Radio Frequency Studies of Soot Loading and Ammonia Storage on a Diesel Particulate Filter with a SCR Catalyst Coating

Abstract

Introduction In the exhaust system of some diesel vehicles, the diesel particulate filter (DPF) is loaded with a selective catalyst reduction (SCR) wash coat. This combined catalyst, referred to as a SDPF, serves two functions by trapping soot and by reducing NOx emissions. The latter is accomplished via reactions of NOx with NH3 stored on the SCR wash coat. The NH3 is dosed, as urea, upstream of the SDPF. The radio frequency (RF) measurement technique has been shown to be a viable method for measurement of the amount of soot loaded onto DPF’s [1]. Additional studies have demonstrated the technique’s utility in determining the amount of NH3 stored on SCR catalysts [2]. In one configuration of the method, a microwave resonant cavity is formed by the metal can encasing the enclosed catalyst. Two metal probes acting as antennas are placed on either side of the catalyst. Power transmitted between the antennas is monitored as the frequency of the signal is swept. At certain frequencies, determined by the cavity’s geometry and the dielectric properties of the catalyst, resonance is achieved. Measurements of the resonance frequencies, the signal amplitude at resonance, and the mean amplitude over a frequency range give information on the dielectric properties of the catalyst. These properties change upon soot and NH3 loading. In this work we show how the RF response is influenced by both the soot and NH3 loading on the SDPF. We demonstrate the comparative responses for each and show how the soot loading can be determined with little influence by the stored NH3. We also demonstrate how for the case of “light” soot loading, before the attenuation of the signal by the soot dominates the RF response, simultaneous measurements of both the soot and NH3 can be achieved. Method The SDPF used in our measurement was 4.4 cm diameter by 17.4 cm in length, cored from a larger piece used in serial application. It was mounted tightly near the center of a stainless cavity, 31 cm long. An antenna was positioned between each face of the SDPF and a stainless screen defining the cavity’s ends. Before RF measurements were done, the empty SDPF was loaded with soot at room temperature using a benchtop Jing miniCAST 5201C soot generator. The mean mobility diameter of the generated soot was approximately 100 nm. The output of the soot generator was diluted with N2, and a flow of 90 l/min was used for loading. The measured soot concentration was near 10 mg/m3. Seven different soot loading values were used, with the maximum 1.37 grams corresponding to 5.2 g/l soot on the SDPF. After each of these loadings, the cavity containing the sooted SDPF was connected to a 60 liter/min heated gas-flow bench, with temperature probes placed at the ends of the cavity. The RF S21 transmission parameter, the power received at the 2nd antenna with respect to the 1st, was measured in lean gas up to 500 °C using an Agilent E5071C network analyzer. Measurements were then done at 250 °C, before and after the SDPF was saturated with NH3 (0.4 g on average) using a 400 ppm input concentration. Afterwards, the SPDF temperature was raised to 650°C to desorb the NH3 and burn off the soot, enabling a subsequent soot loading on an empty catalyst. Results and Conclusions Figure 1 shows the RF spectra over the frequency range 2.7 GHz – 4.7 GHz. Shown are the spectra for seven different soot loading levels, all measured at a mean SDPF temperature of 250 °C and with it empty of NH3. Multiple resonant peaks are seen for the case of 0 g soot loading. For the cases of higher soot loading, the amplitude of the resonant peaks diminishes, and the “baseline” of the signal is reduced. All evidence of the resonant peaks disappears for soot loadings of 0.41 g and greater. Figure 2a shows the influences on one of the resonant peaks measured at 250 °C on both “light” soot loading, up to 0.18 g, and on NH3 loading, measured after saturating the SDPF with 400 ppm NH3. The measured resonance frequency and the signal amplitude at resonance both decrease with soot and NH3 loading. This is also shown in Figs. 2b and 2c. Results for other resonant peaks are qualitatively similar. Using the curves in Figs. 2b and 2c, simultaneous estimation of both the soot and NH3 can be accomplished. This is only possible for cases of “light” soot loading, when the resonance peaks can still be observed. The resonance spectra for the cases of 0.41 g of soot and greater show little influence on NH3 loading. Figure 3 demonstrates that the soot loading can be determined over the full range investigated with little influence by the stored NH3. For this we measure the mean amplitude either in a narrow frequency window away from resonant peaks, or as done in this figure using a wide window averaging over many resonant peaks. References [1] H. Nanjundaswamy, V. Nagaraju, Y. Wu, E. Koehler, A. Sappok, P. Ragaller, L. Bromberg, Advanced RF Particulate Filter Sensing and Controls for Efficient Aftertreatment Management and Reduced Fuel Consumption, SAE Technical Paper 2015-01-0996, 2015, doi:10.4271/2015-01-0996. [2] D. Rauch, D. Kubinski, U. Simon, R. Moos, Detection of the Ammonia Loading of a Cu Chabazite SCR Catalyst by a Radio Frequency-Based Method, Sens. and Act. B: Chemical 205 (2014) 88-93. Figure 1