Performance analysis, modeling, and development of a signal conditioning unit of n-type metal oxide gas sensor for acetone gas detection

Metal oxide semiconductors have been widely used in the field of gas sensor study. Various researches are being done to improve the sensitivity of the sensing material for applications like breath analyzers. In this work, a theoretical investigation and analysis of the n-type metal oxide for acetone gas detection are carried out. The rate of change of resistance of the sensing material with respect to the change in the concentration of the target gas is analyzed. Acetone being a reducing gas, the resistance was found to decrease for n-type material. The simulations were done using COMSOL Multiphysics and results showed that the resistance of the sensing layer varies with the concentration of the target gas. Also, the performance analysis of sensors has been compared with the experimental results. Further, we have also derived a mathematical expression for connecting the relationship between the concentration of gas and the rate of change of resistance. The resistance change is observed to be proportional to the target gas concentration. A signal-conditioning circuit was also designed for providing a user-friendly interface for monitoring the gas concentration. The simulation of the signal conditioning circuit was done using Proteus Design Suite. This work will aid researchers to define and predict the behavior of gas sensors.


Introduction
The increase in pollution in the environment and its adverse effect on living organisms have drawn considerable attention toward the need for gas sensors [1]. There are different methods like resistive, acoustic, ultrasonic, thermoelectric, and electrochemical ones used for gas sensing. Among the different methods for gas sensing, resistive type-based gas sensors have gained a lot of research interest due to low cost, fabrication simplicity, and easy portability [2,3]. Metal oxide semiconductors are sensing materials used for gas sensors with resistive types. Metal oxides semiconductors are classified into two based on the type of charge carriers, namely n-type and p-type. ZnO, SnO 2 , WO 3 , etc are commonly used n-type semiconducting materials, and NiO, CuO, and Co 2 O 3 are the p-type semiconducting materials [4]. They have been used for the detection of different types of gases like Hydrogen ( H 2 ), Hydrogen sulfide ( H 2 S), Ammonia ( NH 3 ), Carbon monoxide (CO), Volatile Organic Compounds (VOCs), etc [5,6].
Metal oxide semiconductors with different geometries like nanotube, nanoflower, nanosphere, and hollow spheres have been synthesized [7,8]. The sensing layer with such structures helps to determine the selectivity and sensitivity of the sensor. Research work is mainly focussing on improving the selectivity and sensitivity of the sensor. Metal ions, metal oxide composites, and carbon materials are doped with the sensing layer, for enhancing the performance of the sensor. Due to the unique effect of temperature on the performance of the sensors, temperature modulation is a possible way for changing selectivity. In static detection, temperature modulation can significantly reduce the temperature drift effect and eliminate interference [9,10].
The transmittal mechanism of these sensors is based on the variation of the resistance of the sensing layer. The redox reactions which are taking place on the surface of the sensors are the reason for the change in resistance [11]. A hightemperature environment is required for the redox reactions to take place. This temperature is provided with the help of a microheater. Different geometries like meander, double meander, spiral, circular, etc have been designed for the heater for optimizing the temperature distribution and also for reducing the power consumption [12]. The change in the resistance of the sensing layer is measured using an interdigitated electrode. The gas sensor performance is evaluated by considering the parameters such as recovery time, selectivity, sensitivity, response time, the limit of detection, and operating temperature, respectively.
VOCs are emitted during day-to-day activities like cooking, building materials, driving a car, painting the house, or using pesticides [13]. These compounds have an adverse effect on human health even when exposed to low concentration. The human breath also contains VOCs which are an important breath marker for many diseases like kidney failure, diabetes, Asthma, etc [14]. Acetone is one of the important VOC present in the human breath which is a breath marker for diabetes mellitus. A healthy person has 0.2 to 1.8 ppm concentration of Acetone in his breath. Acetone concentration above 1.8 ppm indicates diabetes condition. If the concentration is above 25 ppm, then it's the acute case of diabetes mellitus [15][16][17]. Also in laboratories and industries, Acetone is used as a common liquid. When exposed to a high concentration of Acetone leads to nausea, headache, fatigue, and nervous system damage. So as a safety concern, the environmental concentration of Acetone monitoring is necessary [18].
Performance parameters of the gas sensor have been observed by using COMSOL Multiphysics. For n-type the properties of ZnO are used. Theoretical analysis and investigation of the sensing characteristics of ZnO gas sensors toward Acetone gas are performed. The effect of concentration on the rate of variation of resistance of the sensing layer was also analyzed. A signal conditioning circuit was also designed for providing a user-friendly interface for monitoring the gas concentration.

