Sensing of Soot Oxidation by Using Combustion-Type Sensor with Ag-Supported Catalyst

Soot-sensing properties of the combustion-type sensor coated with Ag-supported catalysts (Ag/TiO2, Ag/α-Al2O3, and Ag/CeO2) were investigated. The catalyst with a high carbon black (CB) oxidation activity in the tight contact (TC) mode had a high response speed, although the response speed had no relationship with the CB oxidation activity in the loose contact (LC) mode. Moreover, the difference in CB oxidation activities between TC and LC modes was consistent with the difference in response speed between V10 and V50, where V10 and V50 indicate the response speed of the CB oxidation at 10 and 50% of total output voltage. The catalyst with high intrinsic CB combustion activity improves the response property at initial combustion. On the other hand, the catalyst with the CB combustion activity, which does not depend on the contact state between the CB and the catalyst, improves the response property at later combustion.


Introduction
Particulate matter (PM), which mainly consists of soot and soluble organic fraction (SOF) from diesel engines, is one of the main pollutants causing the PM2.5 problem, along with other serious environmental and health problems. (1)(2)(3) Generally, improving fuel quality, boosting engine performance, and developing diesel exhaust after-treatment systems are the main ways of reducing PM emissions from diesel vehicles. Research on diesel exhaust after-treatment is a promising solution. The catalysed diesel particulate filter (C-DPF) is commonly regarded as the most effective technology for reducing PM emissions from diesel engines, where the PM is trapped and oxidized using a catalyst at low temperatures. (4)(5)(6) Therefore, to guarantee the efficient and stable operation of a diesel engine system with C-DPF after-treatment, on-board PM monitoring technology is required. (3,7,8) For this reason, several soot detection technologies have been proposed by researchers in the last few decades. Potentiometric and amperometric soot sensors based on an oxide ion conductor (zirconia) and electrical insulator were proposed by Vogel and colleagues (8,9) and Moss and colleagues, (10)(11)(12) and these can be used for in situ measurements of soot concentration. Sensors based on the electromagnetic wave method, which includes radio-frequency (RF) (13,14) and optical waves in terms of changing the light transmission, scattering or extinction, have also been used to monitor PM concentration. (15)(16)(17)(18) In addition, a series of PM sensors dependent on polarization resistance resulting from a four-electron electrochemical reaction (C + H 2 O → CO 2 + 4H + + 4e − ) by utilizing a solid electrolyte as well as an electrocatalyst were proposed by Hibino and colleagues. (7,(19)(20)(21) Ag is well known as an efficient partial oxidation catalyst. For instance, Ag has been used industrially for the epoxidation of ethylene (22,23) and the oxdehydrogenation of methanol to formaldehyde. (24)(25)(26) In addition, Ag is well-used for NO x abatement, (27,28) and the oxidation of ammonia, (29) methane, (30) carbon monoxide, (31) organic volatile compounds. (32) Ag with support materials such as ZrO 2 , TiO 2 , Al 2 O 3 and CeO 2 also showed remarkable performance for diesel soot oxidation. (33)(34)(35)(36) In this study, we evaluate the soot oxidation activities of Ag-supported catalysts (Ag/TiO 2 , Ag/ α-Al 2 O 3 , and Ag/CeO 2 ) and discuss the soot sensing properties of the catalytic combustion-type sensor, which converts oxidation heat into an electric signal, with these catalysts. Additionally, we evaluate the relationship between the sensing properties and the soot oxidation activities of the catalysts.

