- Email: [email protected]
Heat transfer control of micro-thermoelectric gas sensor for breath gas monitoring
Accepted Manuscript Title: Heat transfer control of micro-thermoelectric gas sensor for breath gas monitoring Authors: Tomoyo Goto, Toshio Itoh, Takafumi Akamatsu, Yoshitaka Sasaki, Kazuo Sato, Woosuck Shin PII: DOI: Reference:
S0925-4005(17)30530-0 http://dx.doi.org/doi:10.1016/j.snb.2017.03.113 SNB 22025
To appear in:
Sensors and Actuators B
Received date: Revised date: Accepted date:
14-10-2016 13-3-2017 19-3-2017
Please cite this article as: Tomoyo Goto, Toshio Itoh, Takafumi Akamatsu, Yoshitaka Sasaki, Kazuo Sato, Woosuck Shin, Heat transfer control of microthermoelectric gas sensor for breath gas monitoring, Sensors and Actuators B: Chemicalhttp://dx.doi.org/10.1016/j.snb.2017.03.113 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Submitted to: Sensors and Actuators B: Chemical Heat transfer control of micro-thermoelectric gas sensor for breath gas monitoring Tomoyo Goto1,†, Toshio Itoh1, Takafumi Akamatsu1, Yoshitaka Sasaki1,2, Kazuo Sato2, Woosuck Shin1,* 1National
Institute of Advanced Industrial Science and Technology (AIST), 2266-98
Shimo-Shidami, Moriyama-ku, Nagoya 463-8560, Japan 2Aichi
Institute of Technology, 1247 Yachigusa, Yakusa-Cho, Toyota 470-0392, Japan
†Present
address: The Institute of Scientific and Industrial Research (ISIR), Osaka
University, 8-1 Mihogaoka, Ibaraki 567-0047, Japan *Corresponding author: Dr. Woosuck Shin Phone: +81-52-736-7107, Fax: +81-52-736-7244 E-mail: [email protected] Highlights
Catalyst on the micro-TGS lowers the temperature and induces a biased voltage. Additive process of alumina dots changed the heat balance to controls the voltage. Heat transfer analysis by FEM modeling is performed to control the heat balance.
1
Abstract A micro-thermoelectric gas sensor (micro-TGS) uses thermoelectric voltage induced by the catalytic combustion of hydrogen or methane for selective gas detection in breath. This is accomplished under an elevated temperature using a micro-heater built on the same membrane as a hotplate, which enables selective combustion of the target gas. A temperature differential built by the catalyst dot on the membrane induces the offset voltage (Voff) of the micro-TGS, which limits the amplifier circuit application. In this study, we strived to suppress Voff by an additive integration process of heat dissipation dots prepared using -Al2O3 paste. In this paper, we discuss the effects of these Al2O3 dots on the thermal balance over the micro-TGS membrane by finite element method (FEM) modeling. When the dots were deposited in a symmetrical position to the combustion catalyst, Voff was compensated depending on the size, numbers, and locations of the dots. The micro-TGS heat transfer control by the dots was additionally verified by 3-D FEM modeling. The changes in Voff by the -Al2O3 dots in FEM modeling were greater than those of the experiments, suggesting the high thermal conductivity of the micro-TGS membrane. The deviation of the membrane thermal conductivity due to the process non-uniformity significantly influenced the Voff; however, it was effectively reduced by the additive integration of dots. Keywords: Gas sensor; Catalyst; Thermoelectric device; Micro-electromechanical systems (MEMS); Finite element method (FEM)
2
2. Introduction Medical diagnostic technologies for early disease stages have been widely studied in response to global population aging issues and the expansion of the health care industry. Among these technologies, breath analysis techniques have garnered considerable attention as non-invasive and simple health check methods that can be conducted both at home and in a medical facility. Human breath includes several kinds of inorganic and volatile organic compounds (VOCs). These gases are known as markers of disease and human body metabolism [1,2]. To detect the part-per-million (ppm) levels of breath gases, such as H2, CO, and CH4 [3], we developed a unique micro-thermoelectric gas sensor (micro-TGS) with various combustion catalysts for a breath analysis system. Calorimetric gas sensors take advantage of micro-machined devices using the heat of the chemical reactions generated by catalytic metals. Their thermal characteristics must be optimized for better sensing performance. The micro-TGS uses the catalytic combustion of the gas operated at a catalyst temperature over 100 ºC by using a micro-heater built on the membrane. It thereby shows the linear relationship of the voltage signal to the gas concentration [4,5] and the sensing performance for the ppm level of hydrogen in the breath gas analysis [5]. Sensor performance of micro-TGS, a synergistic combination of the catalytic combustion of gas and thermoelectric conversion (from heat to electrical), can be easily modulated by the choice of the combustion catalyst. Previously, micro-TGS for H2, CH4, and CO detection were reported using Pt/-Al2O3 [4,5], Pd/-Al2O3 [6], and Au/Co3O4 [7] catalysts, respectively. To produce an effective design, the heat balance of the sensor during operation must be estimated. The heat balance can predict the inflammable gas combustion energy of the catalyst. The heat as a function of the sensor output is also important for the development of the catalyst. The membrane area of the micro-TGS is comprised of ends of thermoelectric SiGe forming a symmetrical pattern. However, one end of the SiGe pattern is covered by a 3
ceramic combustor, which induces the asymmetrical temperature distribution in the membrane and results in the temperature differential, as shown in Fig 1. Fig. 1(a) depicts an example of a voltage signal of micro-TGS for detection of CH4 at 325 °C. It indicates a problem of micro-TGS in real applications; i.e., voltage from the sensor from the initial stage in air, named the offset voltage (Voff), occurs at a high temperature operation. The fluctuation or drift of the Voff can also be a serious problem when the target gas concentration is low. The voltage signal was thus gradually elevated after 90 s (Voff90) from the initial stage (Voff1). This difference of Voff is shown in Vdirft in Fig. 1(a). After CH4 detection, the voltage continues to rise, and the offset voltage at 270 s (Voff270) is higher than that of the initial stage (Voff1). The origin of this problem was an initial temperature difference (T) formed between point A (catalyst site) and point B:
VS TAB TA TB
(1)
where is the Seebeck coefficient of the thermoelectric film. This V off of micro-TGS increases in high temperature operation because the combustion catalyst on the membrane plays the role of “heat dissipater” (Fig. 1(b)). In addition, non-uniformity of the membrane in micro-TGS also effects the heat balance inside the membrane. Generation of this voltage cannot be ignored because the voltage signal increases with the increasing operating temperature. The suppression of offset voltage of microTGS is important to stabilize the base voltage and improve the quantitative performance. To make an effective design of each part of micro-TGS, especially the catalyst and micro-heater, meanders are important for optimizing the sensor performance. To make effective design of each parts of micro-TGS, especially a double meander-shaped Pt heater and a circular shape catalyst are important to optimize the sensor performance, and moreover it is important to quantitatively evaluate the heat transfer of these parts. Optimization of the micro-heater can involve thermal distribution in the membrane that is aimed at temperature control of a uniform and stable response at elevated 4
temperatures, thereby improving the sensor reliability and sensitivity. The development of a theoretical model for thermo-mechanical calculations and heat transfer mechanisms is important for the design of catalytic sensors, which can include predicting device process non-uniformity. The influence of the micro-hotplate design on the catalytic reaction or sensor response is not fully understood, the model calculation is expected to be effective tool for the sensor development. Most studies have primarily focused on flowmeters [8], while heat and mass transfer in heated systems [9] and thermochemical devices in gas-fueled combustor systems [10] have additionally been investigated. In the case of micro-TGS, the heat of gas combustion at the catalyst film which is thin and smaller than 1 mm, is extremely difficult to directly measure. In our previous study, however, we have quantitively estimated the catalytic combustion heat (power) required for 1 mV of signal voltage to be 46 μW, using the linear calculation and thermal time constants of the experimental signal curves of the micro-TGS [11]. The design of micro-heater meanders can efficiently and uniformly heat the catalyst; moreover, they can prevent the loss of combustion heat though the heater line. Various heater designs have been tested. The power consumption of the micro-heater meanders was reduced to 30–40 mW at the catalyst temperature of 100 ºC, which is approximately 30% less than that of the first generation one [12]. In addition, an asymmetric micro-heater meander enabled different catalyst temperatures for a single sensor with a double catalyst structure. In this study, to control this heat balance and suppress V off of micro-TGS, the -Al2O3 thick film, dummy dots, for heat dissipation were additively integrated on the membrane of micro-TGS. The effects of the -Al2O3 dots on the heat balance are herein described with the quantitative results estimated by 3-D solid modeling of the finite element method (FEM). The effects of thermal conductivity of micro-TGS on the Voff control by -Al2O3 dots are then discussed. Experimental