Gas sensing mechanism
Metal oxide gas sensors detect the target gas due to the changes in their electrical resistance [19]. This change happens due to the various adsorption and desorption process taking place on the sensing layer surface. These results in the change in the concentration of the charge carrier. When the metal oxide comes in contact with the atmosphere, the molecules of Oxygen that are absorbed over the surface educed the electrons from the conduction band. As a result, ions are formed due to the trapped electrons. Due to the falling off of the density of the electrons over the surface, an electron depletion layer is formed and a barrier potential is formed. The chemisorbed ions of oxygen formed depend on the operating temperature of the gas sensors. The following equations summarize the oxygen adsorption mechanism [20].
When the sensor is then disclosed to the target gas, the surface is getting absorbed by molecules of gas and then reacts with adsorbed ions of oxygen. The increase or decrease in resistance depends on the majority carriers in the metal oxide semiconductor [21]. The reaction of Acetone gas with the adsorbed oxygen ions is given by Eq. (5) [22]: When ZnO is exposed to air, oxygen adsorption takes place and the resistance of the metal oxide layer increases due to the removal of electrons [23]. A schematic demonstration of the reaction process is shown in Fig. 1a. As a result, an increase in depletion layer width has occurred as illustrated in Fig. 1b. When Acetone gas is fed through the chamber, the gas molecules react with the adsorbed oxygen ions. Acetone being a reducing gas causes the resistance of the sensing layer to decrease as the adsorbed electrons are injected back into the electron depletion layer [24], which is shown schematically in Fig. 1c. and the corresponding decrease in the potential barrier due to the injection of the adsorbed electrons back into the conduction band is shown in Fig. 1d.

Architecture of gas sensor
The schematic representation of the proposed gas sensor is shown in Fig. 2. The proposed gas sensor consists of the following parts: -The substrate is Silicon with 0.8 x 1 x 0.5 mm dimensions. It should well adhere to metal and have high mechanical stability.
-Meander-shaped Platinum heater with a dimension of 0.8 x 0.5 x 0.001 mm for providing the operating temperature. -An insulating layer made up of silicon dioxide with dimensions 0.8 x 0.5 x 1 mm for preventing direct contact between the heater and the electrode. -The resistance measurement is done using the electrode with a dimension of 0.8 x 0.5 x 0.5 mm. Platinum is the material used for the electrode. -The sensing layer of the sensor consists of a 1 mm thick ZnO for n-type sensor.

Finite element analysis
COMSOL Multiphysics was used for simulating the gas sensor. The gas chamber model, heat transfer model, and conducting model were the models simulated for analyzing the performance of the sensor.

Gas Chamber model
The gas chamber model is shown in Fig. 3a. A cylindrical chamber with a radius and height of 16 mm with one inlet and two outlets of a diameter of about 2 mm and 1 mm, respectively, are designed. The chamber is used to study the conductivity variation of the sensing layer when exposed to the Acetone gas. The gas is applied to the chamber from the upper base and is collected out from the other end with the help of two outlets for the easy outward flow of the gas from the chamber. Figure 3b depicts the gas flow distribution in the chamber.

Heat transfer model
The operating temperature for the working of the gas sensor is provided by using a microheater. The schematic representation of microheater inside the chamber is shown in Fig. 4a.
The simulated microheater showed in Fig. 4b. Platinum is the heating material used for the microheater due to its advantage like low power consumption [25]. To obtain the required temperature, DC voltage in the range of 0.1-5 V is applied. Meander geometry is chosen for the microheater for obtaining a uniform temperature distribution of heat on the sensing layer. In the experimental work [26] the maximum response is obtained at a temperature of 220 °C and the same has been used for finding the variation of resistance.
The platinum properties which are used in the simulation are shown in Table 1 The distribution of heat over the surface of the heater at a voltage of 5 V is shown in Fig. 5a and the variation of temperature with respect to the applied input voltage of the heater shown in Fig. 5b

Conducting model
Metal oxide semiconductors undergo a change in their resistance when exposed to the target gas. This property is used for gas sensors. The variation in resistance depends on the type of metal oxide used for gas sensing and it is achieved by the adsorption and desorption processes. The change in resistance is also analyzed as a function of temperature and concentration of gas.

