Multifunctional Platform with CMOS-Compatible Tungsten ... - MDPI

Article

Multifunctional Platform with CMOS-Compatible Tungsten Microhotplate for Pirani, Temperature, and Gas Sensor Jiaqi Wang * and Jun Yu Received: 26 July 2015 ; Accepted: 20 October 2015 ; Published: 28 October 2015 Academic Editor: Ching-Liang Dai Key Laboratory of Liaoning for Integrated Circuits Technology, School of Electronic Science and Technology, Dalian University of Technology, Dalian 116024, China; [email protected] * Correspondence: [email protected]; Tel.: +86-411-8470-6184; Fax: +86-411-8470-6706

Abstract: A multifunctional platform based on the microhotplate was developed for applications including a Pirani vacuum gauge, temperature, and gas sensor. It consisted of a tungsten microhotplate and an on-chip operational amplifier. The platform was fabricated in a standard complementary metal oxide semiconductor (CMOS) process. A tungsten plug in standard CMOS process was specially designed as the serpentine resistor for the microhotplate, acting as both heater and thermister. With the sacrificial layer technology, the microhotplate was suspended over the silicon substrate with a 340 nm gap. The on-chip operational amplifier provided a bias current for the microhotplate. This platform has been used to develop different kinds of sensors. The first one was a Pirani vacuum gauge ranging from 10´1 to 105 Pa. The second one was a temperature sensor ranging from ´20 to 70 ˝ C. The third one was a thermal-conductivity gas sensor, which could distinguish gases with different thermal conductivities in constant gas pressure and environment temperature. In the fourth application, with extra fabrication processes including the deposition of gas-sensitive film, the platform was used as a metal-oxide gas sensor for the detection of gas concentration. Keywords: multifunctional platform; microhotplate; Pirani sensor; temperature sensor; gas sensor

1. Introduction Recently, complementary metal oxide semiconductor (CMOS) compatible sensors have been widely used in the detection of pressure, temperature, acceleration, and chemical leakage [1–4]. With better control and optimization of CMOS technology, the manufacturing yield of the sensors can be dramatically increased [5]. The noise level of the sensor would be decreased by integrating the sensor and the signal processing circuit in one chip [6]. With the fast development of various kinds of sensors, the idea of multifunctional platforms for the different sensors is generated [7]. With these multifunctional platforms, different kinds of sensors can be implemented, which would save a lot of development time and cost. The microhotplate is a good multifunctional platform. It includes the thin membrane for pressure measurement, the heater and thermister for thermal-related applications, and the free-standing structure with excellent thermal isolation for high temperature platforms. The microhotplate fabricated by microelectromechanical systems (MEMS) technology has been widely used for gas pressure sensors, gas sensors, chemical sensors, flow sensors, and accelerometers [8–13]. It could also be implemented in the standard CMOS process for high yield and low cost. The most important considerations for the CMOS-compatible microhotplate fabrication include the CMOS-compatible materials and the post-CMOS process. The materials of microhotplate mainly

Micromachines 2015, 6, 1597–1605; doi:10.3390/mi6111443

www.mdpi.com/journal/micromachines

Micromachines 2015, 6, 1597–1605

include the heating material and the thermal insulation material. Many different heating materials are available for microhotplate, such as polysilicon [14], aluminum (Al) [15], platinum (Pt) [16], and tungsten (W) [17]. Platinum has been considered the best material for the microhotplate, but it is not the material in the standard CMOS process. Aluminum and polysilicon cannot endure high currents and high temperatures. Tungsten is suggested to be the perfect heating and temperature measurement material for microhotplate due to its properties of high melting point, large temperature coefficient, low cost, and CMOS compatibility. The SiO2 and Si3 N4 as the dielectric layer in the CMOS process are commonly used as the thermal insulation layer for the microhotplate. The post-CMOS processes for the microhotplate mainly include the etching process to suspend the microhotplate from the substrate. Alkaline solutions are used to etch silicon or polysilicon to suspend the microhotplate. The fact that most of the alkaline solutions attack the aluminum bonding pads should be considered. One way is to protect the aluminum pads with some metal, such as Pt, against alkaline corrosion [18]. The improved tetramethylammonium hydroxide (TMAH) has also been used as an effective method to remove the silicon without damaging the aluminum pads [19,20]. In this paper, tungsten was employed as the heating and temperature measurement material for the microhotplate. An improved TMAH solution was used to remove the polysilicon sacrificial layer to suspend the microhotplate. The multifunctional platform for sensors consisted of the tungsten microhotplate and the on-chip operational amplifier. A bias current for the microhotplate was supplied by the operational amplifier. The platform for sensors was implemented in 0.5 µm standard CMOS process. It has been employed to develop a Pirani gas pressure, temperature, and thermal-conductivity gas sensor based on the fact that the gaseous thermal conduction is mainly affected by gas pressure, gas temperature, and the type of the gas. Besides those applications, with a little extra fabrication process, including deposition of the gas-sensing film on the microhotplate, this platform could be used as the heating component for a metal-oxide gas sensor. 2. Experimental Section 2.1. Fabrication Process of Tungsten Microhotplate Tungsten has been traditionally used as a plug material to form via pathways between various metal layers due to its ability to uniformly fill the high aspect ratio vias when deposited by chemical vapor deposition (CVD) methods. In our design, the tungsten microhotplate was fabricated in 0.5 µm standard CMOS process, which featured two polysilicon layers (Poly1 and Poly2) and two aluminum interconnection layers (Metal1, Metal2). The metal plug between Metal1 and Metal2 was tungsten. Dielectric layers surrounding the metal layers were phosphosilicate glass (PSG). The Si3 N4 was more stable than PSG so it was used as the passivation layer on the surface to protect the chip. In the design of the tungsten microhotplate, tungsten was in the form of serpentine resistor instead of via plug. The anchors of the tungsten resistor were connected to Metal2, leaving Metal1 unconnected. A 0.34 µm thick Poly2 was used as a sacrificial layer below the tungsten microhotplate. The etch windows of the tungsten microhotplate were opened during the bonding-pad patterning process in a standard CMOS process, as shown in Figure 1a. Figure 1b shows that the etch windows for the microhotplate and the bonding pads were etched simultaneously during the bonding pad dry-etching in the CMOS process until Poly2 was exposed. Lastly, in our lab, Poly2 was removed to suspend the tungsten microhotplate by an improved TMAH etchant which avoided etching the exposed aluminum bonding pads. Before starting the etching process, the aluminum layer was sputtered at the backside of the die to keep etching substrate from the backside. It took about 8 h to remove the sacrificial layer with the improved TMAH etchant in 85 ˝ C, as shown in Figure 1c [21]. 2.2. Multifunctional Platform with Microhotplate and Operational Amplifier The constant current circuit was adopted to bias the microhotplate [22], as shown in Figure 2. R1 was the tungsten resistor of the microhotplate and R2 was the reference resistor in the substrate.

