Co-Evaporated CuO-Doped In2O3 1D-Nanostructure for Reversible CH4 Detection at Low Temperatures: Structural Phase Change and Properties

In order to improve the sensitivity and to reduce the working temperature of the CH4 gas sensor, a novel 1D nanostructure of CuO-doped In2O3 was synthesized by the co-evaporation of Cu and In granules. The samples were prepared with changing the weight ratio between Cu and In. Morphology, structure, and gas sensing properties of the prepared films were characterized. The planned operating temperatures for the fabricated sensors are 50–200 °C, where the ability to detect CH4 at low temperatures is rarely reported. For low Cu content, the fabricated sensors based on CuO-doped In2O3 showed very good sensing performance at low operating temperatures. The detection of CH4 at these low temperatures exhibits the potential of the present sensors compared to the reported in the literature. The fabricated sensors showed also good reversibility toward the CH4 gas. However, the sensor fabricated of CuO-mixed In2O3 with a ratio of 1:1 did not show any response toward CH4. In other words, the mixed-phase of p- and n-type of CuO and In2O3 materials with a ratio of 1:1 is not recommended for fabricating sensors for reducing gas, such as CH4. The gas sensing mechanism was described in terms of the incorporation of Cu in the In2O3 matrix and the formation of CuO and In2O3 phases.


Introduction
It is known that CH 4 is detected by a few materials at high operating temperatures. However, some literature reported developed materials that give a satisfactory performance to CH 4 at relatively high temperatures [1][2][3][4][5][6][7][8][9]. Among these materials, SnO 2 NR-NP-Gr hybrids and PdPt-SnO 2 -rGO which showed a promising property for methane at as low as 150 • C [2,10]. Also, our previous In 2 O 3 nanostructure film showed a promising performance toward methane at 100 • C [11]. As technology improved, an E-nose for CH 4 at RT was developed by commercially available SnO 2 [12]. However, the operating temperature of the promising In 2 O 3 sensor needs to be reduced to lower temperatures.
Recently, researchers have paid great attention to develop new nanocomposites of binary or ternary metal oxides with a new functionality [13][14][15][16][17]. Heterostructure materials offer more opportunities by the effective combination of two or more chemically dissimilar components in a single nanostructure with multifunctional properties. They are surely led to an innovative application of nanomaterials in various fields. Thus, the creation of heterostructure materials is an effective way to achieve low-temperature gas sensor [18][19][20][21]. N-type In 2 O 3 is a wide bandgap semiconductor with stable physical and chemical properties. Its thin film is important for many applications and it is transparent to the visible light. Currently, In 2 O 3 films are used for solar cells [22], optoelectronic devices [23], and

Preparation Method
The preparation of films was carried out in a vacuum stainless-steel chamber. The evaporation takes place in the evaporation closed system consisted of alumina crucible surrounded by the heating element and covered by a glass substrate. The deposition area is limited to an exposed area of 6.28 cm 2 , keeping the substrate is heated. The temperature of the set is controlled at 900 • C by the supplied current to the heater.

Co-Evaporation and In Situ Oxidation of Nanocomposite
To prepare CuO-mixed In 2 O 3 , Cu, and In pieces of total 3.0 mg with a purity of 99.999% are kept in the evaporation source, which covered by the glass substrate connected with a thermocouple. The chamber was kept under a vacuum of 7 × 10 −2 mbar, keeping the air is flowing into the chamber, which is connected to a rotary pump. Then the specific current was applied on the heating element to heat up the evaporation set up to 900 • C within 5.0 min, which was kept for 30 min, where the temperature of the substrate is maintained at 400 • C. Afterward, the evaporation set is naturally cooled down to room temperature, taking about 90 min to cool. After deposition, it is found that the substrate is coated with a film strongly adhered to the substrate. The thickness of the deposited film was about 1 µm. The color of the deposited film is depending on the ratio Cu:In, where it was light yellowish for In 2 O 3 and then came to be black with increasing Cu content. Three samples are prepared based on the weight ratio in mg of Cu:In granules as 0.5:2.5, 1.0:2.0, and 2.5:0.5, which are assigned here as S1, S2, and S3, respectively.

