Effects of Base Materials (α-Alumina and/or γ-Alumina) on Volatile Organic Compounds (VOCs)-Sensing Properties of Adsorption/Combustion-Type Microsensors

Abstract

The sensing properties of adsorption/combustion-type microsensors using 5 wt% Pt-loaded aluminas, which consist of two kinds of alumina (α-Al₂O₃ and γ-Al₂O₃), as sensing (catalytic) materials for ethanol and toluene, were investigated in air, and the mixing effects of α-Al₂O₃ with γ-Al₂O₃ on the dynamic and static responses of the sensors were discussed in this study. The mixing of 50 wt% α-Al₂O₃ with γ-Al₂O₃ was the most effective in enhancing the dynamic responses to ethanol, which originated from the flash combustion behavior of ethanol and/or their partially decomposed products adsorbed on the sensing films from 150 °C to 450 °C, while further mixing of α-Al₂O₃ with γ-Al₂O₃ tended to increase the dynamic responses to toluene. On the other hand, the static responses to both ethanol and toluene, which arise from their catalytic combustion at elevated temperatures (450 °C), mainly increased with an increase in the addition of α-Al₂O₃ in the 5 wt% Pt-loaded aluminas. These results indicate that the synergistic effects of the catalytic activity and the thermal conductivity of the 5 wt% Pt-loaded aluminas are the most important for the sensing properties of these sensors to ethanol and toluene.

1. Introduction

The large number of volatile organic compounds (VOCs) generated from various building materials, such as paints, adhesives, and resins, has been a serious problem for a long time, and it is widely recognized that these substances have a variety of adverse effects on human health [1,2,3]. In order to reduce these symptoms, so-called “sick house syndrome” and “chemical sensitivity”, it is essential to develop specific VOC-sensing devices that can detect VOCs in indoor environments with high sensitivity. On the other hand, various kinds of VOCs are also contained in human biogases, such as exhaled breath and skin gas, and monitoring changes in their concentrations can help determine the health status and presence or absence of disease [4,5,6]. For example, acetone has been reported to be a biomarker for diabetes and metabolic activity, isoprene and pentane for heart disease, and toluene and 1-nonanal for lung cancer [7,8,9,10,11], and the development of a wearable device that can sensitively and selectively detect these VOCs is significantly attractive in the healthcare and medical fields. Recently, various gas sensors have been developed for the sensitive and selective detection of different types of VOCs [12,13,14,15,16,17,18,19,20,21,22,23,24]. Our group has developed adsorption/combustion-type gas sensors that combine the functions of “VOC adsorption” and “flash VOC combustion” for many years, by dynamically operating kinds of catalytic combustion-type gas sensors, in order to detect such ultra-trace amounts of VOCs [25,26]. After VOCs are adsorbed on the surface of the catalytic materials of the sensors at low temperatures, they are burned by flash heating and the large amount of combustion heat generated becomes output as a sensor signal. We previously reported that the VOC response of adsorption/combustion-type gas sensors was drastically improved by impregnating Pd and Au nanoparticles in an optimal ratio on γ-alumina (γ-Al₂O₃) or by highly dispersing core/shell (Au/Pd) nanoparticles by ultrasonic reduction on γ-Al₂O₃ [27,28]. We also confirmed that the VOC response was improved by co-loading Pt and a metal oxide on γ-Al₂O₃ and that the VOC-sensing characteristics were highly dependent on the type of metal oxide [29,30]. In the process of clarifying the VOC-response mechanism of adsorption/combustion-type gas sensors while closely investigating the VOC oxidation activity and adsorption/desorption characteristics, it was suggested that the thermal conductivity of the catalytic material also had a significant effect on the VOC-response characteristics [31]. In fact, our previous research revealed that the VOC-response characteristics of the sensor were improved when α-alumina (α-Al₂O₃), which has a small specific surface area but high crystallinity, was mixed with γ-Al₂O₃ and that they were loaded with Pd and Au [26]. In this case, it was inferred that the higher thermal conductivity of α-Al₂O₃ compared with γ-Al₂O₃ contributed to the improved VOC-response characteristics. In this study, therefore, we focused on Pt nanoparticles as a noble-metal catalyst for efficient VOC combustion and aimed to clarify the effect of mixing α-Al₂O₃ with γ-Al₂O₃ as the base material of the sensing film on the VOC-sensing properties of adsorption/combustion-type microsensors for two kinds of VOCs, ethanol and toluene.