1598


Micromachines 2015, 2, 6015, In Figure V6, page–page was an ref

external reference voltage and Equation (1) defined the output

Micromachines 2015, voltage V out : 6015, 6, page–page

RR Vout “ VVref p1 +` 11)q = ref (1 out RRR2 = Vout Vref (1 + 1 2) R2

Bonding Pad (etch-window) Bonding Pad (etch-window) Patterning Etching Bonding Pad (etch-window) Bonding Pad (etch-window) Patterning Etching

P-Si

(b)

(a)

P-Si

(a) Photoresist Passivation

PSG

Aluminum

(b) Tungsten

(1) (1)

(1) Sacrificial Layer Etching Sacrificial Layer Etching

P-Si

(c)

P-Si Poly2

(c) Poly1

Tungsten Poly1 Photoresist the Passivation PSGstructure Aluminumthe tungsten Poly2 Figure Figure 1. 1. Fabrication Fabrication of of the freestanding freestanding structure of of the tungsten microhotplate. microhotplate. (a) (a) Opening Opening of of the the etch windows for the tungsten microhotplate; (b) Dry-etching process to expose the sacrificial Figure 1. Fabrication oftungsten the freestanding structure the tungsten microhotplate. oflayer the etch windows for the microhotplate; (b)ofDry-etching process to expose(a) theOpening sacrificial layer for (c) the microhotplate the removal of etch windows for microhotplate; the tungsten microhotplate; (b)of process to exposeby for the the tungsten tungsten microhotplate; (c) Suspending Suspending ofDry-etching the tungsten tungsten microhotplate bythe thesacrificial removal layer of the the sacrificial layer. for the tungsten microhotplate; (c) Suspending of the tungsten microhotplate by the removal of the sacrificial layer.

sacrificial layer.

R1 R1

R2 R2V

ref

Vref

+ +

Vout Vout

Figure 2. Constant current circuit for the microhotplate. Figure Figure2.2.Constant Constantcurrent currentcircuit circuitfor forthe themicrohotplate. microhotplate.

When the Vref was set to constant, the voltage in the minus input of the amplifier was also constant. 2 was in the whose temperature considered equal to thewas ambient When R the Vref was set substrate to constant, the voltage in the was minus input of the amplifier also When the V was set to constant, the voltage in the minus input of the amplifier was ref temperature. During application of the sensor, the ambient temperature would be controlled to be constant. R2 was in the substrate whose temperature was considered equal to the ambient also constant. R was in the substrate whose temperature was considered equal to the ambient 2 as stable as During possible,application so R2 wasofnearly constant the current through R2 was also considered temperature. the sensor, theand ambient temperature would be controlled to be temperature. During application of theequal sensor, ambient temperature would becurrent controlled to ideal be as The currentso through was to the the current of R2 since the input of the asconstant. stable as possible, R2 wasR1nearly constant and the current through R2 was also considered stable asThe possible, soThat R nearly constant and thecurrent current alsoincurrent considered constant. 2 was 2 was amplifier wascurrent zero. temperature compensation technique wasRrequired order of to the avoid the constant. through R1 was equal to the ofthrough R2 since the input ideal The current through R was equal to the current of R since the input current of the ideal amplifier 2 1 influencewas of the ambient temperature compensation fluctuation. Fortechnique the constant circuit, R2 was also used amplifier zero. That temperature was current required in order to avoid the was zero. That temperature compensation technique was required in order to avoid the influence as the temperature compensation resistor. Considering the influence of ambient temperature influence of the ambient temperature fluctuation. For the constant current circuit, R2 was also used of the the temperature ambient temperature fluctuation. ForConsidering the current circuit,ofR2ambient was alsotemperature used as the the output voltage Vout was defined by constant Equation (2):influence asfluctuation, compensation resistor. the temperature compensation resistor. Considering the influence of ambient temperature fluctuation, fluctuation, the output voltage Vout was defined by Equation (2): R1 (1 + α1∆T ) the output voltage V out was defined by Equation (2): = Vout Vref [1 + ] (2) RR1 2(1(1++αα1∆2 ∆ T T) ) ] (2) R p1 ` α ∆Tq Vout “ Vref r1 `R2 (11 + α 2 ∆T1) s (2) where α1 and α2 are the temperature coefficient for resistors in the microhotplate (R 1 ) and in the R2 p1 ` α2 ∆Tq = Vout Vref [1 +