Characterization Tools
CuKα X-ray technique (Philips Type PW 1710) with 1.546 Å was used to investigate the crystal structure of synthesized nanocomposites in diffraction angle (2θ = 4-70 • ). the morphology was observed by using SEM (JEOL JSM-5400LV) (Tokyo, Japan) attached with EDX for elemental analysis of the formed composites. The nanostructured films were also characterized by Raman spectroscopy. Raman shift spectra were recorded from the film samples with FRA 106 (Nd: YAG-Laser. 442 nm) spectrometer.

Sensor Preparation and Characterization
For fabricating the gas sensor, sputtered two electrodes of gold with a gap of 400 µm were deposited on the surface of the sensing layer at room temperature (RT). Because the sensors are investigated at operating temperatures higher than RT (up to 200 • C), and as a providence process, Materials 2019, 12, 4073 3 of 14 they were treated at 350 • C for 30 min in the air in order to avoid any change in contact interface during the sensing measurements. Figure 1 shows the schematic illustration of the gas sensing equipment and the sample stage. The sensing measurements were performed in a quartz chamber, which consists of a controlled electric furnace and electrical terminals connected to a computerized data acquisition (Multi-channel LXI Agilent 34972 A). The low operating temperature applied here lies between 50 and 200 • C. Three sensors are applied to the same conditions where a dry synthetic air (21% O 2 + 79% N 2 ) mixed with 1.0% CH 4 is flowing with 200 SCCM into the testing chamber by using SEC-N112 MGM-Horiba. Figure 1b shows the schematic diagram of the device with the electrical connections. The resistance and output voltage of the sensors is calculated with applying of 5.0-V potential difference to the electric circuit.  Fig. 1b shows the schematic diagram of the device with the electrical connections. The resistance and output voltage of the sensors is calculated with applying of 5.0-V potential difference to the electric circuit.

Structural Characterizations
The morphology of the films obtained of the co-evaporated Cu and In is characterized by SEM images, as shown in Figure 2. The slow evaporation or sublimation of the mixed-phase allows the nanostructures to grow up. The grain-like structure is observed for S1 and S3, whereas, the 1D nanostructure is observed for S2. The surface thermal energy of the heated substrate allows the landing atoms to aggregate and build up nuclei for growing nanostructure. In earlier work, a pure phase of In2O3 prepared under the same condition was sphere-like grains [11]. Thus, the incorporation of Cu atoms in the present evaporation may affect the dimension growth. With low Cu content, the seeds of the one-dimensional structure started as shown in the S1 sample. With moderated Cu content, 1D nanostructures are growing, as shown in the S2 sample ( Figure 2b). The 1D nanostructure is inclined or vertical to the substrate surface and ended with needle-like at the tips. It seems that the excessive Cu content in the composite suppressed the growth of the 1Dnanostructure, as shown in the S3 sample ( Figure 2c).