2. Experimental Section

2.1. Preparation of Sensing Materials and Characterization

γ-Al₂O₃ powder (JGC Catalysts and Chemicals Ltd., Kanagawa, Japan) and α-Al₂O₃ powder (TM-DA, Taimei Chemical Chemicals Co., Ltd., Tokyo, Japan) were mechanically mixed in an agate mortar, and then Pt particles were loaded onto their surfaces by general impregnation with hydrogen hexachloroplatinate(IV) hexahydrate (FUJIFILM Wako Pure Chem. Corp., Osaka, Japan), followed by heat treatment at 500 °C for 1 h. The obtained powder is denoted as nPt/γ(r)α(t)-Al₂O₃ [n: the amount of Pt loaded (wt%), r: the weight ratio of γ-Al₂O₃ to all aluminas (wt%), t: the weight ratio of α-Al₂O₃ to all aluminas (wt%)], and nPt/γ(100)α(0)-Al₂O₃ and nPt/γ(0)α(100)-Al₂O₃ are abbreviated and simply expressed as nPt/γ-Al₂O₃ and nPt/α-Al₂O₃, respectively. The pore size distribution and the specific surface area (SSA) of the obtained powders were measured by Barrett-Joyner-Halenda (BJH) and Brunauer-Emmett-Teller (BET) methods using nitrogen adsorption-desorption isotherms (TriStar 3000, Micromeritics Inst. Corp., Norcross, GA, USA), respectively [32,33,34]. The crystal phases of the powders were characterized by X-ray diffraction analysis (XRD; Minflex 600-DX, Rigaku Corp., Tokyo, Japan) using Cu Kα radiation (40 kV, 40 mA), and their crystallite sizes were calculated by using the Scherrer equation. The chemical states of Al, O, and Pt on the surface of the representative powders were characterized by X-ray photoelectron spectroscopy using Al Kα radiation (XPS; ACIS-TLATRA DLD, Kratos Analytical, Manchester, UK). The binding energy was calibrated using the C 1s level (285.0 eV) of the usual contamination.

The dispersibility, surface area, and average particle size of the Pt nanoparticles on the surface of typical 5Pt/γ(r)α(t)-Al₂O₃ powders were calculated using the CO-pulse method (BELCAT II, MicrotracBEL Corp., Osaka, Japan). The nanostructure of the typical 5Pt/γ(r)α(t)-Al₂O₃ powder was observed by transmission electron microscopy (TEM; JEM-ARM200F, JEOL Ltd., Tokyo, Japan), and scanning transmission electron microscopy with energy dispersive X-ray spectroscopy for elemental mapping (STEM-EDS; JEM-ARM200F and JED-2300T, JEOL Ltd., Tokyo, Japan). The thermal conductivity of representative nPt/γ(r)α(t)-Al₂O₃ powders was measured by using a modified transient plane source sensor (TRIDENT, C-Therm Technologies, Fredericton, NB, Canada). The nPt/γ(r)α(t)-Al₂O₃ powders were pelletized by compressing at 1000 kg for 8 min, followed by annealing at 500 °C for 2 h in ambient air, and the obtained disks with a diameter of ca. 2 cm and a thickness of ca. 1 mm were used for the thermal conductivity measurement. The heat treatment condition was the same as that of the sensor fabrication described in the following section. The catalytic combustion behavior of ethanol and toluene over typical 5Pt/γ(r)α(t)-Al₂O₃ powders was measured using a handmade catalytic activity evaluation system. The 5Pt/γ(r)α(t)-Al₂O₃ powders were pelletized by compressing at 1000 kg for 8 min, followed by annealing at 500 °C for 2 h in ambient air, and the obtained disks with a diameter of ca. 1 cm were crushed to obtain irregularly shaped agglomerates with a size of 250–850 μm. They were packed into the region of ca. 12 mm in width in the center of a fixed bed reactor with an inner diameter of 2.0 cm, and the oxidation activity of ethanol and toluene over the typical nPt/γ(r)α(t)-Al₂O₃ powders was measured in synthetic air at a flow rate of 20 cm³ min⁻¹ (space velocity: 8159 h⁻¹) and in compressed air at a flow rate of 20 cm³ min⁻¹ (space velocity: 40,795 h⁻¹).