substrate (R2),αrespectively; ΔT was the ambient for temperature fluctuation. If R1 and(R R12) adopted the where α1 and 2 are the temperature coefficient resistors in the microhotplate and in the where α and α are the temperature coefficient for resistors in the microhotplate (R ) and in the 2 1 1 same material, α1 was equal α2. the Thus, R2 was also a tungsten resistor, but in the substrate, substrate (R2), respectively; ΔTtowas ambient temperature fluctuation. If R 1 and R2 adopted theto substrate (R ), respectively; ∆T was the ambient temperature fluctuation. If R and R adopted 2 αinfluence 2 compensate the of the variation the substrate. In this Equation same material, 1 was equal to αtemperature 2. Thus, R2 was also aoftungsten resistor, butsituation, in 1the substrate, to(2) the same material, α was equal to α . Thus, R was also a tungsten resistor, but in the substrate, 2 2 1 would change to Equation and the variation output voltage was not In influenced by the ambient compensate the influence of the(1), temperature of the substrate. this situation, Equation (2) to compensate the influence of the temperature variation of the substrate. In this situation, temperature would changefluctuation. to Equation (1), and the output voltage was not influenced by the ambient Equation (2) 3would change Equation (1),scanning and the output voltage was not influenced by the ambient Figure shows the tooptical and electron microscope (SEM) pictures of the temperature fluctuation. temperature fluctuation. multifunctional sensorthe platform. tungsten microhotplate with a square area of 60 μm ×of60 the μm Figure 3 shows opticalThe and scanning electron microscope (SEM) pictures Figure 3 shows the optical and scanning electron microscope (SEM) pictures of the was suspended by four beams which had a length of 30 μm and a width of 15 μm. The width of the multifunctional sensor platform. The tungsten microhotplate with a square area of 60 μm × 60 μm multifunctional sensor0.8 platform. tungsten microhotplate with a square area 60 gap µm ˆ 60 µm tungsten resistor μm. which The The thickness of the microhotplate about μm.of The was suspended bywas four beams had a length of 30 μm and a was width of 154 μm. The widthbetween of the was suspended by four beams which had a length of 30 µm and a width of 15 µm. The width as of the the the tungsten microhotplate substrate 0.34 μm according to the 4thickness Poly2 tungsten resistor was 0.8 μm.and Thethe thickness of was the microhotplate was about μm. Theofgap between tungsten resistor was 0.8 µm. The thickness of the microhotplate was about 4 µm. The gap between sacrificial The end and of the substrate sacrificialwas layer removal process couldof be identified the tungstenlayer. microhotplate 0.34(Poly2) μm according to the thickness Poly2 as the the tungsten microhotplate and the substrate was 0.34 µm according to the thickness of Poly2 as the accordinglayer. to the The colorend difference the Poly1 layer concentric square after Poly2 removal, sacrificial of theofsacrificial (Poly2) removal process could asbecompared identifiedin Figure 3a,b. Figure shows the SEM picture of thesquare microhotplate. 3d as shows the SEM according to the color3c difference of the Poly1 concentric after Poly2Figure removal, compared in picture of the microhotplate platform, consisting of the tungsten microhotplate array, the reference Figure 3a,b. Figure 3c shows the SEM picture 1599 of the microhotplate. Figure 3d shows the SEM resistors, andmicrohotplate the constant current circuit with theofon-chip operational amplifier.array, the reference picture of the platform, consisting the tungsten microhotplate resistors, and the constant current circuit with the on-chip operational amplifier. 3


sacrificial layer. The end of the sacrificial layer (Poly2) removal process could be identified according to the color difference of the Poly1 concentric square after Poly2 removal, as compared in Figure 3a,b. Figure 3c shows the SEM picture of the microhotplate. Figure 3d shows the SEM picture of the microhotplate platform, consisting of the tungsten microhotplate array, the reference resistors, and the constant current circuit with the on-chip operational amplifier. Micromachines 2015, 6015, 6, page–page

Tungsten Microhotplate

Poly2 Layer

20 μm

Poly1 Stripe

(a)

20 μm

(b)

(c)

(d) Figure 3. Pictures of the multifunctional sensor platform. (a) Optical microscopy picture of the

Figure 3. Pictures of the multifunctional sensor platform. (a) Optical microscopy picture of the tungsten microhotplate before etching; (b) Optical microscopy picture of the tungsten microhotplate after tungsten microhotplate before etching; (b) Optical microscopy picture of the tungsten microhotplate etching; (c) scanning electron microscope (SEM) picture of the tungsten microhotplate; (d) SEM afterpicture etching; (c) scanning electron microscope (SEM) picture of the tungsten microhotplate; (d) SEM of the platform for sensors. picture of the platform for sensors.