Catalysts for soot oxidation
TiO 2 [JRC-TiO-4(2), Catalysis Society of Japan], α-Al 2 O 3 (Taimei Chemicals), and CeO 2 (JRC-CEO-2, Catalysis Society of Japan) were used as the catalyst supports. To change the particle size, CeO 2 was calcined at 500 and 1300 °C. Ag-loaded catalysts were prepared by ordinary impregnation of the catalyst support material with an aqueous solution of silver nitrate [AgNO 3 (Kishida Chemical)]. The Ag content in the catalyst was fixed at 4.5 wt% on the basis of our previous research. (36) The suspended solutions were evaporated to dryness under stirring at 350 °C and then baked at 350 °C for 2 h, followed by calcination at 500 °C for 5 h in air.
The Brunauer, Emmett, and Teller (BET) specific surface areas of the catalysts were measured by nitrogen adsorption at 77 K (BELSORP-mini gas adsorption analyzer, BEL Japan, Inc.). Thermogravimetry (TG) and differential thermal analysis (DTA) measurements (TG/DTA 7300, HITACHI) were carried out to evaluate the soot oxidation activity of the catalysts at a heating rate of 10 °C min −1 from 30 to 800 °C in synthetic air (21 vol.% O 2 and 79 vol.% N 2 , 100 ml min −1 ) after the pre-heat treatment at 120 °C for 2 h. In this study, commercially available carbon black (CB, Sigma-Aldrich), which has a particle size of 5 μm, was used as a substitute for soot. All the catalysts and CB were mixed by grinding together in a mortar with pestle and spatula for 10 min, denoted as tight contact (TC) mode and loose contact (LC) mode, respectively.

Fabrication and evaluation of sensor
The sensor element and fabrication of the sensing device are shown in Fig. 1. Sensor elements, a combination of a Pt coil and γ-Al 2 O 3 bead, were coated with a slurry of Ag/TiO 2 , Ag/α-Al 2 O 3 , and Ag/CeO 2 catalysts, respectively. They were then heated at 600 °C for 2 h in air by using an internal Pt heater of the sensor element to remove the binder resin. After that, the CB slurry was coated on the catalyst-coated sensor element, followed by heating at 200 °C for 2 h in air. The catalyst slurry was prepared by mixing catalysts and terpineol (C 10 H 18 O, Wako) at a weight ratio of 9:1 for 3 min using a mixer (AR-100, Thinky) for 3 min. On the other hand, the CB slurry was prepared by mixing CB and terpineol at a weight ratio of 12:1.
The sensing principle and measurement system used in this study were the same as those described in previous research. (3) A Wheatstone bridge circuit, which includes a sensor element coated with the catalyst and a reference element without the catalyst, was employed. The temperature dependence of output voltage for CB oxidation was measured from 50 to 600 °C at intervals of 50 °C in synthetic air at a gas flow rate of 50 ml min −1 . The temperature of the sensor element was controlled and kept constant for 3 min at each measurement temperature by the Pt heater. On the other hand, the response property was investigated at a constant temperature of 600 °C by changing the gas from N 2 to air.

CB oxidation activity of Ag/TiO 2 , Ag/α-Al 2 O 3 , and Ag/CeO 2 catalysts
The TG-DTA curves for CB oxidation of Ag/TiO 2 , Ag/α-Al 2 O 3 , and Ag/CeO 2 catalysts in TC and LC modes are shown in Fig. 2. Table 1 shows the specific surface area and CB oxidation activities in TC and LC modes. T ig and T max indicate the starting temperature of weight loss in the TG curve and the peak temperature of the DTA curve, respectively.
In the TC mode, the Ag/TiO 2 catalyst showed higher CB oxidation activity, which means lower oxidation temperatures of T ig and T max , than the Ag/α-Al 2 O 3 and Ag/CeO 2 catalysts. Although both Ag/α-Al 2 O 3 and Ag/CeO 2 catalysts showed the same T ig , T max with the Ag/α-Al 2 O 3 catalyst was lower than that with the Ag/CeO 2 catalyst and had a sharp DTA peak, which means quick combustion of CB. In the LC mode, both Ag/α-Al 2 O 3 and Ag/CeO 2 catalysts showed the same T ig and T max , and this suggests that the CB oxidation activities of these catalysts are not different in the state in which CB and the catalysts contacted weakly. Furthermore, the CB oxidation activity of the Ag/TiO 2 catalyst was low and T max was higher than those of the Ag/α-Al 2 O 3 and Ag/CeO 2 catalysts.
The Ag/TiO 2 catalyst has an intrinsically high activity for CB oxidation and showed a large difference in catalytic activity between the TC and LC modes. On the other hand, the Ag/CeO 2 catalyst is different from the other two catalysts and showed similar CB oxidation activity between the TC and LC modes.   Figure 3 shows the temperature dependence of output voltage for CB oxidation using the sensor elements coated with Ag/TiO 2 , Ag/α-Al 2 O 3 , and Ag/CeO 2 catalysts. Here, the increase in output voltage corresponds to CB combustion with different catalysts. For the sensor element with the Ag/TiO 2 catalyst [ Fig. 3(a)], the output voltage began to increase from 300 °C and then increased    Fig. 3, is shown in Fig. 4(a) as a function of temperature. The ∆V corresponds to the combustion heat of CB, which indicates the amount of combusted CB on the element. The ∆V of the sensor element with the Ag/ TiO 2 catalyst began to increase at 300 °C and increased rapidly at over 400 °C until most of the CB was burned off. At 600 °C, the ∆V of the sensor element with the Ag/TiO 2 catalyst decreased because there was little residual CB. In these cases, the contact condition between the catalyst layer and the CB is indicated by the schematic in Fig. 4(b). First, some CB that had a tight contact with the catalyst (such as the TC mode) burned between 300 and 400 °C, as shown in Fig. 4(b-A). At over 400 °C, the CB that had a loose contact with the catalyst (such as the LC mode) burned, as shown in Fig. 4(b-B). Then, at 600 °C, a small amount of residual CB, which had hardly any contact with the catalyst (much looser than the LC mode), burned [ Fig. 4