Structural Characterizations
The morphology of the films obtained of the co-evaporated Cu and In is characterized by SEM images, as shown in Figure 2. The slow evaporation or sublimation of the mixed-phase allows the nanostructures to grow up. The grain-like structure is observed for S1 and S3, whereas, the 1D nanostructure is observed for S2. The surface thermal energy of the heated substrate allows the landing atoms to aggregate and build up nuclei for growing nanostructure. In earlier work, a pure phase of In 2 O 3 prepared under the same condition was sphere-like grains [11]. Thus, the incorporation of Cu atoms in the present evaporation may affect the dimension growth. With low Cu content, the seeds of the one-dimensional structure started as shown in the S1 sample. With moderated Cu content, 1D nanostructures are growing, as shown in the S2 sample ( Figure 2b). The 1D nanostructure is inclined or vertical to the substrate surface and ended with needle-like at the tips. It seems that the excessive Cu   Figure 3 illustrates the XRD charts of the prepared nanostructures for S1, S2, and S3 with adopted CuO (no. 04-015-5877) and In2O3 (no. 04-021-4783). The analysis of XRD showed that the films are polycrystalline with a cubic bixbyite-type structure In2O3 and monoclinic structure of CuO. It is observed that the preferred orientation planes are (222) for In2O3 and (−111) and (111) for CuO. From XRD charts several observations can be concluded. The first observation is the shift in the peaks (222), (400), (440), (622), and (444) to high diffraction angles, indicating the incorporation of Cu atoms into the In2O3 matrix. By comparing the ionic radii of In 3+ (0.81 Å) and Cu 2+ (0.73 Å), the substitution is expected mechanism for incorporating Cu 2+ ions into In2O3 lattice at low concentrations. The substitution causes a shrinkage in the lattice dimensions, which is observed from the shift of peak positions to large diffraction angles. The second observation is the appearance of new peaks of (110), (111), and (−202) which may be close to the CuO phase. They have a shift also to the large diffraction angles. Nonetheless, it is expected for CuO peak positions to be shifted to lower diffraction angles because d-spacing should expand due to In 3+ radii. A final observation is the dominance of the orientation plane of (222) for the sample S2 overall other orientations, confirming the formation of vertical-like structure observed for S2. From these observations, we can claim that a new phase nanocomposite of CuO-mixed In2O3 is formed. Celref UNIT CELL software was used to calculate the lattice parameters using the raw data obtained from XRD. From the XRD chart analysis, the lattice parameters and d-spacing of In2O3 shrinkage respective to the standard values recorded in card no. 04-021-4783 are listed in Table 1.  Figure 3 illustrates the XRD charts of the prepared nanostructures for S1, S2, and S3 with adopted CuO (no. 04-015-5877) and In 2 O 3 (no. 04-021-4783). The analysis of XRD showed that the films are polycrystalline with a cubic bixbyite-type structure In 2 O 3 and monoclinic structure of CuO. It is observed that the preferred orientation planes are (222) for In 2 O 3 and (−111) and (111) for CuO. From XRD charts several observations can be concluded. The first observation is the shift in the peaks (222), (400), (440), (622), and (444) to high diffraction angles, indicating the incorporation of Cu atoms into the In 2 O 3 matrix. By comparing the ionic radii of In 3+ (0.81 Å) and Cu 2+ (0.73 Å), the substitution is expected mechanism for incorporating Cu 2+ ions into In 2 O 3 lattice at low concentrations. The substitution causes a shrinkage in the lattice dimensions, which is observed from the shift of peak positions to large diffraction angles. The second observation is the appearance of new peaks of (110), (111), and (−202) which may be close to the CuO phase. They have a shift also to the large diffraction angles. Nonetheless, it is expected for CuO peak positions to be shifted to lower diffraction angles because d-spacing should expand due to In 3+ radii. A final observation is the dominance of the orientation plane of (222) for the sample S2 overall other orientations, confirming the formation of vertical-like structure observed for S2. From these observations, we can claim that a new phase nanocomposite of CuO-mixed In 2 O 3 is formed. Celref UNIT CELL software was used to calculate the lattice parameters using the raw data obtained from XRD. From the XRD chart analysis, the lattice parameters and d-spacing of In 2 O 3 shrinkage respective to the standard values recorded in card no. 04-021-4783 are listed in Table 1.   Figure 3 illustrates the XRD charts of the prepared nanostructures for S1, S2, and S3 with adopted CuO (no. 04-015-5877) and In2O3 (no. 04-021-4783). The analysis of XRD showed that the films are polycrystalline with a cubic bixbyite-type structure In2O3 and monoclinic structure of CuO. It is observed that the preferred orientation planes are (222)    For supporting the data observed in XRD, EDX spectra of the obtained nanostructures are shown in Figure 4. EDX spectra exhibited the existence of CuK line of low and high Cu content. The emission of CuK spectra is very strong at high Cu content, as shown for S3 in Figure 4. The InL spectra are dominant at low Cu content. The elemental analysis of the nanostructures of the actual weight and atomic percentage for Cu and In is listed in the table attached in Figure 4. EDX analysis shows that Cu's weight percent in the samples increases from 0.81%, to 2.49%, up to 52.1% with increasing Cu weight for S1, S2, S3, respectively. In other words, the atomic percentage of oxygen is dependent on the nanostructure or composite obtained, where it is recorded as 72.74, 54.86, and 59.68% for S1, S2, S3, respectively. This change in oxygen percentage gives an indicator of the composite formation as follows. (1) At low Cu content, the oxygen atoms are mostly linked to the In ions in In 2 O 3 structure, where Cu atoms involved in the In 2 O 3 lattice are less. (2) with increasing Cu ions in In 2 O 3 lattice with substitution mechanism, the bridging oxygen atoms are missing, and the atomic oxygen ratio decreases. Moreover, the building up for 1D nanostructure may increase the oxygen vacancies on its surface. (3) with exceeding Cu content, more substitution of Cu is expected to grow up, and oxygen ions are linked with both Cu 2+ and In 3+ , keeping the oxygen percentage less due to increasing of Cu 2+ content (CuO).   The measured Raman spectra of S1, S2, and S3 are shown in Figure 5. The shown spectrum for each sample is the summation of 64 spectra taken from different points on the surface in order to get a full image about the nature and the uniformity of the surface. The comparison of the measured frequencies is also listed in Table 2. For the S1 sample, where the Cu content is very less about 0.8%, the vibrations of Cu or CuO are not detected, and the only vibrations detected for cubic-In 2