3. Results and Discussions

3. Results andon Discussions Based the microhotplate platform, different kinds of sensors were developed. Without any extra processes, thisplatform, platform different was developed a Pirani were gas pressure sensor, a gas any Basedfabrication on the microhotplate kinds for of sensors developed. Without temperature sensor, and a thermal-conductivity gas sensor. With a few fabrication processes, extra fabrication processes, this platform was developed for a Pirani gas pressure sensor, a gas including the deposition of aluminum electrodes and sensitive chemical film, the platform was used temperature sensor, and a thermal-conductivity gas sensor. With a few fabrication processes, as the metal-oxide gas sensor. Thus, experiments were carried out in four aspects. First, for the same including the deposition of aluminum electrodes and sensitive chemical film, the platform was used type of gas, the gas pressure was measured when the gas temperature was constant. Second, for the as the metal-oxide gasthe sensor. Thus, experiments were when carried four aspects. First, for the same same type of gas, gas temperature was measured theout gasinpressure was constant. Third, type the of gas, the gas pressure was measured when the gas temperature was constant. Second, for the responses to the different types of gases with different thermal conductivities were introduced sameattype of gas, the gas temperature was measured when the gas pressure was constant. Third, the constant temperature and gas pressure. Fourth, after deposition of sensitive material on the the microhotplate, the concentration the alcohol gas was detected the platform.were All the results responses to the different types ofofgases with different thermalwith conductivities introduced at showed that the microhotplate an effective platform to deposition develop a gas a gas the constant temperature and gaswas pressure. Fourth, after of pressure sensitivesensor, material on the temperaturethe sensor, a thermal-conductivity gas sensor, and a metal-oxide microhotplate, concentration of the alcohol gas was detected with gas the sensor. platform. All the results

showed that the microhotplate was an effective platform to develop a gas pressure sensor, a gas 3.1. Measurement of Pirani Gas Pressure Sensor temperature sensor, a thermal-conductivity gas sensor, and a metal-oxide gas sensor. The working principle of the Pirani gas pressure sensor is based on the fact that the heat loss of

the hotplate toofits substrate through Sensor gas conduction is a function of the gas thermal conductivity. 3.1. Measurement Pirani Gas Pressure

It is widely used in rough vacuum measurement. In the experiment, the platform configured as the The working principle of the Pirani gas pressure sensor is based on the fact that the heat loss Pirani sensor was put into the vacuum chamber. The temperature in the chamber was 30.3 °C, as of the hotplatebytoa its substrate gas conduction is a function the gas thermal conductivity. measured PT100 thermalthrough resistor (Hayashi Denko, Tokyo, Japan). of The bias current through the It is microhotplate widely used was in rough vacuum measurement. In the experiment, configured as 3.85 mA. The measurement procedure was divided intothe twoplatform steps. First, the gas the Pirani sensor was put into vacuumtochamber. The temperature the chamber was 30.3 ˝ C, pressure was decreased fromthe atmosphere 10−1 Pa. Second, when the gasinpressure of the vacuum system was 1 Pa,thermal the pumpresistor was isolated, then the air slowly bledJapan). into the chamber the desired as measured bybelow a PT100 (Hayashi Denko, Tokyo, The biastocurrent through pressure, and the real-time output voltage of the circuit was recorded with a personal computer using the A/D card. Figure 4 shows the typical response curve of the sensor to the gas pressure 1600 ranging from 10−1 to 105 Pa, which made the sensor platform suitable for the Pirani vacuum sensor. 4


the microhotplate was 3.85 mA. The measurement procedure was divided into two steps. First, the gas pressure was decreased from atmosphere to 10´1 Pa. Second, when the gas pressure of the vacuum system was below 1 Pa, the pump was isolated, then the air slowly bled into the chamber to the desired pressure, and the real-time output voltage of the circuit was recorded with a personal computer using the A/D card. Figure 4 shows the typical response curve of the sensor to the gas pressure ranging from 10´1 to 105 Pa, which made the sensor platform suitable for the Pirani Micromachines vacuum sensor.2015, 6015, 6, page–page 4400

gas pressure rise gas pressure drop

Output Voltage(mV)

4200 4000

Chamber Temperature=30.3 Bias Current=3.85 mA

3800

℃

3600 3400 3200

0

20000 40000 60000 80000 100000 gas pressure(Pa)

Figure 4. The response of the sensor platform to the gas pressure from 10−1 to 105 Pa.

Figure 4. The response of the sensor platform to the gas pressure from 10´1 to 105 Pa.