2.2. Fabrication of Sensors and Measurement of Sensing Properties to Ethanol and Toluene

Figure 1a shows a stereomicroscope photograph of a microsensor platform with a couple of Pt microheaters, which was fabricated by MEMS technology, and Figure 1b shows a SEM photograph of a Pt microheater, which was covered with a thin alumina film, on a thin insulating substrate (ca. 70 × 70 μm²) with a diaphragm structure. The Pt microheater also served as a detector. The microsensor platform, which was recently downsized from the previous one [16,17,18,19,20,21,22,23], was designed by Yazaki Energy System Corp. The nPt/γ(r)α(t)-Al₂O₃ and unloaded γ(r)α(t)-Al₂O₃ powders were used as the sensing and reference materials, respectively. They were mixed with an appropriate amount of an organic vehicle consisting of polyvinyl butyral resin (mean polymerization degree: 700)], di-n-butyl phthalate, and terpineol by ball milling for 30 min. The obtained nPt/γ(r)α(t)-Al₂O₃ and unloaded γ(r)α(t)-Al₂O₃ pastes were applied over both the Pt microheaters, as a sensing film or a reference film, respectively, by drop coating employing an air-pulse fluid dispenser (MS-10DX, Musashi Eng., Inc., Tokyo, Japan) with a suitable syringe. The microsensor chips attached to these films were then heat-treated at 500 °C for 2 h in ambient air. The obtained sensors were referred to as sensing materials (nPt/γ(r)α(t)-Al₂O₃). Figure 1c shows the SEM photographs of a typical sensing 5Pt/γ(r)α(t)-Al₂O₃ film on a substrate with a Pt microheater. The thickness at the center of the film was approximately 30–40 μm, and the thickness gradually decreased toward the edge of the film. In addition, the surface of the film was relatively smooth, and the morphology of the film was formed with relatively good reproducibility.

2.3. Measurement of Sensing Properties to Ethanol and Toluene

The sensing properties of ethanol and toluene were measured in an acrylic chamber (internal volume: approximately 50 dm³) in which 10, 100, and 1000 ppm of ethanol or toluene were vaporized just on a compact ceramic heater. The nPt/γ(r)α(t)-Al₂O₃ sensors consist of sensing (nPt/γ(r)α(t)-Al₂O₃) and reference (γ(r)α(t)-Al₂O₃) films with Pt microheaters, and these films were assembled into a bridge circuit, as shown in Figure S1a. All sensors were consecutively pulse-heated during a cycle of 10 s from low temperature (LT, 150 °C, 9.6 s) to high temperature (HT, 450 °C, 0.4 s) in this study. A typical signal profile of the sensor is shown in Figure S1b. Ethanol, toluene, and their related components (intermediates), which are produced by partial oxidation and/or condensation, are adsorbed on the oxide surfaces of both films at 150 °C in the target gas. These adsorbates burn mainly on the oxide surface of the sensing film the moment the sensor is heated to 450 °C. At that time, the sensor-signal profile typically has one dynamic signal originating from the flash catalytic combustion of all the adsorbates mainly on the oxide surface of the sensing film. A static signal originating from the general catalytic combustion of ethanol or toluene mainly on the oxide surface of the sensing film is constantly confirmed after the instantaneous dynamic signal disappears during pulse heating at 450 °C. The magnitude of the general response (ΔV_MAX) was defined as the largest difference between the output voltage in air containing ethanol or toluene and that in air (ΔV). Two responses, which were calculated by integrating the dynamic and static signals of the sensor-signal profiles, were defined as the integrated dynamic response (IDR) and integrated static response (ISR), respectively (see Figure S1b).

