Gas-Sensing Performance of M-Doped CuO-Based Thin Films Working at Different Temperatures upon Exposure to Propane

Cupric oxide (CuO) thin films are promising materials in gas sensor applications. The CuO-based gas sensors behaved as p-type semiconductors and can be used as part of an e-nose or smart sensor array for breath analysis. The authors present the investigation results on M-doped CuO-based (M = Ag, Au, Cr, Pd, Pt, Sb, Si) sensors working at various temperatures upon exposure to a low concentration of C3H8, which can be found in exhaled human breath, and it can be considered as a one of the biomarkers of several diseases. The films have been deposited in magnetron sputtering technology on low temperature cofired ceramics substrates. The results of the gas sensors’ response are also presented and discussed. The Cr:CuO-based structure, annealed at 400 °C for 4 h in air, showed the highest sensor response, of the order of 2.7 at an operation temperature of 250 °C. The response and recovery time(s) were 10 s and 24 s, respectively. The results show that the addition of M-dopants in the cupric oxide films effectively act as catalysts in propane sensors and improve the gas sensing properties. The films’ phase composition, microstructure and surface topography have been assessed by the X-ray diffraction (XRD), scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDX) methods.

and 100 ppm, and ΔR/R0 increased regularly by approximately 0.02 at each 25 ppm concentration step. The sensitivity under exposure to 1 ppm and 100 ppm of propane is approximately 0.08 and 0.16, respectively. Saberi et al. [19] have reported results on the dual selective Pt/SnO2 sensor to propane. However, the authors have focused on an automotive application due to their measured sensors being under higher propane concentrations (1000-10,000 ppm). Even though, the maximum sensor response Ra/Rg (where Ra is the resistance of the sensor in the presence of air and Rg is the resistance in the presence of propane gas) was approximately five. Sun et al. [20] have reported results on a Zn-M-O (M = Sn, Co) sensing electrode for selective propane sensors. The highest sensitivity (defined as mV/decade) was obtained for a Zn-Sn-O composited sensing electrode with 50% Pt coverage, and it was approximately 17. Liu et al. [21] have presented investigation results on a Pt-CeO2 nanofiber-based high-frequency impedancemetric gas sensor. The sensor response defined as log(Zg/Z0), where Z0 is the impedance of the sensor in the presence of air and Zg is the impedance in the presence of propane gas, was approximately one under 20 ppm C3H6. However, it is still too high of a range in comparison to exhaled propane levels.
In this paper, novel M-doped CuO-based (M = Ag, Au, Cr, Pd, Pt, Sb, Si) sensors with enhanced sensitivity to propane have been presented. The long-term stability is described and discussed. The films' phase composition, microstructure and surface topography have been assessed by X-ray diffraction (XRD), scanning electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDX) methods.

Preparation of Films
The reported films were deposited on silicon and LTCC (low temperature cofired ceramic) substrates, previously reported in [22] by an MF (medium frequency) magnetron co-sputtering system, schematically shown in Figure 1. The metallic copper target (purity 99.995%) and metallic dopants targets (99.95%) of 50 mm and 10 mm in diameter were used, respectively. The pressure was set to 2.0 × 10 −6 mbar and 4.0 × 10 −2 mbar for the base vacuum and the working pressure, respectively. The target to substrate distance was set to 50 mm. After the standard cleaning process, the gas-sensitive layers were deposited at 100 °C and then annealed at 500 °C for 4 h in air. The deposition and annealing step parameters were chosen based on previous results. The substrate temperature was controlled by PID Eurotherm 2408. The targets were pre-sputtered for 10 min to eliminate target surface contamination and to obtain a stable plasma density. Sputtering was then performed under both pure argon (Ar) and argon/oxygen (90% Ar/10% O2). The sputtering was completed with a low power of 40 W. The sputtering time was changed to yield different film thicknesses. However, the highest gas responses were obtained for a 50-nm film thickness. The responses of the other sensors were around 20%-50% lower than observed in the case of the sensor with an optimal thickness of 50 nm. The sputtering parameters were chosen based on previously-reported results [23] and controlled by homemade software with MFCs (mass flow controllers) and a Baratron pressure gauge (MKS Instruments ® , Andover, MA, USA). Figure 2 shows the schematic view of preparing the gas sensors. Furthermore, the deposition parameters, such as pre-sputtering time, deposition temperature, power, etc., were the same for CuO-and M:CuO-based thin films.