According to the measurement results, the whole rough vacuum range could be divided into According to the theconstant whole current rough vacuum rangesensitivities could be divided several ranges. Themeasurement Pirani sensor results, biased by had different to each into vacuum range, shown in Table 1. In by the constant high vacuum fromhad 0.1 to 4 Pa, thesensitivities sensitivity oftothe sensor several ranges. TheasPirani sensor biased current different each vacuum 4 to 105 Pa, the sensitivity of sensor decreased to was 1.8 mV/Pa, while in the low vacuum from 10 range, as shown in Table 1. In the high vacuum from 0.1 to 4 Pa, the sensitivity of the sensor 4 to 105current mV/Pa.while Thus, in thethe Pirani biased had high sensitivity in was 0.0025 1.8 mV/Pa, low sensor vacuum fromby10constant Pa, thecircuit sensitivity of sensor decreased high vacuum.

to 0.0025 mV/Pa. Thus, the Pirani sensor biased by constant current circuit had high sensitivity in high vacuum. Table 1. The sensitivity of the Pirani sensor in different vacuum ranges. Vacuum Range Average Sensitivity Table 1. The sensitivity of the Pirani sensor in different vacuum ranges. 0.1–4 Pa 1.8 mV/Pa 4–100Range Pa 0.7 mV/Pa Vacuum Average Sensitivity 100–1000 Pa 0.28 mV/Pa 0.1–4 Pa 1.8 mV/Pa 1000–3500 Pa 0.081 mV/Pa 4–100 Pa 0.7 mV/Pa 3500–10,000 0.013 100–1000 PaPa 0.28mV/Pa mV/Pa 10,000–100,000 0.0025 1000–3500 PaPa 0.081mV/Pa mV/Pa 3500–10,000 Pa 0.013 mV/Pa 3.2. Measurement of Temperature Sensor 10,000–100,000 Pa 0.0025 mV/Pa Thermal conduction of gas is also influenced by the gas temperature, which is the principle of

3.2. Measurement of Temperature Sensor the temperature sensor discussed in detail in our previous research [23]. The gas in the gap between the microhotplate and the substrate was employed as the sensory component for the temperature Thermal conduction of gas is also influenced by the gas temperature, which is the principle sensor. In the test, the platform configured as the temperature sensor was placed in the of the temperature sensor discussed in detail in our previous research [23]. The gas in the gap temperature-controlled oven filled with air at atmospheric pressure and the temperature was between the microhotplate and the was the sensory for the adjusted each Celsius degree fromsubstrate −20 to 70 °C.employed The outputasvoltage for thecomponent corresponding temperature sensor. In the test, platform The configured as theresults temperature sensor placed temperature was recorded by the a multimeter. measurement were shown in was Figure 5. Thein the temperature-controlled oven filled with atmospheric pressure andtemperature the temperature was was adjusted sensor showed a good linearity over air the at temperature range, and the coefficient ˝ 0.303% when the bias current through the microhotplate was 3.85 mA. each Celsius degree from ´20 to 70 C. The output voltage for the corresponding temperature was withoutThe themeasurement etching processresults and suspended gap have also been tested. For these recorded The by asamples multimeter. were shown in Figure 5. The sensor showed a samples, since there was no suspended gap, the tungsten resistor in the microhotplate was alsothe in bias good linearity over the temperature range, and the temperature coefficient was 0.303% when the silicon substrate. Therefore, as shown in Figure 2, both R1 (the resistors in microhotplate) and R2 current through the microhotplate was 3.85 mA. (the resistors in substrate) had the same response to the temperature changes. According to The samples without the etching process and suspended gap have also been tested. For these Equation (1), the output voltage had little response to the temperature change, measured about samples, there from was −20 no to suspended gap, the resistor inerror, the microhotplate was also in 1 mVsince fluctuation 70 °C, probably due tungsten to the measurement noise, and temperature the silicon Therefore, as as shown R1 (thedifference resistors for in microhotplate) drift ofsubstrate. the operational amplifier, shownininFigure Figure2, 5. both The obvious the response of and the microhotplate with and without the suspended gap was the result of the temperature-related thermal conduction of the gas [23]. As the1601 sensory component was the gas, the inflatable sealed package filled with N2 or inert gas would be adopted as the ideal package for the temperature sensor.


R2 (the resistors in substrate) had the same response to the temperature changes. According to Equation (1), the output voltage had little response to the temperature change, measured about 1 mV fluctuation from ´20 to 70 ˝ C, probably due to the measurement error, noise, and temperature drift of the operational amplifier, as shown in Figure 5. The obvious difference for the response of the microhotplate with and without the suspended gap was the result of the temperature-related thermal conduction of the gas [23]. As the sensory component was the gas, the inflatable sealed package filled with Micromachines N2 or inert2015, gas6015, would be adopted as the ideal package for the temperature sensor. 6, page–page Micromachines 2015, 6015, 6, page–page

Output Voltage(mV) Output Voltage(mV)

3490 3480 Temperature Coefficient=0.303% 3490 3470 3480 Temperature Coefficient=0.303% 3460 3470 With suspended gap 3450 3460 3440 With suspended gap 3450 3430 3440 3420 3430 3410 3420 Without suspended gap 3400 3410 3390 Without suspended gap 3400 3380 3390 3370 3380 -20 -10 0 10 20 30 40 50 60 70 80 3370 Temperature( ℃) -20 -10 0 10 20 30 40 50 60 70 80 Figure 5. The response of the sensor platform −20 to 70 °C. ℃)from Figure 5. The response of theTemperature( sensor platform from ´20 to 70 ˝ C.

