Capacitive Soot Sensor

Available online at www.sciencedirect.com

ScienceDirect Procedia Engineering 120 (2015) 241 – 244

EUROSENSORS 2015

Capacitive Soot Sensor G. Hagena*, G. Rießa, M. Schuberta, M. Feulnera, A. Müllera,b, D. Brüggemannb, R. Moosa Bayreuth Engine Research Center (BERC), Zentrum für Energietechnik (ZET), University of Bayreuth, 95440 Bayreuth, Germany a Department of Functional Materials, b Department of Engineering Thermodynamics and Transport Processes

Abstract A capacitive soot sensor was built up on basis of a known resistive device by covering the electrode area with a thin but dense alumina layer manufactured by the aerosol-deposition-method (ADM). Deposited soot cannot short-circuit the electrodes but leads to an increasing capacitance of the sensor. Experiments were conducted in real exhaust. Increasing soot concentrations lead to a more fast and sharp increase of the capacitance and the blind time of the sensor decreased to almost zero. The performance of such sensors could be enhanced when a voltage is applied at the electrodes during soot collection. © by Elsevier Ltd. This an open Ltd. access article under the CC BY-NC-ND license © 2015 2015Published The Authors. Published by isElsevier (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of EUROSENSORS 2015. Peer-review under responsibility of the organizing committee of EUROSENSORS 2015

Keywords: Aerosol-deposition-method (ADM); interdigital electrodes; real exhaust test; dosimeter-type sensor; on-board diagnostics (OBD)

1. Motivation To meet the stringent emission limits for particulate matter (PM) in automotive exhausts, ceramic wall-flow filters are in serial use for diesel-fueled vehicles (Diesel Particulate Filter, DPF). Trapped soot has to be burned off from time to time to avoid clogging, but filter regeneration at high exhaust gas temperatures is fuel consuming and should be carried out only when necessary. Therefore, appropriate techniques for an efficient operation are needed [1]. As state-of-the-art, the filter loading is estimated from a differential pressure sensor signal going along with model data to start regeneration of the filter. The efficiency of such aftertreatment systems depend on the accuracy of the information about the real filter loading [2]. A more precise insight into the filter loading state may also be

* Gunter Hagen. Tel.: +49-921-55-7406; fax: +49-921-55-7405. E-mail address: [email protected]

1877-7058 © 2015 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of EUROSENSORS 2015

doi:10.1016/j.proeng.2015.08.590

242

G. Hagen et al. / Procedia Engineering 120 (2015) 241 – 244

possible by direct in-situ microwave-based methods [3,4,5]. Furthermore, it is required to monitor the correct functionality of all emission-relevant parts of the exhaust gas aftertreatment system during operation. For this socalled on-board diagnostics (OBD), soot sensors are applied downstream of a DPF. In most applications, resistive (also called “conductometric”) sensors are in use. When soot deposits on its electrode structure (where a voltage is applied), conductive paths are formed, leading to a current increase. Evaluation of the time until this threshold is reached (blind time) gives a measure on the amount of soot in the gas flow and indicates a filter failure [6,7,8,9,10]. In former studies, we showed that an application upstream of a DPF could basically enable soot mass detection [11]. Here, we evaluated the slope of the current increase for each loading cycle as the sensor signal. In general, increasing the sensors sensitivity is of high interest, especially the reduction of the sensors blind time. One possibility on that topic is described in [12]. By covering the electrodes with a thin conductive layer, the initiation time for detecting soot was reduced by about 40 % compared to an uncoated sensor. Another approach for selectivity enhancement is given in [13]. Here, additional electrodes are operated at higher voltage for soot collection. A special sensor design including a cavity enables measurements of low soot concentrations by the capacitive principle. In comparison to the resistive principle a smaller temperature dependency is expected. In the present contribution, we evaluated the potential of a planar capacitive soot sensor. It bases on a formerly used resistive sensor by covering the electrodes with a thin alumina layer, which was deposited by the novel room temperature impact consolidation process (RTIC), also known as aerosol deposition method (ADM). 2. Experimental Basic sensor devices were manufactured as follows: Pt-IDE electrodes (line = space = 100 μm) were screenprinted on the front side of an alumina substrate, comprising a thick-film heater (also Pt, LPA88-S, Heraeus) on the reverse side. Electrical feed lines are covered by glass-ceramic dielectrics (QM42, DuPont). Normally, i.e. for resistive measurements, the electrode area would remain uncovered. To obtain a capacitive signal, the electrodes were insulated from each other to avoid an electrical short-circuit by soot deposition. The insulation must be hightemperature stable, dense and preferably thin. For that purpose, we applied an aerosol-deposited film (Fig. 1). By this novel method, dense ceramic layers can be deposited directly from the ceramic powder onto nearly any kind of substrate material in a totally cold process, i.e. without a sintering step [14,15].