5
2.1. Design of micro-TGS Details of the preparation procedure of micro-TGS for H2 and CH4 detection were described in previous reports [3-5]. Here, the p-type B-doped SiGe thermoelectric pattern was used for the micro-TGS device. As shown in Fig. 2, the 4 4 mm2 dimension of the micro-TGS device was comprised of the thermoelectric pattern, Pt heater, and electrode line pattern on a double-sided polished Si substrate. The optimal geometrical design comprised of the most advantageous materials was implemented and integrated on a suspended membrane—a double meander-shaped Pt heater of resistance ca. 100 —to heat the 0.6-mm-diameter catalyst at point A in Fig. 2. Details of the dimensions are given in Table 1. The micro-TGS had an active membrane area (heated area, i.e. covered with the heater and catalyst), which was determined by the mask size of the back-side of the silicon substrate of 2.45×1.40 mm. After the KOH wet etching of Si substrate, the membrane size became ca. 2.0×1.4 mm. It was deviated by the process parameters and non-uniformity of the wet etching process. 2.2. Catalyst deposition process of micro-TGS For H2 detection, the Pt/-Al2O3 catalyst powder was prepared by referencing the method of Nishibori et al. [3]. The Pd/-Al2O3 catalyst powder was prepared for CH4 detection of micro-TGS by referencing the work of Nagai et al. [6]. Pt or Pd colloid solutions were purchased from Tanaka Kikinzoku Kogyo K.K., Tokyo, Japan, respectively. Alumina powders, the -Al2O3 and -Al2O3, Taimicron TM-D and TM-100, are supplied by TAIMEI Chemicals Co., Ltd., Nagano, Japan with the specific surface area of 13.5 and 128 m2/g, respectively. In the case of the preparation of catalyst for H2 sensor, 30 wt% Pt was added to -Al2O3 powder in distilled water. Suspension was stirred at 70 ºC until the water evaporated. Mixed powder was calcined at 300 ºC in an electric furnace for 2 h under air. For the CH4 sensor, the 10 wt% Pd/-Al2O3 catalyst powder was prepared by the same procedure as described above. Then, the calcined powder was dispersed in vehicle containing terpineol, ethyl cellulose, and distilled water (9:1:5 wt% ratio). The catalyst pastes were then deposited by air dispenser
6
(Musashi Engineering, Inc., Tokyo, Japan) on the device. Beside the chemical composition of the catalyst, the viscosity of the paste is important parameter for the deposition process. The viscosity determines not only the achieved size of the drop, the final thickness of the thick film, but also affects the dispensing conditions, air pressure and needles. Actually, in the viscosity level of our catalyst pastes, 0.3 Pa・s at shear rate 357/s, the bigger dot is much controllable than small one [13]. The dispenser robot process very precisely equipped with laser distance meter, Z axis, and a camera system for x-y alignment with the precision below 0.05 mm. After the catalyst deposition, the devices were calcined at 300 ºC for 2 h under air. After that, the devices were mounted on stem and connected using gold-bonding wire to prepare the micro-TGS. 2.3. Evaluation of additive integration process of -Al2O3 dots In the case of alumina dots deposition, micro-manipulator is applied and the sensor device with the catalyst deposited already mounted on stem with wire bonding, while the big catalyst dot is deposited before the chip mounting. The -Al2O3 powder, which did not show the catalyst activity, was used as the heat dissipation film (-Al2O3 dots) in this study. The -Al2O3 powder was mixed in a vehicle and deposited on the membrane of the micro-TGS. Then, the micro-TGS was dried at 90 ºC for 30 h. After drying, in the case of the H2 sensor, the offset voltage (Voff) of the micro-TGS was measured under the applied voltage of 1, 2, and 3 V with air-flowing of 200 cm3/min. After measurement, the deposition of -Al2O3 dots and the measurement were repeated to become closer to Voff = 0 mV. In the case of the microTGS for CH4 detection, the micro-TGS was aged at 6 V of applied voltage for 6 h to stabilize the Voff. After this treatment, similar to the above micro-TGS for H2, effects of the position of -Al2O3 dots on Voff were measured at 6 V of applied voltage. The schematic presentation of this evaluation procedure is shown in Fig. 3. 2.4. Simulation of effects of dummy catalyst by FEM modeling
7
A simulation of the micro-TGS heat balance was performed with the idealized multilayer model constructed and analyzed by FEM using a Femtet (Murata Software Co,. Ltd., Tokyo, Japan) environment. In Fig. 4, the geometry of the simplified structures is depicted. Table 1 shows the size of the micro-TGS device. The material property values for FEM modeling are listed in Table 2, where several values are referenced in terms of previous reports [14]. The membrane material was simplified to SiO, which was actually a multilayer of Si3N4 and SiO2. The mesh size of the finite elements was smaller than 0.1 mm for the Pt line, catalyst, and -Al2O3 dots; smaller than 0.2 mm for SiGe and membrane; and smaller than 0.3 mm for the bulk Si rim. The solver of Femtet for thermal analysis called "Watt" carried out electric field analysis and combine into thermal analysis by calculating current density and the Joule's loss. The final Joule heating steady state stage combines the temperature dependent resistivity of the Pt heater. The thermal expansion of the membrane and the Si substrate has been also simulated but not shown here. 3. Results and discussion 3.1. Evaluation of -Al2O3 dots on changes in Voff of micro-TGS Fig. 5 shows the images of (a) the micro-TGS for CH4 and (b-d) the deposited -Al2O3 dots on this sensor, as well as (e) the offset voltage (Voff) at applied 6 V of micro-TGS, as shown in Figs. 5 (a-d). To make small dot is much controllable than large dots in our experiments, however, this made a few dots rather ellipsoidal shape. We have increased the number of dots step by step to check the heat transfer. To investigate the change of Voff by -Al2O3 dots, the micro-TGS for CH4 with an initial value of Voff = 2.41 mV was used. The -Al2O3 dots were deposited in the order of one (single), two (double), and three (triple) from point B to point A (Figs. 5(b-d)). The Voff of micro-TGS were measured after aging for 6 h. The Voff of micro-TGS changed from 2.41 mV to 2.69, 2.38, and 2.25 mV, as shown in Fig. 5(e) and Table 3. From this result, when the dots were deposited on the opposite side of the combustion catalyst (point B), the Voff of micro-TGS slightly increased. In the case of 8
proximity to the catalyst side (point A), the Voff of micro-TGS slightly decreased. It is considered a voltage is induced by heating the membrane and the catalyst, by the Seebeck effect from the temperature difference developed between the A and B points, and the point A is expected to be lower temperature, from the negative sign of the voltage. When the dot is deposited across the Si substrate and membrane, the dot thermally connected them and thermal insulation around the point B becomes worse and the increase of temperature at B point becomes less. We consider that this explanation can be a reasonable to understand the change of the Voff in this study. This tendency was also observed on other micro-TGS (Sensor-2), as shown in Table 3. These results indicated that the Voff with micro-TGS for CH4 detection could be controlled by -Al2O3 dots. Fig. 6 shows the optical images of the micro-TGS for CH4 and deposited alumina, Al2O3, dots on this sensor (a), and the changes in offset voltage by -Al2O3 dot deposition (b), and the changes of V response of micro-TGS for H2 and CH4 (c). For this sensor of gas detection, large alumina dot is deposited. As discussed above with the results of Fig. 5, the additive deposition of alumina dots lowered the V off effectively, and the initial large positive Voff reduced to around 0 by the first dot, and again to negative value by the second dot. The ΔV of the micro-TGS for gas of CH4 and H2 are not affected as shown in Fig.5(c), showing the detection performance is not disturbed by alumina dots. The ΔV of the micro-TGS for 200 ppm CH4 is around 0.1 mV, which is reasonable value from the results of previous reports [11, 15], which is successful demonstration of additive dot deposition. Fig. 6(d) shows the response of the micro-TGS for CH4 and the response after the dot deposition changed. From this, we could suggest that the combustion around the catalyst is slightly affected by the dots. However, in the case of H2 detection, Fig. 6(e), the responses of the micro-TGS are exactly the same with or without the alumina dots.