Figure 5. The responseGas of the sensor platform from −20 to 70 °C. 3.3. Measurement of Thermal-Conductivity Sensor

3.3. Measurement of Thermal-Conductivity Gas Sensor

For the gases with different thermal conductivities, this platform also had the response due to 3.3. Measurement of Thermal-Conductivity Gas Sensor For gases thermal with different thermal conductivities, this platform had out the response due to the the the different conduction of each gas. The experiment wasalso carried at atmospheric For the gases with different thermal conductivities, this platform also had the response due to different thermal conduction of each gas. including The experiment was(ethanol), carried out atmospheric pressure, pressure, in 25 °C. The various gases, CH3CH2OH CHat 3COOH (acetic acid), ˝ the different thermal conduction of each gas. The experiment was carried out at atmospheric 3 (acetone), NHincluding 3 (ammonia), (formaldehyde) mixed (acetic with air, respectively, in 25 CH C.3COCH The various gases, CHand CH3 COOH acid), CH3 COCH3 3 CHHCHO 2 OH (ethanol), pressure, inslowly 25 °C.toThe various gases, including CH3aCH 2OH (ethanol), CH3COOH (acetic acid), were bled the platform in the chamber with syringe. Figure 6 presents the responses of (acetone), NH (ammonia), and HCHO (formaldehyde) mixed with air, respectively, were bled slowly 3 CH 3 COCH 3 (acetone), NH 3 (ammonia), and HCHO (formaldehyde) mixed with air, respectively, the platform to the different gases. The responses to the organic gases were similar due to their to the platform in the chamber with a syringe. Figure 6 presents the responses of the platform to were bledthermal slowly conductivity. to the platform the chamber with aconductivity syringe. Figure 6 presents the responses of similar NHin 3 had larger thermal compared to the organic gases, the different gases. The responses to the organic gases were similar due to their similar thermal the the different responses to the decreased organic gases similar duevoltages to their soplatform it had an to obvious response.gases. WhenThe the gas concentration over were time, the output conductivity. NH3 had larger thermal conductivity compared to the organic gases, so it had an similar conductivity. NH3 to had thermal conductivity compared the organic of the thermal sensor gradually got closer thelarger conditions in the air's atmosphere. This to experiment wasgases, not obvious response. When the gas concentration decreased over time, the output voltages ofsensor, the sensor soso it had an obvious response. the gas concentration over time, the output precise; nevertheless, the When platform could be used fordecreased the thermal-conductivity gas voltages gradually got for closer toHe the conditions inconditions the air's This experiment was not was so precise; ofespecially the sensor gradually got closer due to the in the air's atmosphere. This experiment not H2 or detection to their high atmosphere. thermal conductivity [24,25]. nevertheless, platform could be used for the be thermal-conductivity gas sensor, especially for H2 or so precise;the nevertheless, the platform could used for the thermal-conductivity gas sensor, He especially detection for dueHto their high3400 thermal [24,25]. 2 or He detection due toconductivity their high thermal [24,25]. Temperature = 25conductivity °C Gas Pressure= 105 Pa

Output Voltage (mV) Output Voltage (mV)

3390 3400

Temperature = 25 °C Gas Pressure= 105 Pa

3380 3390 3370 3380 3360 3370 3350 3360

3350

0

200

400

600

800

CH3COCH3 CH3COOH CH3CH2OH CH3COCH3 HCHO CH3COOH NH 3 CH3CH2OH HCHO NH3 1200 1000

1400

Time (s)

Figure 6. The response of the platform different including 0 200 to the 400 gases 600with800 1000 thermal 1200 conductivities, 1400 ethanol, acetic acid, acetone, ammonia, and formaldehyde mixed with air at atmosphere pressure, Time (s) in 25 °C, respectively. Figure 6. The response of the platform to the gases with different thermal conductivities, including Figure 6. The response of the platform to the gases with different thermal conductivities, including ethanol, acetic of acid, acetone, ammonia, and formaldehyde mixed with air at atmosphere pressure, 3.4. Measurement Metal-Oxide Gas Sensor ethanol, acetic acid, acetone, ammonia, and formaldehyde mixed with air at atmosphere pressure, in in 25 °C, respectively. ˝

25 C,The respectively. metal-oxide gas sensor requires working in high temperature for better performance. This platform could ofbeMetal-Oxide used as the 3.4. Measurement Gasheating Sensor plate for metal-oxide gas sensors with extra chemically sensitive material on the surface of the microhotplate [6,9,26]. In our experiment, two aluminum 1602 The metal-oxide gas sensor in high temperature betterprocess. performance. This electrodes were fabricated on therequires surface working of the microhotplate during thefor CMOS Thin SnO 2 platform could be used as the heating plate for metal-oxide gas sensors with extra chemically film was sputtered on the microhotplate in our lab as the gas-sensitive material, as shown in Figure 7a. sensitive material of the microhotplate [6,9,26]. In our experiment, two aluminum The resistance of on thethe SnOsurface 2 film was measured as 0.3 MΩ. During the gas detection measurement, electrodes were fabricated on the surface of the microhotplate during the CMOS process. Thin SnO2


3.4. Measurement of Metal-Oxide Gas Sensor

SnO2 Film

Electrode

Resistance Changement (%)

