Lean burn diesel engines operate at a high air-to-fuel ratio to improve fuel efficiency and reduce the emissions of hydrocarbons (HC) and CO. The higher amounts of NO_x (NO and NO₂) produced under lean-burn conditions require a complex catalytic treatment system where an initial diesel oxidation catalyst eliminates HC and CO and an ammonia-mediated selective catalytic reduction (SCR) system converts NO_x to N₂. Excess ammonia is then oxidized to N₂ and H₂O through an ammonia oxidation catalyst.^1,2 To enable the real-time tuning of the SCR system, sensors are required which can simultaneously quantify NH₃ and NO_xunder exhaust gas conditions.

We have designed three-electrode mixed-potential sensors for real-time detection of NO_x and NH₃. The sensors were manufactured by ESL ElectroScience on an insulator-yttria-stabilized zirconia (YSZ)-insulator substrate. Dense electrodes of La_0.8Sr_0.2CrO₃ (LSCO), Au_0.5Pd_0.5 (Au/Pd), and Pt and a porous layer of YSZ electrolyte have been deposited by screen printing. The Au/Pd+Pt pair is sensitive to NH₃ under open circuit conditions, while under a negative current bias, the sensitivity of the LSCO+Pt pair to NO_x is enhanced and the sensitivity to NH₃ is suppressed. The tunable sensitivity allows us to quantify the concentration of NO_x and NH₃ separately. These mixed potential sensors are also expected to have fast response times, sufficient robustness in the exhaust gas conditions, and are manufactured in processes that are readily scaled up for mass production.

We take advantage of the ability of artificial neural networks (ANNs) to learn continuous functional relationships without human intervention to decode the sensor signals in the presence of cross interference. ANNs have also been successfully used in a broad range of mobile applications which makes them an attractive tool for developing portable analyzers and on-board vehicle diagnostic systems. We have previously demonstrated this technique can quantify mixtures of NO_x, C₃H₈, and CO.³ The ANN model used for NO_x/NH₃ in Figure 1(a) consists of a three-layer feed-forward structure with 6 input neurons, 16 hidden layer neurons, and 2 output neurons. A peak test error in Figure 1(b) of less than 5% was achieved for decoding the signals from NO_x/NH₃ mixtures in the 50-200 ppm range. For practical use in automotive systems, stability on the order of 10³ hours must be studied. Long term signal drift and changes in impedance response were monitored under repeated exposure to 50-200 ppm of NO_x, C₃H₈, and NH₃ in alternating biased and unbiased conditions at 530^oC over a period of over 100 days.

References

[1] G. Busca, L. Lietti, G. Ramis, F. Berti, Appl. Catal. B Environ., 18, 1–36 (1998).

[2] M. Koebel, M. Elsener, M. Kleemann, Catal. Today, 59, 335–345 (2000).

[3] L. Tsui, A.D. Benavidez, P. Palanisamy, L. Evans, F.H. Garzon, ECS Trans., 75, 9–22 (2016).

Figure 1. (a) Artificial neural network model used to decode the signals from a three-electrode sensor operated in open circuit and unbiased mode. (b) Test error for the ANN shows a peak at 2.5% and 90% of the data points are confined to less than 15% error.