9
Fig. 7 shows the optical images of (a-e) the micro-TGS for H2 and deposited -Al2O3 dots on this sensor, and (f) the changes in offset voltage by -Al2O3 dot deposition. From the results of micro-TGS for CH4 (Fig. 5), we strived to control the Voff of the micro-TGS for H2. In this experiment, the micro-TGSs of negative Voff were tested to validate this additive process. Fig. 7(f) shows the change of Voff of micro-TGS by heating with increasing the applied voltage and with four alumina dots (Fig. 7(e)). From the results of Fig. 7(f), the Voff increased around 0.2 mV by first (Fig. 7(b)) and third dots (Fig. 7(d)), and second (Fig. 7(c)) and firth (Fig. 7(e)) dot reduced the V off around 1 mV. These results indicated that the dots effects become larger by cover area and position of dots. By increasing the applied heater voltage, which was roughly proportional to the power of joule heating, the negative Voff became high. However, as the number of dots on the counter side close to point B increased, the negative V off became low. This tendency was additionally confirmed by other micro-TGSs. As shown in Fig. 8, the Voff of four additional micro-TGSs with the applied voltage 3.0 V were modulated by the dots and tested. On the evaluation of H2 sensors, 5 sensors were evaluated for investigate the effect of dots. These sensors were numbered from TGS-1 to TGS-5. To investigate the effects of number of dot deposition, all sensors were deposited four points in the same procedure as shown in Fig. 7. Their results were analogous to the shifting of their Voff close to 0 V. In the case of TGS-3, Voff value become around 0 by deposition of one dots, and Voff of TGS-3 was almost constant regardless of additional deposition. In contrast, TGS-4 and TGS-5 showed around -3 mV after deposition of four dots. A deviation of the Voff of different sensors occurred. It is considered that this deviation originated from the individual differences produced in the fabrication process, such as the membrane size or partial etching during the wet etching process, or the catalyst diameter or thickness during the catalyst deposition process. The controllability of the alumina dot deposition is important and we have controlled the size and thickness of the dots, but their variance easily leads to the deviation. However, in this study, the position of the dots seems much important parameter makes the difference, especially whether the dot is across thin membrane to Si
10
substrate or not. From these results, to control the Voff, produce dots in the most effective location or with similar shape is important. By the additive process of alumina dot deposition, this deviation by as much as -3 to -11 mV was successfully reduced to as small as 0 to -3 mV. 3.2. Simulation of effects of -Al2O3 dots on heat balance of micro-TGS In the FEM modeling, two dummy points were deposited on the opposite side of the catalyst (point B), as shown in Fig. 9. The solver of Femtet for thermal analysis called "Watt" carried out electric field analysis and combine into thermal analysis by calculating current density and the Joule's loss. The final Joule heating steady state stage combines the temperature dependent resistivity of the Pt heater. The thermal expansion of the membrane and the Si substrate has been also simulated but not shown here. The temperature profile of the suspended heater is shown in Fig. 9 for an SiO2 (1 m thick) membrane with a fixed thermal conductivity (10 W/mK), a fixed value of applied voltage of 3 V to the two heater electrodes, and a temperature of the Si substrate bottom of 25°C. The SiGe pattern in the center had the highest temperatures. Meanwhile, the connecting Pt leads remained unheated, and the temperature uniformity in the center heated area was high but low at the corner. The local heating confinement of heat in the high resistive area was demonstrated. The maximum temperature reached by applying 3 V was 221°C without dots and decreased to 202 °C and 181 °C by single dot and double dots, respectively. These effectively transferred the heat generated by the Pt heater joule heating on the membrane to the Si rim. The FEM results in Fig. 9 are summarized in Table 4. Although the amount of changes of Voff by -Al2O3 dots is large compared with that of the experimental results, more details of the device operation can be depicted. Furthermore, the data again indicates that the Voff of micro-TGS is reduced by deposition of -Al2O3 dots. The FEM results also indicated that the Voff of micro-TGS was changed by deposition of -Al2O3 dots and the Voff was negative; that is, the temperature of the catalyst side (point A) 11
was relatively lower than the other parts. Therefore, the micro-TGS of negative Voff could be controlled by the -Al2O3 dots. Details of the thermal resistance analysis by heat transfer evaluation are omitted in this study because the minimization of power consumption was not our goal, but investigating the effect of alumina dots is the objective. The heater power of the TGS is similar to the real heater power of the device, which is around 125 mW, where the calorific value of the device in the simulation varies from 129 to 136 mW. When the heat transfer from the membrane to the Si rim through dots increases, the temperature of point B becomes relatively lower than that of A, and the Voff transitions to a positive value. The simulation results clearly show this effect and most devices showed same trends even though their offset values are deviated as shown in Fig. 8. To take account the thermal resistance of alumina is important, because the thermal resistance is a parameter which express the thermal flow per unit area between two systems, alumina and gas flow. However, the thermal resistance of the membrane including catalyst (alumina particle deposit with Pt particle), which is considered almost same specific surface area [6] to that of alumina dot is estimated experimentally to be 230 K/mW [11] and this value is not changed much (within 1%) from the membrane without the catalyst. 3.3. Adjusting the thermal conductivity of the membrane by comparing FEM and the experimental values of micro-TGS with dots In this simulation, the most important parameter was the thermal conductivity of the membrane of micro-TGS. Its effect was quantitatively investigated by FEM simulation. Previous analysis by FEM was calculated using SiO2 and Si3N4 as the singlelayer membrane. However, in a real sensor device, the membrane consists of multilayer structures comprised of two SiO2 layers and an Si3N4 layer (SiO2 250+250 nm and Si3N4 330 nm, respectively). Therefore, the thermal conductivity (κeff) can be written as the following calculation:
12
eff = SiO2×42.5% + Si3N4×57.5% Using the reported value of SiO2 and Si3N4 [14, 16], eff was estimated to be 2.714 W/mK. However, considering that the thickness of the multi-layer was not uniform, and the wet etching process could remain as Si underneath the Si 3N4 layer, the thermal conductivity of the membrane was unpredictable and difficult to be measured. The changes in Voff with the addition of -Al2O3 dots, as shown in Fig. 4, were simulated. The thermal conductivity of the membrane was set to be varied from 1.38 to 13.8 W/mK. The potential between the electrodes was set to be 2.75 V, which led to a close value of the catalyst temperature of micro-TGS at approximately 125 ºC. Fig. 10 shows the results of the FEM simulation of changes in V off of micro-TGSs based on the number of -Al2O3 dots deposited. The Voff shifts from negative to positive with the addition of -Al2O3 dots, which is analogous to the above results. Their change is reduced by increasing the thermal conductivity of the membrane. In addition, the effects of -Al2O3 dots on Voff slightly decrease with the decrease of thermal conductivity. The width of the Voff change of the TGS without -Al2O3 dots was fairly large at 7.8 mV, which corresponded to theV response of micro-TGS for over 5,000 ppm H2 in air. This width became narrow with the addition of dots to 2.4 and 1.4 mV for single and double dots, respectively. The width of Voff decreased by more than five times as the thermal conductivity of the membrane increased ten times. This confirmed the feasibility of the proposed concept. In comparing these results to the voltages of real sensors, the closest or average value of the thermal conductivity of the multi-layer membrane was around 8 W/mK. Although the amount of change in Voff of the FEM simulation was not particularly comparable to that of the experimental results, the much higher value of the thermal conductivity of the membrane implied the rather poor precision of the wet etching process. This can provide guidance in understanding their effect on the heat transfer in the sensor device. The accuracy of the Voff and, more fundamentally, the temperature 13
distribution in the membrane could be improved by optimizing the parameters of heat dissipation or thermal resistance values in the FEM simulation. The lack of reproducibility of wet etching is considered to be the origin of the deviation. Not only additive alumina dots shown in this study but dry process technique such as deep RIE which can produce much cleaner membrane can overcome this thermal valance problem. Nonetheless, that approach would be beyond the scope of this study. The heat dissipations of the micro hotplates with the various catalyst or sensing materials of thicknesses of 5-10 m have been calculated [17] and there is no clear differences in thermal conduction below 600 ºC . 3.4. Gas response of micro-TGS with dots In the application of breath analysis, operating the sensor at a lower catalyst temperature is critical to ensuring the selectivity. However, the catalyst becomes inactive for long-term operation at a lower temperature. To activate the catalyst, temporal heating, a so-called flashing operation, is necessary before the normal detection state. Fig. 11(a) shows the gas sensing performance of “TGS-1” with quadruple -Al2O3 dots (As shown in Fig. 8) with the catalyst of 30wt% Pt/-Al2O3 for 200 ppm H2 in air. The catalyst of TGS-1 is elevated at the temperature of 100, 125, and 150 °C. The applied voltages and the corresponding Voff are plotted in Fig. 11 (b). The first response of TGS-1 at catalyst temperature 100 °C shows a delayed unsaturated response, suggesting that the catalyst is inactive. To activate it, the temperature is increased to 150 °C and hydrogen gas is again detected. After this point, the catalyst becomes active and then the TGS-1 is operated at 125 °C, when it shows normal detection performance. Although we have controlled the Voff for CH4 sensing, its response was not stable and not shown here. It is considered to contain other problems of instability of catalytic activity, which is currently under investigation. For the specific pursuit of our prototype breath hydrogen detector, the Voff should be less than 0.5 mV and the Voff of micro-TGS is reduced by the dots to satisfy this 14
criterion. The voltage signal Vs for 200 ppm H2 is fairly stable for the use of the prototype detector. Its level is sufficiently high at 0.2 mV for 200 ppm, which is the same as the normal micro-TGS 0.1 mV for 100 ppm at the catalyst temperature of 125 °C, as reported in [18]. 4. Conclusion In this study, the additive process of alumina dots on the membrane of the microTGS was performed. The heat transfer analysis by 3-D FEM modeling was carried out to quantitatively validate the heat valance control. The micro-TGS with and without dots at an operational temperature at 325 ºC and 125 ºC was tested as a proof-of-concept. The voltage was negatively biased because the catalyst lowered the temperature of the surrounding membrane. When the dot was deposited on the opposite side of the combustion catalyst, the negative Voff of micro-TGS was reduced. This tendency was also verified by the subsequent model results of the FEM analysis using the designs of dots on the micro-TGS membrane. The implementation of a membrane composed of low thermal conductive materials, i.e., SiO2 and Si3N4, yielded a reduction of thermal losses to the silicon substrate. It thus improved the thermal efficiency of the system. Nevertheless, the actual thermal conductivity of the membrane material had to be considered to enable the wet etching process to retain the Si. The main heat loss mechanism in the micro-TGS was via the gas phase of air atmosphere. However, the relative temperature change was mainly affected by the thermal valance of heat transfer from the membrane to the Si rim. The heat transfer could be controlled by the dots, which increased the heat transfer to the rim and reduced the temperature of the opposite side. Consequently, the absolute value of V off was reduced. The conductivity of the membrane significantly influenced the Voff, and the deviation of the Voff was thereby strongly affected. However, their width was reduced by the integration of dots.
15
Acknowledgement This work was supported by the Knowledge Hub Aichi (the priority research project: P3-G3-S1) of Aichi Prefecture, Japan. The authors are grateful to AIST Laboratory members, Atsuko Hotta, Michi Yamaguchi, and Takaomi Nakashima for their technical support in the preparation of the devices and sensor tests used in this study. The authors also wish to thank Assoc. Prof. Maiko Nishibori of Kyushu University, Japan for her comments on this study.
References [1] A. Amann, G. Poupart, S. Telser, M. Ledochowski, A. Schmid, S. Mechtcheriakov, Applications of breath gas analysis in medicine, Int. J. Mass Spectrom. 239 (2004) 227-233. [2] Y. Broza, P. Mochalski, V. Ruzsanyi, A. Amann, H. Haick, Hybrid volatolomics and disease detection, Angew. Chem. 54 (2015) 11036-11048 [3] W. Shin, Medical applications of breath hydrogen measurements, Anal.
Bioanal.