Film Characterization
The structural analysis of the films was carried out by an X-ray diffraction technique using PANalytical X'Pert Pro MDP with CuKα (λ = 1.5406 Å) at a step size of 0.04° over the 2Ɵ range of 30-80°. The chemical composition of the films was confirmed by energy dispersive X-ray (EDX) analysis using the FEI VERSA 3D system. The microstructure of the samples was characterized by scanning electron microscopy FEI VERSA 3D.

Gas Sensing
The gas-sensing performance of the deposited and annealed thin films for propane was examined using a homemade computer-controlled system. The gas concentrations were controlled by changing the mixing ratio of dry air and propane using mass flow controllers (MKS Instruments). The total flow rate was set to 500 sccm. The gas sensing response (R) was defined as R = Rgas/Rair, where Rgas and Rair are the electrical resistances in the presence and absence of propane, respectively. Figure 3 shows the XRD patterns of the films after deposition. All of the diffraction peaks were well indexed to the references. Table 1 shows crystallographic parameters for films after the annealing process. As can be notice, the films exhibit mostly a monoclinic or cubic crystallographic system. The silicon substrates used for material characterizations were previously coated by thick Al film (1000 nm) to avoid peaks from silicon for the Si:CuO-based sensor. Due to the characteristics, peaks from Al (200) can be observed. Figure 4 shows the XRD patterns of the films after annealing at 500 °C for 4 h in air. As seen from Figure Table 1. Crystallographic parameters of thin films after annealing at 500 °C for 4 h in air.

Electrical Resistance Measurements
It is well known that electrical resistance decreases when temperature increases for semiconductor gas sensors. The basic approaches for metal oxide gas sensors are widely presented in the literature, e.g., [24,25]. In this study, the electrical resistance of the samples was measured in the amperometric mode by using the Keithley 6517 electrometer. Table 2 shows the baseline resistances in air (RH: 50%) at various temperatures for CuO-and M:CuO-based sensors. The resistance can be written as: R(T) = R0[1 + α (T − T0)], where R0 is the resistance at temperature T0 and α is the resistance temperature coefficient. The calculated temperature coefficient of resistance (TCR) was in the range of 0.0045-0.4400 (1/°C). However, the TCRs of M:CuO films were at least ten times higher than for pure CuO. Cheng et al. [26] have reported the TCR of single-crystalline CuO nanowires to be approximately 0.0075 (1/°C).

