarly, the NO2 sensor was found to respond reproducibly at 564, 611, and 658 °C over the entire range of NO2 concentrations. It is important to examine the effect of potentially interfering gaseous species on the response of the sensor, i.e., to obtain a measure of selectivity. Therefore, in this study, the EMF of the sensor was measured as a function of oxygen concentration between 16.2 and 4.2 vol %. The test gas was prepared by diluting air with nitrogen. The response of the sensor to changes in the concentration of O2 at 703 °C is shown in Figure 5. The response of the NO2 sensor to variations of oxygen concentration was found to be insignificant when NiFe1.9Al0.1O4 was employed as the sensing electrode at temperatures above 700 °C. The cross-sensitivity of the NO2 sensor using NiFe1.9Al0.1O4 as the sensing electrode to other gases such as CO and CH4 at 703

Figure 4. Time dependence of the response of the NO2 sensor using NiFe1.9Al0.1O4 as a sensing electrode at (a) 703 and (b) 740 °C. The total flow rate of gas was 250 mL min-1.

Figure 5. Response of the NO2 sensor as a function of oxygen concentration at 703 °C. The total flow rate of gas was 250 mL min-1.

and 740 °C at a total gas flow rate of 250 mL min-1 has also been tested and the data are plotted in Figure 6. The concentration of NO2, CO, and CH4 was varied between 100 and 500 ppm, approximately. The variation of the measured EMF of the NO2 sensor as a function of the concentration of NO2, CO, and CH4 is shown in Figure 6, parts a and b, at 703 and 740 °C, respectively. Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

3565

Figure 7. Stability and reproducibility of the EMF of the NO2 sensor using NiFe1.9Al0.1O4 at (a) 741 °C and (b) a magnified view of the local region marked (1) in (a).

Figure 6. Variation of the EMF of the NO2 sensor as a function of the concentration of NO2, CO, and CH4 gases at (a) 703 and (b) 740 °C. The total flow rate of gas was 250 mL min-1.

It can be seen in Figure 6 that the variation of the EMF of the NO2 sensor as a function of the logarithm of NO2 concentration was found to be linear between 100 and 500 ppm. It can also be clearly seen from Figure 6, parts a and b, that the measured data fits very well with the response of the sensor based on the mixedpotential theory (eqs 1-3) discussed earlier suggesting that the electrochemical reactions that are responsible for establishing the mixed potential of cell I are governed by the Tafel relation for NO2 concentrations between 100 and 500 ppm. Further, the sensitivity of the NO2 sensor (Figure 6) was found to be significantly higher for NO2 compared to that for CO and CH4. For example, the sensor EMF at 740 °C to 462.5 ppm NO2 was 22.14 mV, while the response to 408 ppm CO and 412 ppm CH4 was -1.71 and -5.04 mV, respectively. The negative response of the sensor to reducing gases such as CO and CH4 is in agreement with the similar response obtained to NO by other researchers.13,14 Similarly, a high selectivity to NO2 and high degree of discrimination in sensitivity between NO2 and CO/CH4 was observed at 703 °C. The trend in measured EMF of the NO2 sensor shown in Figure 6 as a function of increasing and decreasing concentration of NO2 clearly demonstrates that the 3566 Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

planar sensor responded reproducibly to NO2 at 703 and 740 °C with negligible interference from CO and CH4. The stability of the measured EMF of the NO2 sensor using NiFe1.9Al0.1O4 was periodically tested at 741 °C by introducing rectangular pulses of 463 ppm NO2 in an oxygen-rich atmosphere, and the results are presented in Figure 7. The measured EMF of the sensor was found to be highly reproducible and reversible as shown in Figure 7a although a slight increase in the measured EMF was observed with time over a period of 30 min. The response time, defined as the time required by the sensor to attain 90% of the stable EMF at a fixed temperature, was found to be 8 s, whereas the recovery time, defined as the time required by the sensor to attain within 10% of the initial EMF value, was found to be 10 s at 741 °C as seen in Figure 7b. The smallest time interval for data collection in all the measurements was approximately 1.9 s. The rapid and stable response characteristic of the NO2 sensor using NiFe1.9Al0.1O4 obtained in this investigation is mainly attributed to the phase composition, cation distribution, and microstructure of the novel NiFe1.9Al0.1O4 sensing electrode. A schematic diagram of electrode reactions occurring on Pt and the NiFe1.9Al0.1O4 sensing electrode is shown in Figure 8. The porous microstructure provides the high accessibility of the gas phase to reach the TPB formed between the gas phase, the ScSZ electrolyte, and the NiFe1.9Al0.1O4 sensing electrode with the least possible hindrance for gas adsorption-desorption phenomena and

Figure 8. Schematic diagram of electrode reactions occurring on Pt and the NiFe1.9Al0.1O4 sensing electrode, assuming only NO2 reduction and oxygen ion oxidation occur.

electrode reactions on the NiFe1.9Al0.1O4 sensing electrode, thereby enhancing the sensor sensitivity. CONCLUSIONS A novel solid-state NO2 sensor using ScSZ electrolyte and a NiFe1.9Al0.1O4 oxide electrode was tested for the detection of NO2 at temperatures higher than 700 °C for emission control and monitoring applications. The sensor was found to be selective to NO2. The sensing characteristics were studied in the temperature range of 564-740 °C in an oxygen-rich atmosphere. The NO2 sensor was found to respond rapidly, reproducibly, and selectively to NO2 at 703 and 740 °C. The response time of the sensor was approximately 8 s, and the recovery time was 10 s at 741 °C. The response of the sensor was found to be logarithmically dependent on the concentration of NO2 between 100 and 500 ppm. The response of the sensor was highly reproducible to change in the concentration of NO2 and also showed negligible cross-sensitivity to O2, CO, and CH4 in the gas stream above 700 °C. The rapid and stable response characteristic of the NO2 sensor using

NiFe1.9Al0.1O4 obtained in this investigation is mainly attributed to the phase composition, cation distribution, and microstructure of the novel NiFe1.9Al0.1O4 sensing electrode. The fast response characteristics, high selectivity, and high reproducibility of the NO2 sensor indicate the promising future application of the NiFe1.9Al0.1O4 sensing electrode-based NO2 sensor for on-board exhaust gas monitoring. ACKNOWLEDGMENT G.M.K. thanks EPSRC for supporting this research through a LANTERN JIF Grant (GR/M88167). W.X. thanks the ORS committee for awarding the ORS and the Tetley-Lupton scholarships from 2002-2005 and IMR during the course of work.

Received for review October 16, 2006. Accepted March 5, 2007. AC061944V

Analytical Chemistry, Vol. 79, No. 10, May 15, 2007

3567