Chem. 406 (2014) 3931–3939. [4] M. Nishibori, W. Shin, L. F. Houlet, K. Tajima, T. Itoh, N. Izu, N. Murayama, I. Matsubara, New structural design of micro-thermoelectric sensor for wide range hydrogen detection, J. Ceram. Soc. Japan 114 (2006) 853-856. [5] W. Shin, M. Nishibori, N. Izu, T. Itoh, I. Matsubara, K. Nose, A. Shimouchi, Monitoring breath hydrogen using thermoelectric sensor. Sens. Lett. 9 (2011) 684-687. [6] D. Nagai, M. Nishibori, T. Itoh, T. Kawabe, K. Sato, W. Shin, Ppm level methane detection using micro-thermoelectric gas sensors with Pd/Al2O3 combustion catalyst films, Sens. Actuator B: Chem. 206 (2015) 488-494. [7] M. Nishibori, W. Shin, N. Izu, T. Itoh, I. Matsubara, CO oxidation performance of
16
Au/Co3O4 catalyst on the micro gas sensor device, Catal. Today 201 (2013) 85-91. [8] F. Udrea, J. Gardner, D. Setiadi, J. Covington, T. Dogarua, C. Lua, W. Milne, Design and simulations of SOI CMOS micro-hotplate gas sensors, 78 (2001) 180–190 [9] P. van Malea, M. de Croona, R. Tiggelaarb, A. van den Bergb, J. Schouten, Heat and mass transfer in a square microchannel with asymmetric heating, [10] S. Karagiannidis, K. Marketos, J. Mantzaras, R. Schaeren, K Boulouchos, Experimental and numerical investigation of a propane-fueled, catalytic mesoscale combustor, Catalysis Today 155 (2010) 108–115 [11] D. Nagai, T. Akamastu, T. Itoh, N. Izu, W. Shin, Thermal balance analysis of a microthermoelectric gas sensor using catalytic combustion of hydrogen, Sensors, 14 (2014) 1822-1834 [12] M. Nishibori, W. Shin, N. Izu, T. Itoh, I. Matsubara, Microheater Meander configurations for combustion catalysts in thermoelectric gas sensor, Sens. Lett. 8 (2010) 792-800 [13] W. Shin, M. Nishibori, M. Ohashi, T. Itoh, N. Izu, I. Matsubara, Ceramic catalyst combustors of Pt-loaded-alumina on microdevices, J. Ceram. Soc. Jpn 117 (2009) 659-665 [14] Y.S. Touloukian, Thermophysical Properties of Matter, Vol. 12, Thermal Expansion, IFI/Plenum, New York, 1975. SiN data from; http://accuratus.com/silinit.html. [15] N. Park, T. Akamatsu, T. Itoh, N. Izu, W. Shin,“Calorimetric Thermoelectric Gas Sensor for the Detection of Hydrogen, Methane, and Mixed Gases”, Sensors 14 (2014) 8350-8362 [16] E.Vereshchagina, R. Tiggelaar, R. Sanders, R.Wolters, J. Gardeniers, Low power micro-calorimetric sensors for analysis of gaseous samples, Sens. Actuator B: Chem.
17
206 (2015) 772-787 [17] M. Nishibori, W. Shin, N. Izu, T. Itoh, I. Matsubara, Sensing performance of thermoelectric hydrogen sensor for breath hydrogen analysis, Sens. Actuator B: Chem. 137 (2009) 524–528. [18] F. Biró, C. Dücső, Z. Hajnal, F. Riesz, AE. Pap, I. Barsony, Thermo-mechanical design and characterization of low dissipation micro-hotplates operated above 500 ºC, Microelectronics Journal 45 (2014)1822–1828
18
Biographies Tomoyo Goto received her PhD in crystalline materials science from Nagoya University (2012), Japan. From 2012 to 2014, she was a postdoctoral fellow at Kyushu University in Japan. She worked at the Electroceramics Processing Group, Advanced Manufacturing Research Institute (AMRI), AIST, in Japan from 2014 to 2015. She is currently an assistant professor at ISIR, Osaka University in Japan. Her research interests include bioceramics, purification materials, and gas sensor using combustion catalysts. Toshio Itoh studied organic–inorganic hybrid materials and finished his master’s course in 2002 at Nagoya University in Japan and received a doctorate in 2005 in crystalline materials science from the same institution. He has been a member of the research staff in Electroceramics Group (formerly Electroceramics Processing Group) at Inorganic Functional Materials Research Institute (IFMRI), AIST in Japan since 2005. His research activities include VOC gas sensing devices using organic–inorganic hybrid materials. Takafumi Akamatsu received his PhD in Engineering from Nagoya Institute of Technology, Japan in 2008. From 2008 to 2012, he worked at Nihon Yamamura Glass Co., Ltd. He is currently a research member of Electroceramics Group at IFMRI, AIST, in Japan. His research interests include metal oxide semiconductors for NOx sensor and glass-derived hydrogels for proton conductor. Yoshitaka Sasaki studied mechanical engineering and finished his master’s course in 2016 at Aichi Institute of Technology under the supervision of Prof. K. Sato. He joined AIST group in 2015 for 1 year as a research assistant working for the fabrication and FEM simulation of micro gas sensors. Kazuo Sato received B.S. from Yokohama National University in 1970. He worked with Hitachi, Ltd., during 1970–1994. He received Ph.D. from the University of Tokyo in 1982. He started MEMS research since 1983 in Hitachi. He joined Nagoya University in 1994 as a full professor of MEMS/Micromachining Lab. He became an emeritus in 2012. He joined Aichi Institute of Technology in 2012 as a professor. His research area includes; micromachining, KOH/TMAH etching of Si, etching of Quartz, micro/nano physics of wet anisotropic etching of single crystals, characterization of mechanical properties of thin film Si, micro/nanophysics, microactuators, applied microsystems. Woosuck Shin received his BS and MS in material science and engineering in 1992 and 1994, respectively, at KAIST, Korea. After receiving his PhD in applied chemistry from Nagoya University in 1998, Japan, he has been employed by AIST, Nagoya, Japan. He is currently the group leader of the Electroceramics Group of IFMRI, AIST in Japan. His research interests include a functional thin film for gas sensor technology with micro-systems.
19
Figure captions Fig. 1
(a) Voltage signal (Vs) of micro-TGS for detection of CH4 and (b) schematic illustration of the micro-TGS.
Fig. 2
Optical image of the micro-TGS.
Fig. 3
Schematic presentation of evaluation procedure of offset voltage (Voff) of microTGSs.
Fig. 4 CAD images for simulation model of single and double -Al2O3 dots on microTGS by FEM modeling. Fig. 5 Optical images of (a) the micro-TGS for CH4 sensor, and deposited (b) single, (c) double, and (d) triple -Al2O3 dots on this micro-TGS. (e) The offset voltage signal of micro-TGS with or without -Al2O3 dots. Fig. 6
(a) Optical images of the micro-TGS for CH4 and deposited alumina, -Al2O3, dots on this sensor, (b) Changes in Voff of the micro-TGS with the 10 wt% Pd/Al2O3 catalyst by number of -Al2O3 dot deposition, (c) changes of ΔV against the number of alumina dots of micro-TGS for CH4 and H2 responses, (d) gas response of the micro-TGS with -Al2O3 dots for 200 ppm CH4 and (e) H2 in air.
Fig. 7 Optical images of (a) the micro-TGS for H2 and (b) single, (c) double, (d) triple, and (e) quadruple deposited -Al2O3 dots on this sensor. (f) The changes in the offset voltage (Voff) by -Al2O3 dot deposition. Fig. 8 Changes in offset voltage (Voff) of micro-TGSs by number of -Al2O3 dot deposition. Five sensors were applied at 3 V. Fig. 9
FEM models of heat balance of the micro-TGS device. The surface temperature
20
of the devices of (a) no dot, (b) single, and (c) two -Al2O3 dots on the membrane are shown. Fig. 10
Analysis of the variation of Voff of the micro-TGS device with -Al2O3 dots for thermal valance, where the thermal conductivity of the membrane is varied from 1.38 to 13.8 W/mK.
Fig. 11 (a) Gas response of “TGS-1” with quadruple -Al2O3 dots with for 200 ppm H2 in air, and (b) catalyst temperature and Voff of TGS-1 with the catalyst of 30wt% Pt/ -Al2O3 and Al2O3 dots.
21
22
23
24
25
26
27
28
29
30
31
Tables Table 1. Dimensions used in the 3-D model of the micro-device Parameter
description
Size (m )
AM
Area of membrane (mm)
1.0x2.0
tM
Thickness of membrane
1.0
wTE
Width of thermoelectric, TE, pattern
120
tTE
Thickness of TE pattern
0.5
wPt
Width of Pt heater pattern
60
tPt
Thickness of Pt heater
0.3
DC
Diameter of catalyst
600
tC
Thickness of catalyst
5.0
DD
Diameter of dummy
200
tD
Thickness of dummy
5.0
32
Table 2. Summary of material properties in the model at 300K. Resistance of Pt resistive heater is temperature dependent with the temperature coefficient of resistance (TCR) of 3.80×10−4 /K. Domain
Material
Thermal conductivity (W/mK)
Thermal expansion coefficient (ppm/K) *
Note
Heater
Pt
71.6
8.8*
TCR is applied
Membrane
Si3N4
16-33*
3.3*
*Not used for simulation
Membrane
SiO2
0.8-13.8**
2.8*
**Thermal conductivity is variable parameter
Thermoelectric
SiGe
10.0
0.35*
Substrate
Si
149
2.60*
Catalyst
Al2O3
3.30
7.1*
Table 3. Changes in Voff of micro-TGS for CH4 detection (Sensor-1 and Sensor-2) by deposition of -Al2O3 dots under the applied 6 V. Normal
Single dot
Double dots
Triple dots
Notation Offset voltage Voff (mV) Sensor-1
2.41
2.69
2.38
2.25
Sensor-2
3.57
4.48
4.28
4.02
33
Table 4.
Effects of -Al2O3 dots on simulated Voff of micro-TGS by FEM modeling.