CH 4 Detection at Low Temperatures
The advantage of the mixed-phase of CuO and In 2 O 3 is the low operating temperature sensor for methane gas. Many attempts for researchers were focused to fabricate a low-temperature methane sensor for safety control and reducing power consumption, but most of the proposed sensors are working at relatively high temperatures [1][2][3][4][5][6][7][8][9]. In the present study, we have successfully developed a sensor working at relatively low temperatures. Figures 6-8 show the sensor signals toward the methane gas at low operating temperatures (50-200 • C). We should point out that the fabricated sensor did not show a response to CH 4 at room temperature (signal not shown). Figure 6 shows the sensor signals at 50-200 • C for the S1 layer. At 50 • C, the sensor responds quickly once the methane is introduced into the measurement chamber, and it shows an n-type behavior of the semiconductor. Although the sensor response at this temperature is slightly low, the sensor recorded a response time constant about 75.0 s and recovery time constant about 2.0 s. The response/recovery time constant is the time taken to reduce or increase the sensor resistance to 90% of its initial value. The response of the sensor increased at an operating temperature of 100 • C, and then the response and recovery time constants also increased. Interesting is the behavior of the sensor at 150 • C, where the sensor does not respond to the gas at all. Then the sensor showed a behavior of p-type semiconductor at 200 • C. The resistance of the sensor increased upon exposure to the methane gas, it is due to that the sensing layer consists of n-and p-type materials.   Figure 8 shows the output voltage of sensors fabricated with S1, S2, and S3 films at 100 °C. These signals clearly distinguish between the sensing ability of the fabricated sensors, confirming that S3 is not sensitive to CH4.
The sensor response is calculated as where Ra is the resistance in air and Rg the resistance in the presence of gas. The negative sign in the definition of the sensor response differentiates between the sensor response of n-and p-type. Figure   Figure 6. Change in resistance at operating temperatures of 50-200 • C for the sensor fabricated by the S1 sensing layer.   Figure 8 shows the output voltage of sensors fabricated with S1, S2, and S3 films at 100 °C. These signals clearly distinguish between the sensing ability of the fabricated sensors, confirming that S3 is not sensitive to CH4.
The sensor response is calculated as where Ra is the resistance in air and Rg the resistance in the presence of gas. The negative sign in the definition of the sensor response differentiates between the sensor response of n-and p-type. Figure showing a better sensitivity than that fabricated by S1. Although these two sensors respond to CH4 at 50 °C, the optimal temperature is 100 °C. The sensor fabricated by the S3 sensing layer did not show any response toward methane with increasing the operating temperature from 50 to 200 °C. It seems that the formation of p-and n-type semiconductor composite is not favored to be with 1:1 ratio. This reason can be explained through the point of view of the sensing mechanism.  Figure 10 illustrates a schematic diagram of the gas sensing mechanism, showing two effects due to incorporating Cu into In2O3 grains and the formation of the p-CuO/n-SnO2 phase. As it is known, the sensing property is a surface property depending on how the gas reacts with the material. In the air, the adsorbed oxygen species create a depletion layer on the surface of oxide due to the capture of the conduction electrons. In other words, the adsorbed oxygen molecules on the surface of oxide pass through chemisorbed reaction reported by Ruhland et al [38].