3. Results and Discussion

3.1. Characterization of Sensor Materials

Figure S2 shows the nitrogen adsorption/desorption isotherms and pore size distributions of all the 5Pt/γ(r)α(t)-Al₂O₃ powders. Figure 2 shows the dependence of the SSA of the 5Pt/γ(r)α(t)-Al₂O₃ powders on the amount of α-Al₂O₃ (t). The amount of nitrogen adsorbed on the 5Pt/γ-Al₂O₃ powder was the largest among all the 5Pt/γ(r)α(t)-Al₂O₃ powders, and thus both the pore volume and the SSA (ca. 219 m² g⁻¹) of the 5Pt/γ-Al₂O₃ powder were also the largest among them. In addition, the large hysteresis of the nitrogen adsorption-desorption isotherms was clearly confirmed under a relative pressure of 0.6–1.0, indicating that they were classified mainly as type-IV isotherms in the models of IUPAC [32]. Thus, the mesopore distribution arising from nitrogen adsorption was significantly different from that of nitrogen desorption. The SSA monotonically decreased with an increase in the amount of α-Al₂O₃, as shown in Figure 2, and the ratio of the small mesopores (diameter: ≤ca. 5 nm) to large mesopores (diameter: ≥ca. 5 nm) gradually increased with the decrease in the SSA. The SSA of the 5Pt/α-Al₂O₃ powder was the smallest among them (ca. 12 m² g⁻¹), while the adsorption and desorption isotherms had little hysteresis and the diameters of the mesopores were mainly less than 5 nm. Figure S3 shows the XRD spectra of all the 5Pt/γ(r)α(t)-Al₂O₃ powders, and Figure 3 shows the dependence of the ratio of the normalized integral intensity of the (400) peak of γ-Al₂O₃ to that of the (116) peak of α-Al₂O₃ (I_400,γ/I_116,α) on the amount of α-Al₂O₃ in the 5Pt/γ(r)α(t)-Al₂O₃ powders. All the peaks of the 5Pt/γ-Al₂O₃ powder were assigned to γ-Al₂O₃ (JCPDS No. 10-0425), while all the peaks of the 5Pt/α-Al₂O₃ powder were assigned to either Pt (JCPDS No. 04-0802) or α-Al₂O₃ (JCPDS No. 71-1123). In addition, the peaks of both γ-Al₂O₃ and α-Al₂O₃ were confirmed in the XRD spectra of the other 5Pt/γ(r)α(t)-Al₂O₃ powders. On the other hand, the small peaks of Pt were detected only in the 5Pt/γ(75)α(25)-Al₂O₃ powder among them, which indicates that the crystallinity and/or crystallite size of Pt nanoparticles loaded to the γ(r)α(t)-Al₂O₃ powders increased with an increase in the amount of α-Al₂O₃ in the 5Pt/γ(r)α(t)-Al₂O₃ powders. Furthermore, the I_400,γ/I_116,α value monotonically decreased with an increase in the amount of α-Al₂O₃ in 5Pt/γ(r)α(t)-Al₂O₃. The dependences of both SSA (Figure 2) and I_400,γ/I_116,α (Figure 3) values on the amount of α-Al₂O₃ in the 5Pt/γ(r)α(t)-Al₂O₃ obviously show that α-Al₂O₃ was uniformly mixed with γ-Al₂O₃ in the 5Pt/γ(r)α(t)-Al₂O₃ powders.

Figures S4 and S5 show the XPS spectra of Pt 4f, Al 2s, and O 1s of all 5Pt/γ(r)α(t)-Al₂O₃ powders, and variations in the ratio of Pt or Al to all Pt and Al species and the ratio of each Pt component to all Pt species with the amount of α-Al₂O₃ in the 5Pt/γ(r)α(t)-Al₂O₃ powders are shown in Figure 4. All the ratios were calculated by the deconvolution of their XPS spectra. The half-widths of both the Al 2s and O 1s spectra gradually narrowed with an increase in the amount of α-Al₂O₃ (“from ca. 2.7 eV to ca. 2.0 eV” and “from ca. 2.8 eV to ca. 1.8 eV”, respectively). In addition, the binding energy of the Al 2s spectra was hardly dependent on the amount of α-Al₂O₃, whereas the binding energy of the O 1s spectra was negatively shifted with an increase in the amount of α-Al₂O₃. On the other hand, the amount of Pt⁰ (Pt metal) increased and the amount of Pt²⁺ (such as PtO) decreased with the amount of α-Al₂O₃, while the amount of Pt⁴⁺ (such as PtO₂) remained almost unchanged. Nevertheless, the binding energy of Pt positively shifted with the amount of α-Al₂O₃, while the amount of Pt²⁺ was much larger than those of Pt⁰ and Pt⁴⁺. These results indicate that the electron density in all the Pt components decreased with an increase in the amount of α-Al₂O₃ and that the electrons were transferred to the oxygen species of the Al₂O₃-components due to the interaction between Pt and Al₂O₃. Moreover, the ratio of Pt to all Pt and Al species on the 5Pt/γ(r)α(t)-Al₂O₃ surface increased with an increase in the amount of α-Al₂O₃, even though the amount of Pt loaded onto the mixed aluminas was only 5 wt% for all 5Pt/γ(r)α(t)-Al₂O₃ powders. In other words, the amount of Pt loaded per unit SSA increased with an increase in the amount of α-Al₂O₃ because the SSA of α-Al₂O₃ was much smaller than that of γ-Al₂O₃, as shown in Figure 2.

The dispersibility, surface area, and average particle size of Pt loaded onto the aluminas of representative 5Pt/γ(r)α(t)-Al₂O₃ (5Pt/γ-Al₂O₃, 5Pt/γ(50)α(50)-Al₂O₃, and 5Pt/α-Al₂O₃) powders were estimated from the CO adsorption/desorption properties, as shown in

Table 1. Dispersibility, surface area, and average particle size of Pt loaded onto the aluminas of typical 5Pt/γ(r)α(t)-Al₂O₃ powders.