The Seebeck coefficient of SiGe line is assumed to be 0.15mV/K. Temp. B (ºC) 192.5
Temp. A (ºC) 151.9
Calorific value, Q (W) 0.129
Temp. difference, T (ºC) 40.55
Voff (mV) -6.08
1
151.9
151.9
0.132
0.01
0.00
2
116.7
142.9
0.136
-26.20
3.93
Dot No. None
34
S0925-4005(17)30530-0 http://dx.doi.org/doi:10.1016/j.snb.2017.03.113 SNB 22025
To appear in:
Sensors and Actuators B
Received date: Revised date: Accepted date:
14-10-2016 13-3-2017 19-3-2017
Please cite this article as: Tomoyo Goto, Toshio Itoh, Takafumi Akamatsu, Yoshitaka Sasaki, Kazuo Sato, Woosuck Shin, Heat transfer control of microthermoelectric gas sensor for breath gas monitoring, Sensors and Actuators B: Chemicalhttp://dx.doi.org/10.1016/j.snb.2017.03.113 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Submitted to: Sensors and Actuators B: Chemical Heat transfer control of micro-thermoelectric gas sensor for breath gas monitoring Tomoyo Goto1,†, Toshio Itoh1, Takafumi Akamatsu1, Yoshitaka Sasaki1,2, Kazuo Sato2, Woosuck Shin1,* 1National
Institute of Advanced Industrial Science and Technology (AIST), 2266-98
Shimo-Shidami, Moriyama-ku, Nagoya 463-8560, Japan 2Aichi
Institute of Technology, 1247 Yachigusa, Yakusa-Cho, Toyota 470-0392, Japan
†Present
address: The Institute of Scientific and Industrial Research (ISIR), Osaka
University, 8-1 Mihogaoka, Ibaraki 567-0047, Japan *Corresponding author: Dr. Woosuck Shin Phone: +81-52-736-7107, Fax: +81-52-736-7244 E-mail: [email protected] Highlights
Catalyst on the micro-TGS lowers the temperature and induces a biased voltage. Additive process of alumina dots changed the heat balance to controls the voltage. Heat transfer analysis by FEM modeling is performed to control the heat balance.
1
Abstract A micro-thermoelectric gas sensor (micro-TGS) uses thermoelectric voltage induced by the catalytic combustion of hydrogen or methane for selective gas detection in breath. This is accomplished under an elevated temperature using a micro-heater built on the same membrane as a hotplate, which enables selective combustion of the target gas. A temperature differential built by the catalyst dot on the membrane induces the offset voltage (Voff) of the micro-TGS, which limits the amplifier circuit application. In this study, we strived to suppress Voff by an additive integration process of heat dissipation dots prepared using -Al2O3 paste. In this paper, we discuss the effects of these Al2O3 dots on the thermal balance over the micro-TGS membrane by finite element method (FEM) modeling. When the dots were deposited in a symmetrical position to the combustion catalyst, Voff was compensated depending on the size, numbers, and locations of the dots. The micro-TGS heat transfer control by the dots was additionally verified by 3-D FEM modeling. The changes in Voff by the -Al2O3 dots in FEM modeling were greater than those of the experiments, suggesting the high thermal conductivity of the micro-TGS membrane. The deviation of the membrane thermal conductivity due to the process non-uniformity significantly influenced the Voff; however, it was effectively reduced by the additive integration of dots. Keywords: Gas sensor; Catalyst; Thermoelectric device; Micro-electromechanical systems (MEMS); Finite element method (FEM)
2
2. Introduction Medical diagnostic technologies for early disease stages have been widely studied in response to global population aging issues and the expansion of the health care industry. Among these technologies, breath analysis techniques have garnered considerable attention as non-invasive and simple health check methods that can be conducted both at home and in a medical facility. Human breath includes several kinds of inorganic and volatile organic compounds (VOCs). These gases are known as markers of disease and human body metabolism [1,2]. To detect the part-per-million (ppm) levels of breath gases, such as H2, CO, and CH4 [3], we developed a unique micro-thermoelectric gas sensor (micro-TGS) with various combustion catalysts for a breath analysis system. Calorimetric gas sensors take advantage of micro-machined devices using the heat of the chemical reactions generated by catalytic metals. Their thermal characteristics must be optimized for better sensing performance. The micro-TGS uses the catalytic combustion of the gas operated at a catalyst temperature over 100 ºC by using a micro-heater built on the membrane. It thereby shows the linear relationship of the voltage signal to the gas concentration [4,5] and the sensing performance for the ppm level of hydrogen in the breath gas analysis [5]. Sensor performance of micro-TGS, a synergistic combination of the catalytic combustion of gas and thermoelectric conversion (from heat to electrical), can be easily modulated by the choice of the combustion catalyst. Previously, micro-TGS for H2, CH4, and CO detection were reported using Pt/-Al2O3 [4,5], Pd/-Al2O3 [6], and Au/Co3O4 [7] catalysts, respectively. To produce an effective design, the heat balance of the sensor during operation must be estimated. The heat balance can predict the inflammable gas combustion energy of the catalyst. The heat as a function of the sensor output is also important for the development of the catalyst. The membrane area of the micro-TGS is comprised of ends of thermoelectric SiGe forming a symmetrical pattern. However, one end of the SiGe pattern is covered by a 3
ceramic combustor, which induces the asymmetrical temperature distribution in the membrane and results in the temperature differential, as shown in Fig 1. Fig. 1(a) depicts an example of a voltage signal of micro-TGS for detection of CH4 at 325 °C. It indicates a problem of micro-TGS in real applications; i.e., voltage from the sensor from the initial stage in air, named the offset voltage (Voff), occurs at a high temperature operation. The fluctuation or drift of the Voff can also be a serious problem when the target gas concentration is low. The voltage signal was thus gradually elevated after 90 s (Voff90) from the initial stage (Voff1). This difference of Voff is shown in Vdirft in Fig. 1(a). After CH4 detection, the voltage continues to rise, and the offset voltage at 270 s (Voff270) is higher than that of the initial stage (Voff1). The origin of this problem was an initial temperature difference (T) formed between point A (catalyst site) and point B:
VS TAB TA TB
(1)
where is the Seebeck coefficient of the thermoelectric film. This V off of micro-TGS increases in high temperature operation because the combustion catalyst on the membrane plays the role of “heat dissipater” (Fig. 1(b)). In addition, non-uniformity of the membrane in micro-TGS also effects the heat balance inside the membrane. Generation of this voltage cannot be ignored because the voltage signal increases with the increasing operating temperature. The suppression of offset voltage of microTGS is important to stabilize the base voltage and improve the quantitative performance. To make an effective design of each part of micro-TGS, especially the catalyst and micro-heater, meanders are important for optimizing the sensor performance. To make effective design of each parts of micro-TGS, especially a double meander-shaped Pt heater and a circular shape catalyst are important to optimize the sensor performance, and moreover it is important to quantitatively evaluate the heat transfer of these parts. Optimization of the micro-heater can involve thermal distribution in the membrane that is aimed at temperature control of a uniform and stable response at elevated 4
temperatures, thereby improving the sensor reliability and sensitivity. The development of a theoretical model for thermo-mechanical calculations and heat transfer mechanisms is important for the design of catalytic sensors, which can include predicting device process non-uniformity. The influence of the micro-hotplate design on the catalytic reaction or sensor response is not fully understood, the model calculation is expected to be effective tool for the sensor development. Most studies have primarily focused on flowmeters [8], while heat and mass transfer in heated systems [9] and thermochemical devices in gas-fueled combustor systems [10] have additionally been investigated. In the case of micro-TGS, the heat of gas combustion at the catalyst film which is thin and smaller than 1 mm, is extremely difficult to directly measure. In our previous study, however, we have quantitively estimated the catalytic combustion heat (power) required for 1 mV of signal voltage to be 46 μW, using the linear calculation and thermal time constants of the experimental signal curves of the micro-TGS [11]. The design of micro-heater meanders can efficiently and uniformly heat the catalyst; moreover, they can prevent the loss of combustion heat though the heater line. Various heater designs have been tested. The power consumption of the micro-heater meanders was reduced to 30–40 mW at the catalyst temperature of 100 ºC, which is approximately 30% less than that of the first generation one [12]. In addition, an asymmetric micro-heater meander enabled different catalyst temperatures for a single sensor with a double catalyst structure. In this study, to control this heat balance and suppress V off of micro-TGS, the -Al2O3 thick film, dummy dots, for heat dissipation were additively integrated on the membrane of micro-TGS. The effects of the -Al2O3 dots on the heat balance are herein described with the quantitative results estimated by 3-D solid modeling of the finite element method (FEM). The effects of thermal conductivity of micro-TGS on the Voff control by -Al2O3 dots are then discussed. Experimental
5