Temperature dependence of output voltage for CB oxidation using sensor elements
The ∆V of the sensor element with the Ag/CeO 2 catalyst began to increase at 400 °C and kept increasing rapidly until 600 °C. It is considered that CB, which had a loose contact with the catalyst [ Fig.  4(b-B)], burned immediately after the combustion of CB, which had a tight contact with the catalyst [ Fig. 4(b-A)]. On the other hand, the ∆V of the sensor element with the Ag/α-Al 2 O 3 catalyst hardly changed at the temperature of 450 to 550 °C and increased rapidly at 600 °C. It is considered that much of the CB on the sensor element burned at approximately 600 °C, which is close to the temperature of the natural combustion, although a small amount of CB, which had a tight contact with the catalyst [ Fig. 4(b-A)], burned.
These results suggest that the contact state is different between the catalyst and the CB, which can be measured according to different sensor signals (output voltage). Moreover, it is clear that the contact state between the catalyst and the CB has a large effect on the sensing performance.

Sensing property in exchanged gas, N 2 to air, under constant temperature
The sensing properties of the sensor elements with Ag/TiO 2 , Ag/CeO 2 , and Ag/α-Al 2 O 3 catalysts under constant temperature are shown in Fig. 5. The sensor element coated with the catalyst and the CB was kept in N 2 for 5 min at 600 °C and then the gas was changed to air at the same temperature. It is clear that almost all the CB coated with different catalysts had burned after changing N 2 to air. Here, the peak output voltage after changing N 2 to air is caused by the CB combustion in the TC mode, indicating that the combustion heat of CB conducted through the sensor element to the Pt coil. The increase in output voltage after that peak is caused by CB combustion in the LC mode and indicates the decrease in the amount of residual CB with high heat capacity because the combustion heat hardly conducts to the Pt coil. In Fig. 5(a), CB coated on Ag/TiO 2 was oxidized suddenly after changing N 2 to air, and the difference in output voltage (∆V, ∆V = V air -V N , V air is the output voltage in air and V N is the output voltage in N 2 ) was about 10 mV during the first 5 min, and the contact condition between the catalyst and the CB was near the TC mode [ Fig. 4(b-A)]. The output voltage continued to increase with time, and the contact condition between the catalyst and the CB is near the LC mode [ Fig. 4(b-B)]. The residual CB continued to be burned while the output voltage increased slightly until all the CB burned out. Figure 5(b) shows the sensing property of the sensor element coated with both Ag/ α-Al 2 O 3 and CB. The output voltage returned to the same level as in N 2 following a sharp oxidation peak after changing N 2 to air and then increased slowly until all the CB was consumed, thereby showing a fast combustion speed, because most of the CB was oxidized during the first 5 min. The sensing property of the sensor element coated with both Ag/CeO 2 and CB is shown in Fig. 5(c), and its output voltage gradually increased without the peak after changing N 2 to air, which was different from that of the other two sensors.
The difference in sensing properties among these sensor elements coated with different catalysts is mainly due to the different internal properties of the catalysts. To understand the relationship between the sensing properties and the oxidation performance of the CB, we evaluated the response speed of the CB oxidation. The response speed of the CB oxidation at 10 and 50% of the total output voltage was denoted as V 10 and V 50 , respectively, as shown in Fig. 6. The total output voltage ∆V is the difference between the output voltage at which all the CB was oxidized and the output voltage in N 2 before oxidation. Figure 7 shows the comparison of the response speeds for CB oxidation estimated from the equation described above. The response speeds V 10 of CB oxidation were 50, 34, and 21 mV·min −1 for sensor elements with Ag/TiO 2 , Ag/α-Al 2 O 3 , and