Gas Sensing Mechanism
Thus, in the case of n-type materials the resistivity increases, while the resistivity decreases in the case of p-type. The incorporation of Cu 2+ into the In2O3 matrix causes the lattice mismatching because of the difference in the ionic radii. This may cause an increase in the donor density in In2O3, allowing the adsorbed oxygen molecules to capture more electrons, which creates a depletion layer inside the grain due to the dangling bond, as proposed in Figure 10a. While the high Cu concentration formed a grain to grain interaction. In other words, p-type grain and n-type grain are formed, and it is not layered or film junction. The CuO/In2O3 heterojunction is insignificant compared to the effect of the bulk size. The expected boundary reaction between In2O3 and CuO is shown in Figure 10b. Based on this structure and the above results obtained from the gas sensing measurements, we could report the following. The current target gas is a reducing gas, which reacts with adsorbed oxygen species and produces H2O and CO2 [39][40][41].  Figure 7 shows the change in resistance for the S2 sensor at various operating temperatures, showing n-type behavior. The S2 sensor is good sensitive to methane even at a low temperature of 50 • C, the sensor response may be improved, but the response and recovery time constants slightly increased up to 170 and 25 s, respectively. On the contrary, with the S1 sensor, S2 sensor response time became shorter down to 61.0 s and the sensor response improved at 100 • C, as well. By increasing the temperature up to 150 • C, the sensor takes a long time to recover to its initial resistance. Moreover, it is insensitive to methane at 200 • C. Compared to the response time, the recovery time is small for the sensor at low temperatures. It is expected that, at such low operating temperatures, the surface is just active for gas-thus, upon the exposure to 1% gas-the reaction of gas with surface species to saturate the surface of the sensing layer is slow. However, upon the exposure to the air (switching off the gas), the desorption is fast. We may ascribe this to the weak electric force between the gas and surface species at low temperatures. This can be confirmed by the small change in the sensor resistance at 50 • C within a few ohms. Thus, the response time is longer than the recovery time. Figure 8 shows the output voltage of sensors fabricated with S1, S2, and S3 films at 100 • C. These signals clearly distinguish between the sensing ability of the fabricated sensors, confirming that S3 is not sensitive to CH 4 .
The sensor response is calculated as × 100, f or n-type behavior where Ra is the resistance in air and Rg the resistance in the presence of gas. The negative sign in the definition of the sensor response differentiates between the sensor response of n-and p-type. Figure 9 shows the response as a function of the operating temperature. The sensor fabricated by S2 is showing a better sensitivity than that fabricated by S1. Although these two sensors respond to CH 4 at 50 • C, the optimal temperature is 100 • C. The sensor fabricated by the S3 sensing layer did not show any response toward methane with increasing the operating temperature from 50 to 200 • C. It seems that the formation of p-and n-type semiconductor composite is not favored to be with 1:1 ratio. This reason can be explained through the point of view of the sensing mechanism. Figure 9. Sensor response as a function of operating temperatures for the sensors fabricated with S1, S2, and S3 sensing layers. Note: the negative values here are defined as negative sensor response for p-type behavior to be notable compared to n-type behavior.