Sample	5Pt/γ-Al2O3	5Pt/γ(50)α(50)-Al2O3	5Pt/α-Al2O3
Dispersibility/%	47.8	32.0	14.0
Surface area/m2 g−1-Pt	118	80.0	34.5
Average particle size/nm	2.37	3.55	8.12

. The dispersibility and surface area of the Pt nanoparticles decreased and then the average particle size of Pt increased with an increase in the amount of α-Al₂O₃. These results indicate that the Pt nanoparticles, which were smaller than those on the surface of α-Al₂O₃, were loaded in a highly dispersed state on the surface of γ-Al₂O₃. The dispersibility and surface area of 5Pt/γ(50)α(50)-Al₂O₃ were relatively close to the average values of α-Al₂O₃ and γ-Al₂O₃ (dispersibility: ca. 31%, surface area: ca. 76 m² g⁻¹-Pt). This suggests that the Pt nanoparticles loaded on the surface of α-Al₂O₃ and γ-Al₂O₃ of 5Pt/γ(50)α(50)-Al₂O₃ existed in a relatively similar state to the Pt nanoparticles loaded on the surface of α-Al₂O₃ and γ-Al₂O₃ alone. Therefore, the average particle size of the Pt nanoparticles of 5Pt/γ(50)α(50)-Al₂O₃ calculated from the CO adsorption/desorption properties is not very meaningful when discussing the loading state. However, just the value is also very reasonable because it lies between the Pt nanoparticles on the 5Pt/γ-Al₂O₃ surface, which has a small average particle size, and the Pt nanoparticles on the 5Pt/α-Al₂O₃ surface, which has a large average particle size.

Figure 5 shows TEM photographs of representative 5Pt/γ(r)α(t)-Al₂O₃ (5Pt/γ-Al₂O₃, 5Pt/γ(50)α(50)-Al₂O₃, and 5Pt/α-Al₂O₃) powders. The γ-Al₂O₃ powder consisted of very fine particles, which were quite aggregated. In contrast, the α-Al₂O₃ powder was formed from particles with a diameter of approximately 200 nm. Furthermore, it was confirmed that the particles of both γ-Al₂O₃ and α-Al₂O₃ powders were uniformly mixed in the TEM photograph of 5Pt/γ(50)α(50)-Al₂O₃. Many Pt nanoparticles were confirmed with both γ-Al₂O₃ and α-Al₂O₃ particles in all highly magnified TEM photographs. The sizes of the Pt nanoparticles in the 5Pt/γ-Al₂O₃ and 5Pt/γ(50)α(50)-Al₂O₃ powders were smaller than those of the Pt nanoparticles in the 5Pt/α-Al₂O₃ powder, but their sizes seem to be comparatively smaller than the average sizes of the Pt nanoparticles, which were calculated from the amount of CO adsorption (Table 1).

Figure S6 shows the annular dark-field (ADF)-STEM images and EDS maps (Al K and Pt M) of representative 5Pt/γ(r)α(t)-Al₂O₃ (5Pt/γ-Al₂O₃, 5Pt/γ(50)α(50)-Al₂O₃, and 5Pt/α-Al₂O₃) powders, together with their TEM photographs. In all the samples, the EDS signal of Al K almost overlapped with the morphology of the alumina particles, as confirmed in the TEM photographs. The EDS signal of Pt M was also observed at almost the same location as the EDS signal of Al M in the TEM and ADF-STEM images of 5Pt/γ-Al₂O₃ and 5Pt/γ(50)α(50)-Al₂O₃, which indicates that Pt nanoparticles were uniformly loaded on both γ-Al₂O₃ and α-Al₂O₃ particles in these samples. On the other hand, on the basis of the ADF-STEM images and the EDS maps of the Pt M signal, it was confirmed that not only Pt nanoparticles randomly loaded on the surface of the α-Al₂O₃ particles, but also that a large amount of Pt nanoparticles was agglomerated at the interfaces (grain boundaries) of the α-Al₂O₃ particles. The large average size of the Pt nanoparticles in the 5Pt/α-Al₂O₃ powder (

Table 2. Thermal conductivities of typical samples.