The metal-oxide gas sensor requires working in high temperature for better performance. This platform could be used as the heating plate for metal-oxide gas sensors with extra chemically sensitive material on the surface of the microhotplate [6,9,26]. In our experiment, two aluminum electrodes were fabricated on the surface of the microhotplate during the CMOS process. Thin SnO2 film was sputtered on the microhotplate in our lab as the gas-sensitive material, as shown in Figure 7a. The Micromachines 2015, 6015, 6, page–page resistance of the SnO2 film was measured as 0.3 MΩ. During the gas detection measurement, the sensor waswas placed in a 50 chamber, and ethanol was injected into the chamber. The resistance the sensor placed in La gas 50 L gas chamber, and ethanol was injected into the chamber. The ˝ C with the of the sensor was recorded by the multimeter. The sensing film was heated up to 300 resistance of the sensor was recorded by the multimeter. The sensing film was heated up to 300 °C microhotplate platform, and resistance changeschanges of the SnO ethanol were were measured after 2 film with the microhotplate platform, and resistance of the SnO2tofilm to ethanol measured one week of preheating. Figure 7b shows the typical sensing response curve to ethanol. The gas after one week of preheating. Figure 7b shows the typical sensing response curve to ethanol. The sensor had had a good response to ethanol, and the of the of sensor increased with the ethanol gas sensor a good response to ethanol, andresistance the resistance the sensor increased with the concentration increasing from 30 to 2000 ppm. ethanol concentration increasing from 30 to 2000 ppm.

100 80 60 40 20

10 (a)

100 1000 Ethanol concentration (ppm) (b)

Figure Figure 7.7. Microhotplate Microhotplateplatform platformconfigured configuredas asaametal-oxide metal-oxide gas gas sensor sensor with with SnO SnO22-sensitive -sensitivefilm film and and the the response response to to ethanol. ethanol. (a)(a)Scanning Scanningelectron electronmicroscope microscope(SEM) (SEM) picture picture ofof the the tungsten tungsten microhotplate microhotplate for for metal-oxide metal-oxide gas gas sensor; sensor; (b) (b) Response Response of of metal-oxide metal-oxide gas gas sensor sensor to to ethanol ethanol with with ˝ different concentrations at 300 °C. different concentrations at 300 C.

4.4.Conclusions Conclusions Many ManyCMOS-compatible CMOS-compatible sensors sensors have have been been developed developed in in recent recent years. years. For Forsome somesimilar similarkinds kinds of sensors, it would be effective to develop a uniform platform for them. Based on same of sensors, it would be effective to develop a uniform platform for them. Based on the samethe platform, platform, different sensors could bewith developed withresearch a shortened low cost, and different sensors could be developed a shortened cycle, research low cost, cycle, and similar standard. similar Instandard. this paper, the multifunctional platform for sensors based on the microhotplate was In this paper, multifunctional platform sensors based on the microhotplate was developed. The the tungsten microhotplate couldforafford high temperature and it was fully developed. The tungsten microhotplate could afford high temperature and it was fully CMOS-compatible. Without extra fabrication processes, this platform has been configured as the CMOS-compatible. extra and fabrication processes, this gas platform been configured the Pirani gas pressure,Without temperature, thermal-conductivity sensor,has respectively. Besidesasthose Pirani gas pressure, temperature, and thermal-conductivity gas sensor, respectively. Besides those applications, with a few extra fabrication processes, including the deposition of the sensory film on applications, with a this few platform extra fabrication processes, as including the deposition thesome sensory film on the microhotplate, can be employed the heating componentoffor metal-oxide the microhotplate, this platform can be employed as the heating component for some metal-oxide gas sensors. gas sensors. Acknowledgments: This work was supported by the National Natural Science Foundation of China (Nos. 61201035 and 61274076). Acknowledgments: This work was supported by the National Natural Science Foundation of China (Nos. 61201035 and 61274076). Author Contributions: Jiaqi Wang developed the sensor platform and tested the Pirani gas pressure, and temperature sensor. Jun Yu designed the gas sensor test system and tested the gas sensor. Author Contributions: Jiaqi Wang developed the sensor platform and tested the Pirani gas pressure, and Conflicts of sensor. Interest:Jun The declare no sensor conflicttest of interest. temperature Yuauthors designed the gas system and tested the gas sensor.

Conflicts of Interest: The authors declare no conflict of interest.

References 1.

1603

Dai, C.L.; Lu, P.W.; Chang, C.L.; Liu, C.Y. Capacitive micro pressure sensor integrated with a ring oscillator circuit on chip. Sensors 2009, 9, 10158–10170.


References 1. 2. 3. 4. 5. 6.

7.

8. 9.

10.

11. 12. 13. 14. 15. 16. 17. 18.

19. 20. 21. 22. 23.