Fig. 1. (a) Sketch of the capacitive sensor setup with alumina cover layer on top of the IDE electrodes; (b) Schematic cross-section of the sensor setup for the capacitive sensor principle (soot deposition on the sensor surface should lead to an increasing capacitance due to the change in the electrical field distribution); (c) Photograph of the real sensor (top view) in the IDE area.

In comparison to a resistive device where soot particles have to percolate between the electrodes before a current signal occurs, the blind time should decrease for a capacitive device. Deposited soot should change the electrical field distribution even before the percolation threshold is reached (Fig. 1b). Even non-percolating (low) soot amounts should contribute to the capacitance increase. All measurements were conducted in real exhaust. The sensor was mounted (with its front side facing the gas flow) in the exhaust pipe of a 2.1 l diesel engine that was operated at 25 % load / 1000 rpm. Different soot concentrations in the exhaust were achieved by changing the boost pressure at a constant injection pressure of

243

G. Hagen et al. / Procedia Engineering 120 (2015) 241 – 244

550 bar. Capacitance measurements were done by an LCR-Meter (HP 4284A) at a constant frequency of 10 kHz in the R||C mode, i.e. the measured impedance is interpreted as an equivalent circuit model with R and C in parallel. After a certain time span for soot loading, the sensor was regenerated by adjusting the reverse side heater temperature to 600 °C, to burn off the deposited soot. After switching off the heater, a new cycle started. 3. Results and Discussion In a first experiment, sensor data were taken continuously during engine operation. By decreasing the boost pressure, associated changes in some engine parameters occurred. So, the exhaust gas temperature increased slightly and the exhaust mass flow decreased a little. The air-fuel ratio O decreased therefore, going along with a strong increase in particle mass and number values, which were recorded by a Pegasor sensor over time (Table 1). Table 1. Average values of engine parameters and soot concentrations for different operating points Boost pressure

Exhaust mass flow

Gas temperature

Particle mass

Particle number

3

6

3

O

1.22 bar (OP1)

80 kg/h

271 °C

9 mg/m

34.2 x 10 /m

1.38

1.17 bar (OP2)

78.5 kg/h

294 °C

22 mg/m3

84.8 x 106 /m3

1.32

1.12 bar (OP3)

74.5 kg/h

313 °C

39 mg/m3

152.4 x 106 /m3

1.25

Regarding the sensor signal (Fig. 3), one can observe the expected increase of capacitance during soot loading. The slope of the capacitance change increases with higher soot concentrations. In each cycle, the capacitance saturates at about 40 to 50 pF (which is an increase by a factor of 3 to 4 compared to the unloaded value). This final value is slightly affected by the soot concentration and might be explained by temperature effects. After each loading cycle, a sensor regeneration was initiated. The sensor behavior during that heating phase should be handled with care because the R||C-model for data evaluation does not fit anymore as the alumina substrate as well as the cover layer gets electrically conductive [16]. One could see soot oxidation in the first seconds when the sensor signal jumps up and then decreases again. Oxidation is finished when the sensor signal reaches a constant value of about 20 pF. This behavior is known from resistive type sensors [7] and follows the dosimeter principle [17]. After switching off the heater, the sensor cools down to exhaust gas temperature and its signal comes back to the constant capacitance of about 11 pF which is measured during the blind phase after each regeneration procedure. This time span until a signal change occurs is significantly reduced with increasing soot. 180

OP1

160

OP2

OP3

C / pF

140 120

indicates sensor regeneration

100 80 60

blind phase

40 20 0 500

1500

2500

3500

4500

t / sec Fig. 3. Measurement of a capacitive sensor in real exhaust; the amount of soot increases by changing the boost pressure from 1.22 bar (OP1) to 1.17 bar (OP2) and 1.12 bar (OP3); detailed information in table 1

A further enhancement in the sensors sensitivity or a reduced blind time (i.e. a “faster” performance to measure lower soot concentrations) could be possible by electrophoretic soot collection. In a second experiment (during engine operation near to OP2), we applied a voltage of 34.5 V (dc) for a certain time, measured the sensors

244

G. Hagen et al. / Procedia Engineering 120 (2015) 241 – 244

capacitance, regenerated the sensor, and applied the voltage again for another time. Table 2 shows that the sensor response (increasing capacitance) starts significantly earlier with applied voltage during soot collection compared to the data measured without voltage at the electrodes. Table 2. Sensor capacitance C after soot loading for different time spans with or without applied voltage 0s