→ 2 complete reaction
When CH4 gas reacts with the oxygen species, the electrons are injected back to the conduction band of the oxide. Thus, the electrons density increases in In2O3 grain, reducing its depletion layer. Whereas, other injected electrons recombine with the positive holes inside CuO, extending its depletion layer [1]. Consequently, the depletion layer at the grain-grain boundary is not affected by this reaction (where its depletion thickness increases in p-CuO and decreases in n-In2O3), and its effect on the gas sensing is negligible. Thus, the sensing property here depends only on CuO and In2O3 grains. If the ratio of n-type to p-type is 1:1 as in the present case, both effects may cancel each other. Thus, the net change in the resistance will not be noticeable, where the net resistance approximately is the sum of Rp, Rpn, and Rn, as shown in Figure 10b.
The reversibility of the sensor signal upon exposure to CH4 gas is shown in Figure 11. The signal is recorded for the most sensitive sensor fabricated by the S2 sensing layer. The sensor is very sensitive when the gas is switching on/off. The signal is repeated several times, and there is no observed drift. Moreover, the signal is carried out at a low temperature of 100 °C, with quick response and recovery. The time interval for each signal is about 550 s including the steady time for the resistance. This means that the response and recovery times a constant is as short as ~120 s.  . Sensor response as a function of operating temperatures for the sensors fabricated with S1, S2, and S3 sensing layers. Note: the negative values here are defined as negative sensor response for p-type behavior to be notable compared to n-type behavior. Figure 10 illustrates a schematic diagram of the gas sensing mechanism, showing two effects due to incorporating Cu into In 2 O 3 grains and the formation of the p-CuO/n-SnO 2 phase. As it is known, the sensing property is a surface property depending on how the gas reacts with the material. In the air, the adsorbed oxygen species create a depletion layer on the surface of oxide due to the capture of the conduction electrons. In other words, the adsorbed oxygen molecules on the surface of oxide pass through chemisorbed reaction reported by Ruhland et al. [38]. Figure 9. Sensor response as a function of operating temperatures for the sensors fabricated with S1, S2, and S3 sensing layers. Note: the negative values here are defined as negative sensor response for p-type behavior to be notable compared to n-type behavior.

Gas Sensing Mechanism
When CH4 gas reacts with the oxygen species, the electrons are injected back to the conduction band of the oxide. Thus, the electrons density increases in In2O3 grain, reducing its depletion layer. Whereas, other injected electrons recombine with the positive holes inside CuO, extending its depletion layer [1]. Consequently, the depletion layer at the grain-grain boundary is not affected by this reaction (where its depletion thickness increases in p-CuO and decreases in n-In2O3), and its effect on the gas sensing is negligible. Thus, the sensing property here depends only on CuO and In2O3 grains. If the ratio of n-type to p-type is 1:1 as in the present case, both effects may cancel each other. Thus, the net change in the resistance will not be noticeable, where the net resistance approximately is the sum of Rp, Rpn, and Rn, as shown in Figure 10b.
The reversibility of the sensor signal upon exposure to CH4 gas is shown in Figure 11. The signal is recorded for the most sensitive sensor fabricated by the S2 sensing layer. The sensor is very sensitive when the gas is switching on/off. The signal is repeated several times, and there is no observed drift. Moreover, the signal is carried out at a low temperature of 100 °C, with quick response and recovery. The time interval for each signal is about 550 s including the steady time for the resistance. This means that the response and recovery times a constant is as short as ~120 s.  Thus, in the case of n-type materials the resistivity increases, while the resistivity decreases in the case of p-type. The incorporation of Cu 2+ into the In 2 O 3 matrix causes the lattice mismatching because of the difference in the ionic radii. This may cause an increase in the donor density in In 2 O 3 , allowing the adsorbed oxygen molecules to capture more electrons, which creates a depletion layer inside the grain due to the dangling bond, as proposed in Figure 10a. While the high Cu concentration formed a grain to grain interaction. In other words, p-type grain and n-type grain are formed, and it is not layered or film junction. The CuO/In 2 O 3 heterojunction is insignificant compared to the effect of the bulk size. The expected boundary reaction between In 2 O 3 and CuO is shown in Figure 10b. Based on this structure and the above results obtained from the gas sensing measurements, we could report the following. The current target gas is a reducing gas, which reacts with adsorbed oxygen species and produces H 2 O and CO 2 [39][40][41].
When CH 4 gas reacts with the oxygen species, the electrons are injected back to the conduction band of the oxide. Thus, the electrons density increases in In 2 O 3 grain, reducing its depletion layer. Whereas, other injected electrons recombine with the positive holes inside CuO, extending its depletion layer [1]. Consequently, the depletion layer at the grain-grain boundary is not affected by this reaction (where its depletion thickness increases in p-CuO and decreases in n-In 2 O 3 ), and its effect on the gas sensing is negligible. Thus, the sensing property here depends only on CuO and In 2 O 3 grains. If the ratio of n-type to p-type is 1:1 as in the present case, both effects may cancel each other. Thus, the net change in the resistance will not be noticeable, where the net resistance approximately is the sum of Rp, Rpn, and Rn, as shown in Figure 10b.
The reversibility of the sensor signal upon exposure to CH 4 gas is shown in Figure 11. The signal is recorded for the most sensitive sensor fabricated by the S2 sensing layer. The sensor is very sensitive when the gas is switching on/off. The signal is repeated several times, and there is no observed drift. Moreover, the signal is carried out at a low temperature of 100 • C, with quick response and recovery. The time interval for each signal is about 550 s including the steady time for the resistance. This means that the response and recovery times a constant is as short as~120 s.