Ag/CeO 2 catalysts, respectively. The peak area of the output voltage after changing N 2 to air corresponds to the amount of combustion heat (∆H). The amplitude of the estimated ∆H from Fig. 6 became large in the order of the sensor elements with Ag/TiO 2 , Ag/α-Al 2 O 3 , and Ag/CeO 2 catalysts. Therefore, this suggests that the catalyst with a high CB oxidation activity has a fast response speed at an early stage. On the other hand, the response speeds V 50 of CB oxidation were 1.6, 0.7, and 17.7 mV·min −1 for sensor elements with Ag/ TiO 2 , Ag/α-Al 2 O 3 , and Ag/CeO 2 catalysts, respectively. This suggests that the V 50 of CB oxidation has no relation with the oxidation heat. As shown in Table 1, it is clear that the catalyst with better catalytic activity in T max in the TC mode had a higher response speed V 10 for CB oxidation. However, the V 50 for CB oxidation has no relationship with the catalytic activity in T max in the LC mode. Moreover, although the response curve of the sensor element with the Ag/CeO 2 catalyst increased gradually without an obvious peak after changing the gas (from N 2 to air), its V 50 was highest among the sensor elements with these catalysts.
In the case of the sensor element with the Ag/CeO 2 catalyst, the difference between V 10 and V 50 was very small compared with those of the sensor elements with Ag/TiO 2 and Ag/α-Al 2 O 3 catalysts. In addition, the difference in the CB oxidation activities of the catalysts (∆T max in Table 1) between the TC and LC modes was also consistent with the difference in the response speed between V 10 and V 50 (∆V 50-10 ). For example, the difference in the response speed, ∆V 50-10 , of the CB oxidation for the sensor element with the Ag/CeO 2 catalysts, which showed the smallest ∆T max (8 °C in Table 1) for the CB combustion, was only 3.1 mV·min −1 , which was much smaller than 48.5 and 33.5 mV·min −1 for the sensor elements with the Ag/TiO 2 and Ag/α-Al 2 O 3 catalysts, respectively.
These results suggest that the catalyst with the high intrinsic CB oxidation activity (T max in TC mode) improves the response property (V 10 ) at an early stage. On the other hand, the catalyst that does not depend on the contact state (TC and LC modes) improves the response property of the CB oxidation at the later stage [ Figs. 4(b-B) and 4(c-C)]. In this study, the best catalyst for the soot sensing is Ag/CeO 2 , which showed almost the same response speed of V 10 and V 50 . We suggest that the close response speed of V 10 and V 50 is most important in the real condition, regardless of the active species and the support materials. The contact state between the catalyst and the soot is quite loose under the practical condition. Therefore, the response speed of V 50 will become important for the catalyst used in the combustion-type PM sensor.

Conclusion
The soot-sensing properties of the combustion-type sensor coated with Ag-supported catalysts (Ag/TiO 2 , Ag/α-Al 2 O 3 , and Ag/CeO 2 ) were investigated. The catalyst with an intrinsically high CB oxidation activity had a high response speed although the response speed had no relationship with the CB oxidation activity, regardless of the contact state between the catalyst and the CB. Moreover, the difference in CB oxidation activity between the TC and LC modes was consistent with the difference in the response speed between V 10 and V 50 . The catalyst with high intrinsic CB combustion activity improves the response property at an initial combustion. On the other hand, the catalyst having CB combustion activity which does not depend on the contact state between the CB and the catalyst, improves the response property at later combustion.
In this study, we demonstrated not only the preliminary result of the detection for the soot oxidation under operation temperature but also indicated a catalyst design for soot oxidation. However, because most of our research in this work was focused on qualitative analysis, a more quantitative evaluation of the sensing property and measurements under real exhaust conditions isnecessary in the future.