60 s

90 s

120 s

180 s

voltage of 34.5 V (dc) applied

11.4 pF

19.2 pF

26.2 pF

41.7 pF

42.9 pF

without voltage

10.8 pF

10.8 pF

11.0 pF

11.5 pF

22.4 pF

4. Conclusion and Outlook To form a capacitive soot sensor, the electrode area of a resistive soot sensor was covered with a thin but dense alumina layer utilizing the novel aerosol-deposition method. During soot deposition on the surface, the sensor capacitance increases. The higher the soot concentration, the lower is the blind time of the sensor and the higher is the slope of the capacitance increase. A further reduction of the sensor blind time is achieved by an applied voltage that effects electrophoretic soot collection. Even better results can be expected by an improved setup with reduced width of lines and spaces of the electrode structure to achieve more finger pairs in that area. A thinner alumina layer would increase the sensor effect as well. References [1] M. Twigg, P. Phillips, Cleaning the air we breathe - Controlling diesel particulate emissions from passenger cars, Platinum Metals Review 53 (2009) 27-34, doi: 10.1595/147106709X390977 [2] T. Johnson, Review of Diesel Emissions and Control, SAE Technical Paper (2010) 2010-01-0301, doi: 10.4271/2010-01-0301 [3] A. Sappok, L. Bromberg, Loading and Regeneration Analysis of a Diesel Particulate Filter with a Radio Frequency-Based Sensor, SAE Technical Paper (2010) 2010-01-2126, doi: 10.4271/2010-01-2126 [4] R. Moos, Microwave-based catalyst state diagnosis – state of the art and future perspective, SAE Technical Paper (2015) 2015-01-1042, doi: 10.4271/2015-01-1042 [5] M. Feulner, G. Hagen, A. Piontkowski, A. Müller, G. Fischerauer, D. Brüggemann, R. Moos, In-Operation Monitoring of the Soot Load of Diesel Particulate Filters - Initial Tests, Topics in Catalysis 56 (2013) 483-488, doi: 10.1007/s11244-013-0002-9 [6] T. Ochs, H. Schittenhelm, A. Genssle, B. Kamp, Particulate Matter Sensor for On Board Diagnostics (OBD) of Diesel Particulate Filters (DPF), SAE Technical Paper (2010) 2010-01-0307, doi: 10.4271/2010-01-0307 [7] G. Hagen, C. Feistkorn, S. Wiegärtner, A. Heinrich, D. Brüggemann, R. Moos, Conductometric Soot Sensor for Automotive Exhausts: Initial Studies, Sensors 10 (2010) 1589-1598, doi: 10.3390/s10030158 [8] H. Husted, G. Roth, S. Nelson, L. Hocken, G. Fulks, D. Racine, Sensing of Particulate Matter for On-Board Diagnosis of Particulate Filters, SAE Technical Paper (2012) 2012-01-0372 , doi: 10.4271/2012-01-0372 [9] B. Grob, J. Schmid, N.P. Ivleva, R. Niessner, Conductivity for Soot Sensing: Possibilities and Limitations, Analytical Chemistry 84 (2012) 3586-3592, doi: 10.1021/ac203152z [10] A. Malik, H. Abdulhamid, J. Pagels, J. Rissler, M. Lindskog, P. Nilsson, R. Bjorklund, P. Jozsa, J. Visser, A. Spetz, M. Sanati, A Potential Soot Mass Determination Method from Resistivity Measurement of Thermophoretically Deposited Soot, Aerosol Science and Technology 45 (2011) 284-294, doi: 10.1080/02786826.2010.533214 [11] G. Hagen, A. Müller, M. Feulner, A. Schott, C. Zöllner, D. Brüggemann, R. Moos, Determination of the soot mass by conductometric soot sensors, Procedia Engineering 7 (2014) 244-247, doi: 10.1016/j.proeng.2014.11.64 [12] P. Bartscherer, R. Moos, Improvement of the sensitivity of a conductometric soot sensor by adding a conductive cover layer, Journal of Sensors and Sensor Systems 2 (2013) 95-102, doi: 10.5194/jsss-2-95-2013 [13] A. Kondo, S. Yokoi, T. Sakurai, S. Nishikawa, T. Egami, M. Tokuda, T. Sakuma, New Particulate Matter Sensor for On Board Diagnosis, SAE Technical Paper (2011) 2011-01-0302 [14] M. Schubert, J. Exner, R. Moos, Influence of Carrier Gas Composition on the Stress of Al2O3 Coatings Prepared by the Aerosol Deposition Method, Materials 7 (2014) 5633-5642 [15] J. Exner, M. Hahn, M. Schubert, D. Hanft, P. Fuierer, R. Moos, Powder requirements for aerosol deposition of alumina films, Advanced Powder Technology, in press, doi: 10.1016/j.apt.2015.05.016 [16] J. Kita, A. Engelbrecht, F. Schubert, A. Groß, F. Rettig, R. Moos, Some practical points to consider with respect to thermal conductivity and electrical resistivity of ceramic substrates for high-temperature gas sensors, Sensors and Actuators B: Chemical 213 (2015) 541-546, doi: 10.1016/j.snb.2015.01.04 [16] I. Marr, A. Groß, R. Moos, Overview on Conductometric Solid-State Gas Dosimeters, Journal of Sensors and Sensor Systems 3 (2014) 2946, doi: 10.5194/jsss-3-29-2014