Numerical simulation conditions for the conducting model
Solid mechanics module, heat transfer module, electrical module and chemical transport species module are the physics used for the simulation of gas sensors. The gas flow in the chamber is modeled using Navier Stokes equation given by [27], where , g, U, , and P are gas flux, gravity, velocity field, viscosity, and gas pressure, respectively. The above equation is solved using continuity equation for achieving the gas field velocity which is given by [28], When the gas has an incompressible flow becomes zero. Therefore equation reduces to, The boundary conditions of the inlet and outlet pressure are given by [29],  In the heat transfer model, the heat from the heater of the sensor is transferred to the sensing layer through conduction. The heat transfer in solids is given by [30], where q is the flux density, k is the electrical conductivity and T is the temperature, respectively. Also, there will be heat loss to the surrounding air and the device edges by radiation and convection. This phenomenon is described using Neumann ′ s law of cooling [31], where q 0 is the heat transferred, h is the heat transfer coefficient, T sur and T air is the surface and air temperature, respectively. For electrical model, the reaction between the gas molecules and the sensing surface is modeled using poisson's equation [32], As the sensing layer is shielded from the surrounding environment, the condition at the boundary for the current density is given by [33], where m is the normal vector of the plane of the active layer and J is the current density.

Response of gas sensor
The Scanning Electron Microscopic characterization of synthesized ZnO nanoparticles was found to have spherical structure are shown in Fig. 6 [34]. and the same has been used for simulating the sensing layer. When the oxygen adsorption takes place the electrons are removed from the (9) P inlet = 0 sensing layer and as a result, a depletion layer is formed. As a result, the resistance of the ZnO layer increases. When Acetone gas is fed to the chamber, it reacts with the oxygen adsorbed and releases electrons. This will reduce the electron depletion layer and causes a decrease in the resistance of the ZnO layer. At a temperature of 220 °C the variation of the resistance of the ZnO at different concentrations of the sensing layer is shown in Fig. 7. Table 2 shows the properties of the Zinc oxide Due to the adsorption of Acetone over the surface of the oxide layer the resistance changes. The resistance of the sensing layer in the presence of Acetone is represented by R acetone and the resistance in the presence of air is represented by R air . Hence the response of the gas sensor for n-type is given by Eq. (14) [35], The Time versus Response graphs for ZnO is shown in Fig. 8.

Experimental method
The experimental setup for the sensing of acetone gas is shown in Fig. 9. The metal oxide nanomaterials synthesized were coated on the interdigitated electrode. The sensor was placed in a vacuum chamber. (14) Response ntpye = R air ∕R acetone

3
The desired concentration of acetone vapor is obtained by injecting the Acetone liquid of a particular volume into the chamber. The volume of Acetone in ppm is determined using Eq. (15) [36] Where is the density of target liquid (g/mL), V T is the volume of liquid L, R is the gas constant, T is the temperature (K), M is the molecular weight, P is the chamber pressure(atm) and V is the volume of the chamber in liter. The experimental data are validated with the simulation data for 1000 ppm concentration of acetone and is shown in Fig. 10. The response graph for the same concentration is shown in Fig. 11.

Analytical approach for the rate of change of resistance
The rate of change of resistance is high for a higher concentration of Acetone gas as the adsorption will be more. Moreover, the maximum rate of change will occur immediately after turning on or turning off of Acetone gas flow. An exponential growth behavior is observed in all the resistance-time graphs. Furthermore the conductivity is given by [37], where n is the carrier concentration of electrons, p is the carrier concentration of holes, n is the drift velocity of electrons, p is the drift velocity of holes and q is the charge. As the adsorption takes place, the conductance of the sensing layer can be expressed as, (16) = n n q + p p q   where is the adsorption coefficient and C is the concentration of the target gas. Equation (17) shows the subjection of the sensing layer to an Acetone gas concentration for a short time. As resistivity is the inverse of conductance it can be expressed as, If n = p the rate of change can be expressed as, If Ct is very much smaller than n, Ct<<n, the rate of change can be expressed as As the resistance related to the cross-sectional area A and length l the equation can be written as Since all the parameters are constants, it can be shown that rate of resistance change is proportional to concentration of the gas.