2.1. Design of micro-TGS Details of the preparation procedure of micro-TGS for H2 and CH4 detection were described in previous reports [3-5]. Here, the p-type B-doped SiGe thermoelectric pattern was used for the micro-TGS device. As shown in Fig. 2, the 4 4 mm2 dimension of the micro-TGS device was comprised of the thermoelectric pattern, Pt heater, and electrode line pattern on a double-sided polished Si substrate. The optimal geometrical design comprised of the most advantageous materials was implemented and integrated on a suspended membrane—a double meander-shaped Pt heater of resistance ca. 100 —to heat the 0.6-mm-diameter catalyst at point A in Fig. 2. Details of the dimensions are given in Table 1. The micro-TGS had an active membrane area (heated area, i.e. covered with the heater and catalyst), which was determined by the mask size of the back-side of the silicon substrate of 2.45×1.40 mm. After the KOH wet etching of Si substrate, the membrane size became ca. 2.0×1.4 mm. It was deviated by the process parameters and non-uniformity of the wet etching process. 2.2. Catalyst deposition process of micro-TGS For H2 detection, the Pt/-Al2O3 catalyst powder was prepared by referencing the method of Nishibori et al. [3]. The Pd/-Al2O3 catalyst powder was prepared for CH4 detection of micro-TGS by referencing the work of Nagai et al. [6]. Pt or Pd colloid solutions were purchased from Tanaka Kikinzoku Kogyo K.K., Tokyo, Japan, respectively. Alumina powders, the -Al2O3 and -Al2O3, Taimicron TM-D and TM-100, are supplied by TAIMEI Chemicals Co., Ltd., Nagano, Japan with the specific surface area of 13.5 and 128 m2/g, respectively. In the case of the preparation of catalyst for H2 sensor, 30 wt% Pt was added to -Al2O3 powder in distilled water. Suspension was stirred at 70 ºC until the water evaporated. Mixed powder was calcined at 300 ºC in an electric furnace for 2 h under air. For the CH4 sensor, the 10 wt% Pd/-Al2O3 catalyst powder was prepared by the same procedure as described above. Then, the calcined powder was dispersed in vehicle containing terpineol, ethyl cellulose, and distilled water (9:1:5 wt% ratio). The catalyst pastes were then deposited by air dispenser
6
(Musashi Engineering, Inc., Tokyo, Japan) on the device. Beside the chemical composition of the catalyst, the viscosity of the paste is important parameter for the deposition process. The viscosity determines not only the achieved size of the drop, the final thickness of the thick film, but also affects the dispensing conditions, air pressure and needles. Actually, in the viscosity level of our catalyst pastes, 0.3 Pa・s at shear rate 357/s, the bigger dot is much controllable than small one [13]. The dispenser robot process very precisely equipped with laser distance meter, Z axis, and a camera system for x-y alignment with the precision below 0.05 mm. After the catalyst deposition, the devices were calcined at 300 ºC for 2 h under air. After that, the devices were mounted on stem and connected using gold-bonding wire to prepare the micro-TGS. 2.3. Evaluation of additive integration process of -Al2O3 dots In the case of alumina dots deposition, micro-manipulator is applied and the sensor device with the catalyst deposited already mounted on stem with wire bonding, while the big catalyst dot is deposited before the chip mounting. The -Al2O3 powder, which did not show the catalyst activity, was used as the heat dissipation film (-Al2O3 dots) in this study. The -Al2O3 powder was mixed in a vehicle and deposited on the membrane of the micro-TGS. Then, the micro-TGS was dried at 90 ºC for 30 h. After drying, in the case of the H2 sensor, the offset voltage (Voff) of the micro-TGS was measured under the applied voltage of 1, 2, and 3 V with air-flowing of 200 cm3/min. After measurement, the deposition of -Al2O3 dots and the measurement were repeated to become closer to Voff = 0 mV. In the case of the microTGS for CH4 detection, the micro-TGS was aged at 6 V of applied voltage for 6 h to stabilize the Voff. After this treatment, similar to the above micro-TGS for H2, effects of the position of -Al2O3 dots on Voff were measured at 6 V of applied voltage. The schematic presentation of this evaluation procedure is shown in Fig. 3. 2.4. Simulation of effects of dummy catalyst by FEM modeling
7
A simulation of the micro-TGS heat balance was performed with the idealized multilayer model constructed and analyzed by FEM using a Femtet (Murata Software Co,. Ltd., Tokyo, Japan) environment. In Fig. 4, the geometry of the simplified structures is depicted. Table 1 shows the size of the micro-TGS device. The material property values for FEM modeling are listed in Table 2, where several values are referenced in terms of previous reports [14]. The membrane material was simplified to SiO, which was actually a multilayer of Si3N4 and SiO2. The mesh size of the finite elements was smaller than 0.1 mm for the Pt line, catalyst, and -Al2O3 dots; smaller than 0.2 mm for SiGe and membrane; and smaller than 0.3 mm for the bulk Si rim. The solver of Femtet for thermal analysis called "Watt" carried out electric field analysis and combine into thermal analysis by calculating current density and the Joule's loss. The final Joule heating steady state stage combines the temperature dependent resistivity of the Pt heater. The thermal expansion of the membrane and the Si substrate has been also simulated but not shown here. 3. Results and discussion 3.1. Evaluation of -Al2O3 dots on changes in Voff of micro-TGS Fig. 5 shows the images of (a) the micro-TGS for CH4 and (b-d) the deposited -Al2O3 dots on this sensor, as well as (e) the offset voltage (Voff) at applied 6 V of micro-TGS, as shown in Figs. 5 (a-d). To make small dot is much controllable than large dots in our experiments, however, this made a few dots rather ellipsoidal shape. We have increased the number of dots step by step to check the heat transfer. To investigate the change of Voff by -Al2O3 dots, the micro-TGS for CH4 with an initial value of Voff = 2.41 mV was used. The -Al2O3 dots were deposited in the order of one (single), two (double), and three (triple) from point B to point A (Figs. 5(b-d)). The Voff of micro-TGS were measured after aging for 6 h. The Voff of micro-TGS changed from 2.41 mV to 2.69, 2.38, and 2.25 mV, as shown in Fig. 5(e) and Table 3. From this result, when the dots were deposited on the opposite side of the combustion catalyst (point B), the Voff of micro-TGS slightly increased. In the case of 8
proximity to the catalyst side (point A), the Voff of micro-TGS slightly decreased. It is considered a voltage is induced by heating the membrane and the catalyst, by the Seebeck effect from the temperature difference developed between the A and B points, and the point A is expected to be lower temperature, from the negative sign of the voltage. When the dot is deposited across the Si substrate and membrane, the dot thermally connected them and thermal insulation around the point B becomes worse and the increase of temperature at B point becomes less. We consider that this explanation can be a reasonable to understand the change of the Voff in this study. This tendency was also observed on other micro-TGS (Sensor-2), as shown in Table 3. These results indicated that the Voff with micro-TGS for CH4 detection could be controlled by -Al2O3 dots. Fig. 6 shows the optical images of the micro-TGS for CH4 and deposited alumina, Al2O3, dots on this sensor (a), and the changes in offset voltage by -Al2O3 dot deposition (b), and the changes of V response of micro-TGS for H2 and CH4 (c). For this sensor of gas detection, large alumina dot is deposited. As discussed above with the results of Fig. 5, the additive deposition of alumina dots lowered the V off effectively, and the initial large positive Voff reduced to around 0 by the first dot, and again to negative value by the second dot. The ΔV of the micro-TGS for gas of CH4 and H2 are not affected as shown in Fig.5(c), showing the detection performance is not disturbed by alumina dots. The ΔV of the micro-TGS for 200 ppm CH4 is around 0.1 mV, which is reasonable value from the results of previous reports [11, 15], which is successful demonstration of additive dot deposition. Fig. 6(d) shows the response of the micro-TGS for CH4 and the response after the dot deposition changed. From this, we could suggest that the combustion around the catalyst is slightly affected by the dots. However, in the case of H2 detection, Fig. 6(e), the responses of the micro-TGS are exactly the same with or without the alumina dots.
9
Fig. 7 shows the optical images of (a-e) the micro-TGS for H2 and deposited -Al2O3 dots on this sensor, and (f) the changes in offset voltage by -Al2O3 dot deposition. From the results of micro-TGS for CH4 (Fig. 5), we strived to control the Voff of the micro-TGS for H2. In this experiment, the micro-TGSs of negative Voff were tested to validate this additive process. Fig. 7(f) shows the change of Voff of micro-TGS by heating with increasing the applied voltage and with four alumina dots (Fig. 7(e)). From the results of Fig. 7(f), the Voff increased around 0.2 mV by first (Fig. 7(b)) and third dots (Fig. 7(d)), and second (Fig. 7(c)) and firth (Fig. 7(e)) dot reduced the V off around 1 mV. These results indicated that the dots effects become larger by cover area and position of dots. By increasing the applied heater voltage, which was roughly proportional to the power of joule heating, the negative Voff became high. However, as the number of dots on the counter side close to point B increased, the negative V off became low. This tendency was additionally confirmed by other micro-TGSs. As shown in Fig. 8, the Voff of four additional micro-TGSs with the applied voltage 3.0 V were modulated by the dots and tested. On the evaluation of H2 sensors, 5 sensors were evaluated for investigate the effect of dots. These sensors were numbered from TGS-1 to TGS-5. To investigate the effects of number of dot deposition, all sensors were deposited four points in the same procedure as shown in Fig. 7. Their results were analogous to the shifting of their Voff close to 0 V. In the case of TGS-3, Voff value become around 0 by deposition of one dots, and Voff of TGS-3 was almost constant regardless of additional deposition. In contrast, TGS-4 and TGS-5 showed around -3 mV after deposition of four dots. A deviation of the Voff of different sensors occurred. It is considered that this deviation originated from the individual differences produced in the fabrication process, such as the membrane size or partial etching during the wet etching process, or the catalyst diameter or thickness during the catalyst deposition process. The controllability of the alumina dot deposition is important and we have controlled the size and thickness of the dots, but their variance easily leads to the deviation. However, in this study, the position of the dots seems much important parameter makes the difference, especially whether the dot is across thin membrane to Si
10