Sample	Thermal Conductivity/10−2 W m−1 K−1
γ-Al2O3	9.32
5Pt/γ-Al2O3	9.63
10Pt/γ-Al2O3	9.97
γ(50)α(50)-Al2O3	14.9
α-Al2O3	37.7

), which was calculated from the amount of CO adsorbed, probably originates from their agglomeration. Such agglomeration of Pt nanoparticles was not observed in the 5Pt/γ(50)α(50)-Al₂O₃ powder, because the γ-Al₂O₃ particles were well mixed with the α-Al₂O₃ particles and the γ-Al₂O₃ particles were well inserted among the α-Al₂O₃ particles, as shown in the TEM and ADF-STEM images. The amounts of Pt nanoparticles loaded in 5Pt/γ-Al₂O₃, 5Pt/γ(50)α(50)-Al₂O₃, and 5Pt/α-Al₂O₃, which were evaluated by EDS analysis, were ca. 5.1 wt%, ca. 4.7 wt%, and ca. 4.9 wt%, respectively, and these values were almost the same as the amount of Pt added during the preparation process.

The thermal conductivity of the sensing materials is one of the most important factors for adsorption/combustion-type gas sensors, because the thermal energy generated by the oxidation of VOCs on the catalyst surface needs to be effectively transferred to the Pt microheater (detector) to enhance the magnitude of the VOC response. Table 2 shows the thermal conductivities of representative sample disks fabricated under the same conditions as the sensor manufacturing conditions (heat treatment at 500 °C for 2 h in ambient air). The thermal conductivity of the γ-Al₂O₃ disk was small, and the thermal conductivity largely increased with an increase in the amount of α-Al₂O₃ added to γ-Al₂O₃ and the thermal conductivity of the α-Al₂O₃ disk was very large. On the other hand, the thermal conductivity of the nPt/γ-Al₂O₃ disks also tended to increase with an increase in the amount of Pt loaded, but the increase ratio was small in the range of 0–10 wt%. These results indicate that the addition of α-Al₂O₃ is effective in increasing the thermal conductivity of the sensing film.

3.2. Catalytic Activities

Figure 6a shows the catalytic combustion behavior of 100 ppm ethanol and 100 ppm toluene over representative 5Pt/γ(r)α(t)-Al₂O₃ (5Pt/γ-Al₂O₃, 5Pt/γ(50)α(50)-Al₂O₃, and 5Pt/α-Al₂O₃) powders in dry air at a flow rate of 20 cm³ min⁻¹. The oxidation activities of 5Pt/γ-Al₂O₃ and 5Pt/γ(50)α(50)-Al₂O₃ for ethanol were extremely high, achieving nearly 100% ethanol conversion even at room temperature. However, there was a large difference in the amount of acetaldehyde produced as an intermediate product, and the amount of acetaldehyde produced over 5Pt/γ-Al₂O₃ was clearly smaller than that over 5Pt/γ(50)α(50)-Al₂O₃. The amount of CO₂ produced over both samples did not increase until relatively high temperatures were reached, and the CO₂ concentration reached 200 ppm, which is the concentration produced when ethanol is completely oxidized, at 250 °C. This was because other partially decomposed products except for acetaldehyde were produced over both samples. In fact, previous studies confirmed the production of ethylene and diethyl ether as partially decomposed products at incomplete medium temperatures for ethanol oxidation [22]. No partially decomposed products other than acetaldehyde were confirmed because of the low ethanol concentration used in this study. The catalytic combustion behavior of 100 ppm ethanol over representative 5Pt/γ(r)α(t)-Al₂O₃ (5Pt/γ-Al₂O₃, 5Pt/γ(50)α(50)-Al₂O₃, and 5Pt/α-Al₂O₃) powders in dry air was also investigated at a higher flow rate of 100 cm³ min⁻¹, as shown in Figure S7a. It was not possible to quantify the amount of CO₂ produced in this experiment because compressed air was used as the base gas. Increasing the gas flow rate obviously clarified that the oxidation activity of 5Pt/α-Al₂O₃ for ethanol was higher than that of 5Pt/γ(50)α(50)-Al₂O₃. These results confirmed that the ethanol oxidation activity decreased with increasing amounts of α-Al₂O₃ added to 5Pt/γ(r)α(t)-Al₂O₃, owing to the large specific surface area of 5Pt/γ-Al₂O₃, as well as the high dispersibility of Pt nanoparticles and the large Pt surface area.