Conclusion
In summary, the CuO-doped In2O3 nanocomposite was successfully synthesized through the developed vacuum evaporation method. The morphology, structure, and gas sensing features of the prepared films were well characterized. The samples were prepared with different ratios of the weight of Cu and In granules. At low Cu content, the seeds of 1D nanostructure were grown vertically to the substrate surface with needle-like at the tips. However, the excessive Cu content in the composite suppressed the growth of 1D nanostructure. The sensing performance of CuO-mixed In2O3 film toward methane gas (CH4) was examined at low operating temperatures of 50-200 °C. For low Cu content, the fabricated sensors based on CuO-doped In2O3 showed very good sensing properties at 50 °C. The sensor fabricated by 2.3 mol% CuO-doped In2O3 exhibited the highest responses at 100 °C, as well as showed good reversibility and stability toward CH4 gas. The detection of CH4 at low temperatures shows the potential of the present sensors compared to the reported in the literature. However, the preparation of p-and n-type mixed materials by the ratio of 1:1 is not preferred for CH4 detection. The gas sensing mechanism was described in terms of the incorporation of Cu in the In2O3 matrix and the formation of CuO and In2O3 phases. We can conclude that the present fabricated

Conclusion
In summary, the CuO-doped In 2 O 3 nanocomposite was successfully synthesized through the developed vacuum evaporation method. The morphology, structure, and gas sensing features of the prepared films were well characterized. The samples were prepared with different ratios of the weight of Cu and In granules. At low Cu content, the seeds of 1D nanostructure were grown vertically to the substrate surface with needle-like at the tips. However, the excessive Cu content in the composite suppressed the growth of 1D nanostructure. The sensing performance of CuO-mixed In 2 O 3 film toward methane gas (CH 4 ) was examined at low operating temperatures of 50-200 • C. For low Cu content, the fabricated sensors based on CuO-doped In 2 O 3 showed very good sensing properties at 50 • C. The sensor fabricated by 2.3 mol% CuO-doped In 2 O 3 exhibited the highest responses at 100 • C, as well as showed good reversibility and stability toward CH 4 gas. The detection of CH 4 at low temperatures shows the potential of the present sensors compared to the reported in the literature. However, the preparation of p-and n-type mixed materials by the ratio of 1:1 is not preferred for CH 4 detection. The gas sensing mechanism was described in terms of the incorporation of Cu in the In 2 O 3 matrix and the formation of CuO and In 2 O 3 phases. We can conclude that the present fabricated sensor of low content of Cu has a convenient application for safety control at low temperatures.