Adding donor impurity
Similarly, if we are adding impurities to the sensing layer the equations can be modified. If adding a Donor impurity with donor carrier concentration, N D Eq. (16) can be changes to As the adsorption takes place, the conductance of sensing layer can be expressed as, If N D is very much larger than Ct, N D >> Ct, the rate of change can be expressed as Similar to Eq. (21) (17) sensing = (n − Ct) n q + (p + Ct) p q Since all the parameters are constants, it can be shown that rate of resistance change is proportional to concentration of the gas.
According to Eqs. (22) and (27) the maximum value of dR dt , is directly proportional to the concentration of the target gas. The first-order derivative curve at a particular concentration of Acetone gas for ZnO is shown in Fig. 12.

Signal conditioning unit
The simulated gas sensor resistance varies between a few M ranges depending upon the concentration of the gas present. It needs a dynamic measurement technique for the accurate readout of these resistance changes. A signal conditioning unit is simulated for this purpose can be utilized for n-type gas sensors. It contains a bridge circuit followed by a low signal amplifier shown in Fig. 13. When contact with the target gas, the conductance of the sensing material changes, accordingly it shows a detectable variation in its resistance. These resistance variations are detected by a Wheatstone bridge circuitry, which includes three standard resistors, R 1 , R 2 , R 3 , and one sensor. The bridge output voltage changes according to the resistance of the sensing layer. The output of the bridge is connected to a low signal amplifier. This work uses an AD623 instrumentation amplifier as a low signal amplifier [38]. To ensures the single and dual supply operations it is a modified version of the classic Where V 0 is the output voltage of the amplifier, V x is the non-inverting terminal voltage, V y is the inverting terminal voltage R gain is the external gain controlling resistance. For getting a particular gain it needs to calculate the value of the resistance by equation, The resistance variation is due to the adsorption of Acetone gas on the ZnO is observed in the range of 1-10M , for analyzing the performance of the sensor over this range the gain is selected as 1.5 by choosing the R gain as 200k . The simulation is done using Proteus Design Suite software. The simulated signal conditioning unit with ADC output is shown in Fig. 14. In the simulation, the Arduino Uno board is used for the ADC conversion it has an inbuilt resolution of 4.9 mV per unit 5/1024.
The concentration of acetone gas is varied from 1000 to 62.5 ppm and the corresponding resistance, output voltage, and the ADC output values for the ZnO are shown in Table 3.
It is observed that at a higher Acetone gas concentration of 1000 ppm the resistance was ∼ 1.56M with the corresponding output voltage of 3.18 V and at the lower gas concentration, the resistance was ∼ 8.18M and the corresponding output voltage was 1.09 V. Therefore as the input gas concentration decreases the resistance value increases, the output voltage decreases, accordingly, the ADC value decreases.

Comparison of the proposed work with the reported results of references [24] and [34]
The proposed work developed a feasible method to rapidly determine the concentration of target gas for metal oxide semiconductors considering the donor impurities separately. The work reported in the literature [24] for the detection of reducing gas has not given any significance for the Despite finite element simulation tool COM-SOL being increasingly used for the simulation of the sensing mechanism in the gas sensors, there is no modeling for the rate of change of resistance with the gas concentration and simulation of signal conditioning unit for the measurement of the concentration of target gas. Analytical evaluation is very much useful for determining the general properties of a gas sensor. Chiu Hsien and et al, [34] reported a general approach on the relationship between the rate of the change of resistance dr/dt and gas concentration. In this work, it is separately analyzed for an n-type metal oxide gas sensor for reducing gas.

Conclusion
Performance analysis of metal oxide semiconductors with ZnO as a sensing material has been presented in this work. The optimum operating temperature of the sensing layer is provided by using a microheater with meander geometry. The sensing response for the n-type sensor was observed at different gas concentrations. The performance of the gas sensors has been observed by using the simulation software and has been verified with the experimental data. Developed a mathematical expression for the rate of change of resistance at different concentrations of the Acetone, and it is found to be proportional to the gas concentration. A signal conditioning circuit was also developed for accurately reading the measured resistance change. For n-type metal oxide, a higher ADC value indicates a higher concentration of Acetone. The presented model and signal conditioning unit in this work are capable of analyzing the performance of any n-type of gas sensors made with metal oxide.