Figure 6b shows the catalytic combustion behavior of 100 ppm toluene over representative 5Pt/γ(r)α(t)-Al₂O₃ (5Pt/γ-Al₂O₃, 5Pt/γ(50)α(50)-Al₂O₃, and 5Pt/α-Al₂O₃) powders in dry air at a flow rate of 20 cm³ min⁻¹. The toluene conversion property of 5Pt/γ(50)α(50)-Al₂O₃ was the highest among them, but 5Pt/α-Al₂O₃ also showed a relatively high toluene conversion property. The amount of CO₂ generated, which is an indicator of complete oxidation, over 5Pt/α-Al₂O₃ was larger than that over 5Pt/γ(50)α(50)-Al₂O₃, even at low temperatures, confirming that partially decomposed products of toluene are not easily generated on the surface of 5Pt/α-Al₂O₃. On the other hand, the oxidation activity of 5Pt/γ-Al₂O₃ for toluene was much lower than that of the others, whereas the oxidation activity of 5Pt/γ-Al₂O₃ was comparable to that of the others only at lower temperatures. The catalytic combustion behavior of 100 ppm toluene over representative 5Pt/γ(r)α(t)-Al₂O₃ (5Pt/γ-Al₂O₃, 5Pt/γ(50)α(50)-Al₂O₃, and 5Pt/α-Al₂O₃) powders in dry air was also investigated at a higher flow rate of 100 cm³ min⁻¹, as shown in Figure S7b. The amount of CO₂ produced was not able to be monitored in this experiment, too. The difference in the catalytic activity between 5Pt/γ(50)α(50)-Al₂O₃ and 5Pt/α-Al₂O₃ was clarified by increasing the gas flow rate, and it was confirmed that the toluene oxidation activity of 5Pt/γ(50)α(50)-Al₂O₃ was higher than that of 5Pt/α-Al₂O₃. It is also noteworthy that the oxidation activity of 5Pt/γ-Al₂O₃ for toluene became comparable to that of 5Pt/α-Al₂O₃ when the gas flow rate increased from 20 cm³ min⁻¹ to 100 cm³ min⁻¹, which means that the gas flow rate (the rate at which toluene is supplied to the catalyst surface) is one of the most important factors for the relative comparison of the oxidation activities of the samples.

3.3. VOC-Sensing Properties

Figure 7 shows the sensor-signal profiles of representative 5Pt/γ(r)α(t)-Al₂O₃ (5Pt/γ-Al₂O₃, 5Pt/γ(50)α(50)-Al₂O₃, and 5Pt/α-Al₂O₃) sensors to ethanol and toluene in ambient air. These sensor-signal profiles largely depended on the kind of VOCs and their concentration, as well as the composition of the aluminas, that is, the amount of α-Al₂O₃, in the 5Pt/γ(r)α(t)-Al₂O₃ powders. The 5Pt/γ-Al₂O₃ sensor shows “medium dynamic responses and relatively small static responses to ethanol” and “quite small dynamic responses and medium static responses to toluene”. The 5Pt/γ(50)α(50)-Al₂O₃ sensor showed “quite large dynamic responses and relatively small static responses to ethanol” and “small dynamic response and medium static response to toluene”, but the dynamic responses of the 5Pt/γ(50)α(50)-Al₂O₃ sensor to toluene were much larger than those of the 5Pt/γ-Al₂O₃ sensor. On the other hand, the 5Pt/α-Al₂O₃ sensor showed small dynamic responses and large static responses to both gases. As mentioned above, the mixing of α-Al₂O₃ with γ-Al₂O₃ can significantly change the sensor-signal profiles to ethanol and toluene. Figure 8 summarizes the composition dependencies of the three kinds of responses of all the 5Pt/γ(r)α(t)-Al₂O₃ sensors to ethanol and toluene. The general response, ΔV_MAX, and integrated dynamic response, IDR, to ethanol were the largest when the amount of α-Al₂O₃ was 50 wt%, irrespective of the concentration of ethanol, while the maximum of ΔV_MAX and IDR to toluene tended to shift toward a larger amount of α-Al₂O₃. Therefore, the magnitudes of ΔV_MAX and IDR to ethanol were basically larger than those of toluene, but the differences were relatively smaller when the amount of α-Al₂O₃ was larger. The integrated static response, ISR, to 1000 ppm ethanol increased with an increase in the amount of α-Al₂O₃, while those to 10 and 100 ppm ethanol seemed to be the largest at the amount of α-Al₂O₃ of 50 wt%, as with their ΔV_MAX and IDR. On the other hand, the ISR to toluene increased with an increase in the amount of α-Al₂O₃ in all toluene concentration ranges. In addition, the magnitude of ISR to ethanol was smaller than that of toluene, irrespective of the toluene concentration and the amount of α-Al₂O₃.