substrate or not. From these results, to control the Voff, produce dots in the most effective location or with similar shape is important. By the additive process of alumina dot deposition, this deviation by as much as -3 to -11 mV was successfully reduced to as small as 0 to -3 mV. 3.2. Simulation of effects of -Al2O3 dots on heat balance of micro-TGS In the FEM modeling, two dummy points were deposited on the opposite side of the catalyst (point B), as shown in Fig. 9. The solver of Femtet for thermal analysis called "Watt" carried out electric field analysis and combine into thermal analysis by calculating current density and the Joule's loss. The final Joule heating steady state stage combines the temperature dependent resistivity of the Pt heater. The thermal expansion of the membrane and the Si substrate has been also simulated but not shown here. The temperature profile of the suspended heater is shown in Fig. 9 for an SiO2 (1 m thick) membrane with a fixed thermal conductivity (10 W/mK), a fixed value of applied voltage of 3 V to the two heater electrodes, and a temperature of the Si substrate bottom of 25°C. The SiGe pattern in the center had the highest temperatures. Meanwhile, the connecting Pt leads remained unheated, and the temperature uniformity in the center heated area was high but low at the corner. The local heating confinement of heat in the high resistive area was demonstrated. The maximum temperature reached by applying 3 V was 221°C without dots and decreased to 202 °C and 181 °C by single dot and double dots, respectively. These effectively transferred the heat generated by the Pt heater joule heating on the membrane to the Si rim. The FEM results in Fig. 9 are summarized in Table 4. Although the amount of changes of Voff by -Al2O3 dots is large compared with that of the experimental results, more details of the device operation can be depicted. Furthermore, the data again indicates that the Voff of micro-TGS is reduced by deposition of -Al2O3 dots. The FEM results also indicated that the Voff of micro-TGS was changed by deposition of -Al2O3 dots and the Voff was negative; that is, the temperature of the catalyst side (point A) 11
was relatively lower than the other parts. Therefore, the micro-TGS of negative Voff could be controlled by the -Al2O3 dots. Details of the thermal resistance analysis by heat transfer evaluation are omitted in this study because the minimization of power consumption was not our goal, but investigating the effect of alumina dots is the objective. The heater power of the TGS is similar to the real heater power of the device, which is around 125 mW, where the calorific value of the device in the simulation varies from 129 to 136 mW. When the heat transfer from the membrane to the Si rim through dots increases, the temperature of point B becomes relatively lower than that of A, and the Voff transitions to a positive value. The simulation results clearly show this effect and most devices showed same trends even though their offset values are deviated as shown in Fig. 8. To take account the thermal resistance of alumina is important, because the thermal resistance is a parameter which express the thermal flow per unit area between two systems, alumina and gas flow. However, the thermal resistance of the membrane including catalyst (alumina particle deposit with Pt particle), which is considered almost same specific surface area [6] to that of alumina dot is estimated experimentally to be 230 K/mW [11] and this value is not changed much (within 1%) from the membrane without the catalyst. 3.3. Adjusting the thermal conductivity of the membrane by comparing FEM and the experimental values of micro-TGS with dots In this simulation, the most important parameter was the thermal conductivity of the membrane of micro-TGS. Its effect was quantitatively investigated by FEM simulation. Previous analysis by FEM was calculated using SiO2 and Si3N4 as the singlelayer membrane. However, in a real sensor device, the membrane consists of multilayer structures comprised of two SiO2 layers and an Si3N4 layer (SiO2 250+250 nm and Si3N4 330 nm, respectively). Therefore, the thermal conductivity (κeff) can be written as the following calculation:
12
eff = SiO2×42.5% + Si3N4×57.5% Using the reported value of SiO2 and Si3N4 [14, 16], eff was estimated to be 2.714 W/mK. However, considering that the thickness of the multi-layer was not uniform, and the wet etching process could remain as Si underneath the Si 3N4 layer, the thermal conductivity of the membrane was unpredictable and difficult to be measured. The changes in Voff with the addition of -Al2O3 dots, as shown in Fig. 4, were simulated. The thermal conductivity of the membrane was set to be varied from 1.38 to 13.8 W/mK. The potential between the electrodes was set to be 2.75 V, which led to a close value of the catalyst temperature of micro-TGS at approximately 125 ºC. Fig. 10 shows the results of the FEM simulation of changes in V off of micro-TGSs based on the number of -Al2O3 dots deposited. The Voff shifts from negative to positive with the addition of -Al2O3 dots, which is analogous to the above results. Their change is reduced by increasing the thermal conductivity of the membrane. In addition, the effects of -Al2O3 dots on Voff slightly decrease with the decrease of thermal conductivity. The width of the Voff change of the TGS without -Al2O3 dots was fairly large at 7.8 mV, which corresponded to theV response of micro-TGS for over 5,000 ppm H2 in air. This width became narrow with the addition of dots to 2.4 and 1.4 mV for single and double dots, respectively. The width of Voff decreased by more than five times as the thermal conductivity of the membrane increased ten times. This confirmed the feasibility of the proposed concept. In comparing these results to the voltages of real sensors, the closest or average value of the thermal conductivity of the multi-layer membrane was around 8 W/mK. Although the amount of change in Voff of the FEM simulation was not particularly comparable to that of the experimental results, the much higher value of the thermal conductivity of the membrane implied the rather poor precision of the wet etching process. This can provide guidance in understanding their effect on the heat transfer in the sensor device. The accuracy of the Voff and, more fundamentally, the temperature 13
distribution in the membrane could be improved by optimizing the parameters of heat dissipation or thermal resistance values in the FEM simulation. The lack of reproducibility of wet etching is considered to be the origin of the deviation. Not only additive alumina dots shown in this study but dry process technique such as deep RIE which can produce much cleaner membrane can overcome this thermal valance problem. Nonetheless, that approach would be beyond the scope of this study. The heat dissipations of the micro hotplates with the various catalyst or sensing materials of thicknesses of 5-10 m have been calculated [17] and there is no clear differences in thermal conduction below 600 ºC . 3.4. Gas response of micro-TGS with dots In the application of breath analysis, operating the sensor at a lower catalyst temperature is critical to ensuring the selectivity. However, the catalyst becomes inactive for long-term operation at a lower temperature. To activate the catalyst, temporal heating, a so-called flashing operation, is necessary before the normal detection state. Fig. 11(a) shows the gas sensing performance of “TGS-1” with quadruple -Al2O3 dots (As shown in Fig. 8) with the catalyst of 30wt% Pt/-Al2O3 for 200 ppm H2 in air. The catalyst of TGS-1 is elevated at the temperature of 100, 125, and 150 °C. The applied voltages and the corresponding Voff are plotted in Fig. 11 (b). The first response of TGS-1 at catalyst temperature 100 °C shows a delayed unsaturated response, suggesting that the catalyst is inactive. To activate it, the temperature is increased to 150 °C and hydrogen gas is again detected. After this point, the catalyst becomes active and then the TGS-1 is operated at 125 °C, when it shows normal detection performance. Although we have controlled the Voff for CH4 sensing, its response was not stable and not shown here. It is considered to contain other problems of instability of catalytic activity, which is currently under investigation. For the specific pursuit of our prototype breath hydrogen detector, the Voff should be less than 0.5 mV and the Voff of micro-TGS is reduced by the dots to satisfy this 14
criterion. The voltage signal Vs for 200 ppm H2 is fairly stable for the use of the prototype detector. Its level is sufficiently high at 0.2 mV for 200 ppm, which is the same as the normal micro-TGS 0.1 mV for 100 ppm at the catalyst temperature of 125 °C, as reported in [18]. 4. Conclusion In this study, the additive process of alumina dots on the membrane of the microTGS was performed. The heat transfer analysis by 3-D FEM modeling was carried out to quantitatively validate the heat valance control. The micro-TGS with and without dots at an operational temperature at 325 ºC and 125 ºC was tested as a proof-of-concept. The voltage was negatively biased because the catalyst lowered the temperature of the surrounding membrane. When the dot was deposited on the opposite side of the combustion catalyst, the negative Voff of micro-TGS was reduced. This tendency was also verified by the subsequent model results of the FEM analysis using the designs of dots on the micro-TGS membrane. The implementation of a membrane composed of low thermal conductive materials, i.e., SiO2 and Si3N4, yielded a reduction of thermal losses to the silicon substrate. It thus improved the thermal efficiency of the system. Nevertheless, the actual thermal conductivity of the membrane material had to be considered to enable the wet etching process to retain the Si. The main heat loss mechanism in the micro-TGS was via the gas phase of air atmosphere. However, the relative temperature change was mainly affected by the thermal valance of heat transfer from the membrane to the Si rim. The heat transfer could be controlled by the dots, which increased the heat transfer to the rim and reduced the temperature of the opposite side. Consequently, the absolute value of V off was reduced. The conductivity of the membrane significantly influenced the Voff, and the deviation of the Voff was thereby strongly affected. However, their width was reduced by the integration of dots.
15
Acknowledgement This work was supported by the Knowledge Hub Aichi (the priority research project: P3-G3-S1) of Aichi Prefecture, Japan. The authors are grateful to AIST Laboratory members, Atsuko Hotta, Michi Yamaguchi, and Takaomi Nakashima for their technical support in the preparation of the devices and sensor tests used in this study. The authors also wish to thank Assoc. Prof. Maiko Nishibori of Kyushu University, Japan for her comments on this study.
References [1] A. Amann, G. Poupart, S. Telser, M. Ledochowski, A. Schmid, S. Mechtcheriakov, Applications of breath gas analysis in medicine, Int. J. Mass Spectrom. 239 (2004) 227-233. [2] Y. Broza, P. Mochalski, V. Ruzsanyi, A. Amann, H. Haick, Hybrid volatolomics and disease detection, Angew. Chem. 54 (2015) 11036-11048 [3] W. Shin, Medical applications of breath hydrogen measurements, Anal.
Bioanal.