Dai, C.L.; Lu, P.W.; Chang, C.L.; Liu, C.Y. Capacitive micro pressure sensor integrated with a ring oscillator circuit on chip. Sensors 2009, 9, 10158–10170. [CrossRef] [PubMed] Chen, S.J.; Shen, C.H. A novel two-axis CMOS accelerometer based on thermal convection. IEEE Trans. Instrum. Meas. 2008, 57, 1572–1577. [CrossRef] Souri, K.; Chae, Y.; Makinwa, K.A.A. A CMOS temperature sensor with a voltage-calibrated inaccuracy of ˘0.15 ˝ C (3σ) from ´55 ˝ C to 125 ˝ C. IEEE J. Solid State Circuits 2013, 48, 292–301. [CrossRef] Hagleitner, C.; Hierlemann, A.; Lange, D.; Kummer, A.; Kerness, N.; Brand, O.; Baltes, H. Smart single-chip gas sensor microsystem. Nature 2001, 414, 293–296. [CrossRef] [PubMed] Baltes, H.; Brand, O. CMOS-based microsensors and packaging. Sens. Actuators A Phys. 2001, 92, 1–9. [CrossRef] Afridi, M.Y.; Suehle, J.S.; Zaghloul, M.E.; Berning, D.W.; Hefner, A.R.; Cavicchi, R.E.; Semancik, S.; Montgomery, C.B.; Taylor, C.J. A monolithic CMOS microhotplate-based gas sensor system. IEEE Sens. J. 2002, 2, 644–655. [CrossRef] Hautefeuille, M.; O’Flynn, B.; Peters, F.H.; O’Mahony, C. Development of a microelectromechanical system (MEMS)-based multisensor platform for environmental monitoring. Micromachines 2011, 2, 410–430. [CrossRef] Zhang, F.T.; Tang, Z.; Yu, J.; Jin, R.C. A micro-pirani vacuum gauge based on micro-hotplate technology. Sens. Actuators A Phys. 2006, 126, 300–305. [CrossRef] Ivanov, P.; Stankova, M.; Llobet, E.; Vilanova, X.; Brezmes, J.; Gracia, I.; Cane, C.; Calderer, J.; Correig, X. Nanoparticle metal-oxide films for micro-hotplate-based gas sensor systems. IEEE Sens. J. 2005, 5, 798–809. [CrossRef] Semancik, S.; Cavicchi, R.E.; Wheeler, M.C.; Tiffany, J.E.; Poirier, G.E.; Walton, R.M.; Suehle, J.S.; Panchapakesan, B.; DeVoe, D.L. Microhotplate platforms for chemical sensor research. Sens. Actuators B Chem. 2001, 77, 579–591. [CrossRef] Silvestri, S.; Schena, E. Micromachined flow sensors in biomedical applications. Micromachines 2012, 3, 225–243. [CrossRef] Kuo, J.T.W.; Yu, L.; Meng, E. Micromachined thermal flow sensors-a review. Micromachines 2012, 3, 550–573. [CrossRef] Garraud, A.; Combette, P.; Courteaud, J.; Giani, A. Effect of the detector width and gas pressure on the frequency response of a micromachined thermal accelerometer. Micromachines 2011, 2, 167–178. [CrossRef] Chae, J.; Stark, B.H.; Najafi, K. A micromachined pirani gauge with dual heat sinks. IEEE Trans. Adv. Packag. 2005, 28, 619–625. Paul, O.; Baltes, H. Novel fully CMOS-compatible vacuum sensor. Sens. Actuators A Phys. 1995, 46, 143–146. [CrossRef] Shie, J.S.; Chou, B.C.S.; Chen, Y.M. High-performance pirani vacuum gauge. J. Vac. Sci. Technol. A 1995, 13, 2972–2979. [CrossRef] Ali, S.Z.; Udrea, F.; Milne, W.I.; Gardner, J.W. Tungsten-based SOI microhotplates for smart gas sensors. J. Microelectromech. Syst. 2008, 17, 1408–1417. [CrossRef] Graf, M.; Barrettino, D.; Zimmermann, M.; Hierlemann, A.; Baltes, H.; Hahn, S.; Barsan, N.; Weimar, U. CMOS monolithic metal-oxide sensor system comprising a microhotplate and associated circuitry. IEEE Sens. J. 2004, 4, 9–16. [CrossRef] Yan, G.Z.; Chan, P.C.H.; Hsing, I.M.; Sharma, R.K.; Sin, J.K.O.; Wang, Y.Y. An improved TMAH Si-etching solution without attacking exposed aluminum. Sens. Actuators A Phys. 2001, 89, 135–141. [CrossRef] Tabata, O. PH-controlled TMAH etchants for silicon micromachining. Sens. Actuators A Phys. 1996, 53, 335–339. [CrossRef] Wang, J.Q.; Tang, Z.N.; Li, J.F.; Zhang, F.T. A micropirani pressure sensor based on the tungsten microhotplate in a standard CMOS process. IEEE Trans. Ind. Electron. 2009, 56, 1086–1091. [CrossRef] Wang, J.Q.; Tang, Z.N.; Li, J.F. Tungsten-microhotplate-array-based pirani vacuum sensor system with on-chip digital front-end processor. J. Microelectromech. Syst. 2011, 20, 834–841. [CrossRef] Wang, J.Q.; Tang, Z.A. A CMOS-compatible temperature sensor based on the gaseous thermal conduction dependent on temperature. Sens. Actuators A Phys. 2012, 176, 72–77. [CrossRef]

1604


24. 25.

26.

Sorge, S.; Pechstein, T. Fully integrated thermal conductivity sensor for gas chromatography without dead volume. Sens. Actuators A Phys. 1997, 63, 191–195. [CrossRef] Simon, I.; Arndt, M. Thermal and gas-sensing properties of a micromachined thermal conductivity sensor for the detection of hydrogen in automotive applications. Sens. Actuators A Phys. 2002, 97–98, 104–108. [CrossRef] Andio, M.A.; Browning, P.N.; Morris, P.A.; Akbar, S.A. Comparison of gas sensor performance of SnO2 nano-structures on microhotplate platforms. Sens. Actuators B Chem. 2012, 165, 13–18. [CrossRef] © 2015 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons by Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

1605