Figure 9 summarizes the concentration dependencies of the three kinds of responses of all 5Pt/γ(r)α(t)-Al₂O₃ sensors to ethanol and toluene. The ΔV_MAX and IDR, especially the IDR to both ethanol and toluene, tend to be saturated with an increase in their concentration, which indicates that ethanol, toluene, and their partially decomposed products produced during the low-temperature period (150 °C) were adsorbed on the sensing-material surface and they burned during the flash heating-up period from 150 °C to 450 °C. On the other hand, the ISR to both ethanol and toluene increased approximately linearly with an increase in their concentration. This is because ethanol and toluene directly burned over the sensing-material surface during the static operation at 450 °C, and thus the ISR behavior of these sensors is quite similar to the behavior of general catalytic combustion-type gas sensors.

The variations in the three kinds of responses with the amount of α-Al₂O₃ seem to be influenced by the various characteristics of 5Pt/γ(r)α(t)-Al₂O₃. The ethanol oxidation activity decreased with an increase in the amount of α-Al₂O₃. In addition, the amount of ethanol and/or the partially decomposed products generally decreases with an increase in the amount of α-Al₂O₃, because the specific surface area of 5Pt/γ(r)α(t)-Al₂O₃ powder drastically decreased with an increase in the amount of α-Al₂O₃. However, the thermal conductivity increased significantly with an increase in the amount of α-Al₂O₃. The synergetic effects of these factors must determine the largest ΔV_MAX and IDR to ethanol when the amount of α-Al₂O₃ was 50 wt%. The shift of the largest ΔV_MAX and IDR to toluene toward a larger amount of α-Al₂O₃ in comparison with those to ethanol probably arises from the largest toluene oxidation activity of 5Pt/γ(50)α(50)-Al₂O₃. On the other hand, ISRs to both ethanol and toluene increased with an increase in the amount of α-Al₂O₃, which was almost the same as the dependence of thermal conductivity on the amount of α-Al₂O₃. The highly thermally conductive α-Al₂O₃ effectively transfers the thermal energy generated by the oxidation of ethanol and toluene toward the Pt microheater of the 5Pt/γ(r)α(t)-Al₂O₃ sensors during the high-temperature period.

As mentioned above, the 5Pt/γ(r)α(t)-Al₂O₃ sensors have two important sensing-output information in one signal: dynamic response and static response. Therefore, we expect that these attractive characteristics of adsorption/combustion-type VOC sensors will bring great benefits to highly sensitive and selective VOC detection through accurate analysis using adequate computing technologies, such as machine learning and AI technology.

4. Conclusions

The sensing properties of adsorption/combustion-type 5Pt/γ(r)α(t)-Al₂O₃ sensors to ethanol and toluene were investigated in ambient air, and the effects of mixing α-Al₂O₃ and γ-Al₂O₃ as the base material of the sensing film on the sensing properties were discussed in this study. The α-Al₂O₃ powder was well mixed with the γ-Al₂O₃ powder, and the Pt nanoparticles were basically dispersed on the surface of these aluminas (some of the Pt nanoparticles were agglomerated at the grain boundaries only in the γ-Al₂O₃ powder), which was confirmed by the evaluation of the adsorption/desorption characteristics of N₂ and CO, as well as by the analyses using XRD, XPS, TEM, and STEM-EDS. The mixing of 50 wt% α-Al₂O₃ with γ-Al₂O₃ was the most effective in enhancing both ΔV_MAX and IDR to ethanol, while the maximum ΔV_MAX and IDR to toluene tended to shift toward a larger amount of α-Al₂O₃. The ethanol oxidation activity of 5Pt/γ(r)α(t)-Al₂O₃ decreased with an increase in the amount of α-Al₂O₃, whereas the thermal conductivity increased with an increase in the amount of α-Al₂O₃. These results indicate that the synergistic effects of catalytic activity and thermal conductivity of 5Pt/γ(r)α(t)-Al₂O₃ are quite important in enhancing the ethanol-sensing properties of the sensors. On the other hand, the ISR to both ethanol and toluene mainly increased with an increase in the addition of α-Al₂O₃ in 5Pt/γ(r)α(t)-Al₂O₃, even though the oxidation activities of α-Al₂O₃ for ethanol and toluene were lower than those of the other samples. Therefore, the relatively high thermal conductivity of α-Al₂O₃ must be of great importance in the effective transfer of thermal energy generated by the oxidation of ethanol and toluene toward the Pt microheater of the 5Pt/γ(r)α(t)-Al₂O₃ sensors during the high-temperature period.