Chem. 406 (2014) 3931–3939. [4] M. Nishibori, W. Shin, L. F. Houlet, K. Tajima, T. Itoh, N. Izu, N. Murayama, I. Matsubara, New structural design of micro-thermoelectric sensor for wide range hydrogen detection, J. Ceram. Soc. Japan 114 (2006) 853-856. [5] W. Shin, M. Nishibori, N. Izu, T. Itoh, I. Matsubara, K. Nose, A. Shimouchi, Monitoring breath hydrogen using thermoelectric sensor. Sens. Lett. 9 (2011) 684-687. [6] D. Nagai, M. Nishibori, T. Itoh, T. Kawabe, K. Sato, W. Shin, Ppm level methane detection using micro-thermoelectric gas sensors with Pd/Al2O3 combustion catalyst films, Sens. Actuator B: Chem. 206 (2015) 488-494. [7] M. Nishibori, W. Shin, N. Izu, T. Itoh, I. Matsubara, CO oxidation performance of
16
Au/Co3O4 catalyst on the micro gas sensor device, Catal. Today 201 (2013) 85-91. [8] F. Udrea, J. Gardner, D. Setiadi, J. Covington, T. Dogarua, C. Lua, W. Milne, Design and simulations of SOI CMOS micro-hotplate gas sensors, 78 (2001) 180–190 [9] P. van Malea, M. de Croona, R. Tiggelaarb, A. van den Bergb, J. Schouten, Heat and mass transfer in a square microchannel with asymmetric heating, [10] S. Karagiannidis, K. Marketos, J. Mantzaras, R. Schaeren, K Boulouchos, Experimental and numerical investigation of a propane-fueled, catalytic mesoscale combustor, Catalysis Today 155 (2010) 108–115 [11] D. Nagai, T. Akamastu, T. Itoh, N. Izu, W. Shin, Thermal balance analysis of a microthermoelectric gas sensor using catalytic combustion of hydrogen, Sensors, 14 (2014) 1822-1834 [12] M. Nishibori, W. Shin, N. Izu, T. Itoh, I. Matsubara, Microheater Meander configurations for combustion catalysts in thermoelectric gas sensor, Sens. Lett. 8 (2010) 792-800 [13] W. Shin, M. Nishibori, M. Ohashi, T. Itoh, N. Izu, I. Matsubara, Ceramic catalyst combustors of Pt-loaded-alumina on microdevices, J. Ceram. Soc. Jpn 117 (2009) 659-665 [14] Y.S. Touloukian, Thermophysical Properties of Matter, Vol. 12, Thermal Expansion, IFI/Plenum, New York, 1975. SiN data from; http://accuratus.com/silinit.html. [15] N. Park, T. Akamatsu, T. Itoh, N. Izu, W. Shin,“Calorimetric Thermoelectric Gas Sensor for the Detection of Hydrogen, Methane, and Mixed Gases”, Sensors 14 (2014) 8350-8362 [16] E.Vereshchagina, R. Tiggelaar, R. Sanders, R.Wolters, J. Gardeniers, Low power micro-calorimetric sensors for analysis of gaseous samples, Sens. Actuator B: Chem.
17
206 (2015) 772-787 [17] M. Nishibori, W. Shin, N. Izu, T. Itoh, I. Matsubara, Sensing performance of thermoelectric hydrogen sensor for breath hydrogen analysis, Sens. Actuator B: Chem. 137 (2009) 524–528. [18] F. Biró, C. Dücső, Z. Hajnal, F. Riesz, AE. Pap, I. Barsony, Thermo-mechanical design and characterization of low dissipation micro-hotplates operated above 500 ºC, Microelectronics Journal 45 (2014)1822–1828
18
Biographies Tomoyo Goto received her PhD in crystalline materials science from Nagoya University (2012), Japan. From 2012 to 2014, she was a postdoctoral fellow at Kyushu University in Japan. She worked at the Electroceramics Processing Group, Advanced Manufacturing Research Institute (AMRI), AIST, in Japan from 2014 to 2015. She is currently an assistant professor at ISIR, Osaka University in Japan. Her research interests include bioceramics, purification materials, and gas sensor using combustion catalysts. Toshio Itoh studied organic–inorganic hybrid materials and finished his master’s course in 2002 at Nagoya University in Japan and received a doctorate in 2005 in crystalline materials science from the same institution. He has been a member of the research staff in Electroceramics Group (formerly Electroceramics Processing Group) at Inorganic Functional Materials Research Institute (IFMRI), AIST in Japan since 2005. His research activities include VOC gas sensing devices using organic–inorganic hybrid materials. Takafumi Akamatsu received his PhD in Engineering from Nagoya Institute of Technology, Japan in 2008. From 2008 to 2012, he worked at Nihon Yamamura Glass Co., Ltd. He is currently a research member of Electroceramics Group at IFMRI, AIST, in Japan. His research interests include metal oxide semiconductors for NOx sensor and glass-derived hydrogels for proton conductor. Yoshitaka Sasaki studied mechanical engineering and finished his master’s course in 2016 at Aichi Institute of Technology under the supervision of Prof. K. Sato. He joined AIST group in 2015 for 1 year as a research assistant working for the fabrication and FEM simulation of micro gas sensors. Kazuo Sato received B.S. from Yokohama National University in 1970. He worked with Hitachi, Ltd., during 1970–1994. He received Ph.D. from the University of Tokyo in 1982. He started MEMS research since 1983 in Hitachi. He joined Nagoya University in 1994 as a full professor of MEMS/Micromachining Lab. He became an emeritus in 2012. He joined Aichi Institute of Technology in 2012 as a professor. His research area includes; micromachining, KOH/TMAH etching of Si, etching of Quartz, micro/nano physics of wet anisotropic etching of single crystals, characterization of mechanical properties of thin film Si, micro/nanophysics, microactuators, applied microsystems. Woosuck Shin received his BS and MS in material science and engineering in 1992 and 1994, respectively, at KAIST, Korea. After receiving his PhD in applied chemistry from Nagoya University in 1998, Japan, he has been employed by AIST, Nagoya, Japan. He is currently the group leader of the Electroceramics Group of IFMRI, AIST in Japan. His research interests include a functional thin film for gas sensor technology with micro-systems.
19
Figure captions Fig. 1
(a) Voltage signal (Vs) of micro-TGS for detection of CH4 and (b) schematic illustration of the micro-TGS.
Fig. 2
Optical image of the micro-TGS.
Fig. 3
Schematic presentation of evaluation procedure of offset voltage (Voff) of microTGSs.
Fig. 4 CAD images for simulation model of single and double -Al2O3 dots on microTGS by FEM modeling. Fig. 5 Optical images of (a) the micro-TGS for CH4 sensor, and deposited (b) single, (c) double, and (d) triple -Al2O3 dots on this micro-TGS. (e) The offset voltage signal of micro-TGS with or without -Al2O3 dots. Fig. 6
(a) Optical images of the micro-TGS for CH4 and deposited alumina, -Al2O3, dots on this sensor, (b) Changes in Voff of the micro-TGS with the 10 wt% Pd/Al2O3 catalyst by number of -Al2O3 dot deposition, (c) changes of ΔV against the number of alumina dots of micro-TGS for CH4 and H2 responses, (d) gas response of the micro-TGS with -Al2O3 dots for 200 ppm CH4 and (e) H2 in air.
Fig. 7 Optical images of (a) the micro-TGS for H2 and (b) single, (c) double, (d) triple, and (e) quadruple deposited -Al2O3 dots on this sensor. (f) The changes in the offset voltage (Voff) by -Al2O3 dot deposition. Fig. 8 Changes in offset voltage (Voff) of micro-TGSs by number of -Al2O3 dot deposition. Five sensors were applied at 3 V. Fig. 9
FEM models of heat balance of the micro-TGS device. The surface temperature
20
of the devices of (a) no dot, (b) single, and (c) two -Al2O3 dots on the membrane are shown. Fig. 10
Analysis of the variation of Voff of the micro-TGS device with -Al2O3 dots for thermal valance, where the thermal conductivity of the membrane is varied from 1.38 to 13.8 W/mK.
Fig. 11 (a) Gas response of “TGS-1” with quadruple -Al2O3 dots with for 200 ppm H2 in air, and (b) catalyst temperature and Voff of TGS-1 with the catalyst of 30wt% Pt/ -Al2O3 and Al2O3 dots.
21
22
23
24
25
26
27
28
29
30
31
Tables Table 1. Dimensions used in the 3-D model of the micro-device Parameter
description
Size (m )
AM
Area of membrane (mm)
1.0x2.0
tM
Thickness of membrane
1.0
wTE
Width of thermoelectric, TE, pattern
120
tTE
Thickness of TE pattern
0.5
wPt
Width of Pt heater pattern
60
tPt
Thickness of Pt heater
0.3
DC
Diameter of catalyst
600
tC
Thickness of catalyst
5.0
DD
Diameter of dummy
200
tD
Thickness of dummy
5.0
32
Table 2. Summary of material properties in the model at 300K. Resistance of Pt resistive heater is temperature dependent with the temperature coefficient of resistance (TCR) of 3.80×10−4 /K. Domain
Material
Thermal conductivity (W/mK)
Thermal expansion coefficient (ppm/K) *
Note
Heater
Pt
71.6
8.8*
TCR is applied
Membrane
Si3N4
16-33*
3.3*
*Not used for simulation
Membrane
SiO2
0.8-13.8**
2.8*
**Thermal conductivity is variable parameter
Thermoelectric
SiGe
10.0
0.35*
Substrate
Si
149
2.60*
Catalyst
Al2O3
3.30
7.1*
Table 3. Changes in Voff of micro-TGS for CH4 detection (Sensor-1 and Sensor-2) by deposition of -Al2O3 dots under the applied 6 V. Normal
Single dot
Double dots
Triple dots
Notation Offset voltage Voff (mV) Sensor-1
2.41
2.69
2.38
2.25
Sensor-2
3.57
4.48
4.28
4.02
33
Table 4.
Effects of -Al2O3 dots on simulated Voff of micro-TGS by FEM modeling.
The Seebeck coefficient of SiGe line is assumed to be 0.15mV/K. Temp. B (ºC) 192.5
Temp. A (ºC) 151.9
Calorific value, Q (W) 0.129
Temp. difference, T (ºC) 40.55
Voff (mV) -6.08
1
151.9
151.9
0.132
0.01
0.00
2
116.7
142.9
0.136
-26.20
3.93
Dot No. None
34