Gas Sensing Properties
The dynamic sensing responses were measured using a homemade computer-controlled measurement system. The response (R) was defined as Rgas/Rair, where Rgas and Rair are the resistance of the sensor in propane and air, respectively. The response of CuO-and M:CuO-based (M = Au, Ag, Cr, Pd, Pt, Sb, Si) nanostructure sensors was measured toward 1 ppm C3H8 at 120 °C-380 °C ( Figure 12). It is observed that the gas response of the sensors is greatly influenced by the working temperature due to the temperature-dependent gas adsorption and desorption on the oxide surface. The operating temperature was obtained by applying the power supply to a heater placed inside the sensor substrate. The formation of a uniform temperature distribution in the LTCC gas sensors was previously investigated and reported in [22]. The optimum working temperature was determined at 250 °C for all sensors, except the Ag:CuO-based one, which exhibits a maximum response at 320 °C. Further examining of the gas sensing characteristics has to be performed close to the optimum working temperature to find out the suitable working temperature. As is known, the operating temperature is an important parameter for gas sensors, because it determines the power dissipated by the heater necessary for the achievement of the optimal gas-sensing characteristics, and, through this parameter, influences the reliability and durability of solid-state gas sensors. For practical devices, one wishes to minimize the power needed to operate, so the lowest operating temperature is desired. The power consumption can be also reduced by optimizing the gas sensor substrate dimensions and geometry [22]. In an atmosphere containing flammable gases, a low temperature is favored also for safety reasons. Recently, researchers have investigated the possibility to fabricate sensors working at room temperature, e.g., [27,28]. However, such sensors exhibit low repeatability due to the lower desorption process. The dynamic sensing measurements of CuO-and M:CuO-based sensors were performed to evaluate their gas sensing properties in terms of sensitivity, response and recovery time(s), as well as to determine the reversibility for long-term measurements. The propane-sensing curves show that the CuO-and M:CuO-based sensors exhibit a p-type response to propane. Figure 13 shows the responses (normalized resistance changes) for CuO-, Au:CuO-and Cr:CuO-based sensors (sensors with the highest responses upon exposure to C3H8) toward various propane concentrations measured at 250 °C and at relative humidity ~50%. The resistance of the measured sensors in air is low, which promptly increases and reaches a near plateau upon exposure to propane. The corresponding response of the measured sensors to propane as a function of concentration is shown in Figure 14. With increasing concentration of propane, the responses greatly increase. However, as was mentioned in the Introduction, for medical applications, the responses at the lowest concentration are the most valuable. The sensors have been previously stabilized at the working temperature for 24 h.  Kim et al. [29] investigated a selective detection of NO2 using Cr-doped CuO nanorods. The authors observed that pore volumes over the entire pore size range increased for a higher Cr concentration. This is beneficial for two reasons: first, for the surface area to volume ratio and, second, for the gas accessibility, which can explain the enhanced response of CuO nanostructures [29]. The Cr:CuO-based sensor exhibits higher resistance in air than the CuO-based sensor (Table 2) with the absence of the Cr2O3 peak (Figure 4), which suggests that Cr was incorporated into the CuO lattice. However, a detailed study of the sensing mechanism of Cr:CuO-based sensors would be necessary in order to determine the impact of various Cr concentrations. Based on the literature and the obtained results, the enhancement of responses to propane by Cr doping can be explained in part by the incorporation of Cr 3+ into the CuO lattice and the consequent decrease in its hole concentration.

Response-Recovery Characteristics
Response and recovery times are one of the most important issues that have to be taken into account during the gas sensing measurements; therefore, they are well described in the literature, e.g., [30]. The response and recovery time(s) is the time to reach 90% variation of the sensor resistance upon exposure to an analyte gas and air, respectively. Usually, the times are determined by measuring changes in the electrical resistance from the base level in air (Rair) to a steady level in air containing the analyte gas (Rgas) upon switching the gas atmosphere from air to the analyte gas. However, the dead volume of a conventional chamber is too large to quickly introduce gases into the chamber within a sufficiently short time. The response and recovery time(s) have to be compared to the information of the chamber volume, the temperature, the thickness of the sensitive layer, etc. In this study, the authors used a quartz tube-shaped chamber having a volume approximately 40 cm 3 ± 0.6 cm 3 . Savovic et al. [31] have reported the investigation results on gas diffusion dynamics of the response and recovery processes of a thin film semiconductor gas sensor using an equivalent model of a thin film device and a simple diffusion equation [31]. Park et al. [32] have reported that response and recovery times for the CuO nanocubes with 800 ppb formaldehyde, which were 50 s and 150 s at 250 °C, respectively. Liang et al. [33] have reported that the response time for the CuO-In2O3 nanofiber sensor toward 5 ppm H2S was 150 s at 250 °C. Kim et al. [34] have reported the response and recovery times for the bare CuO and Pd-functionalized CuO nanorods under exposure of H2S (20-100 ppm). The times were in the range of 80-700 s. Briefly, the recovery time of the Pd-functionalized nanorods sensor was 5-8-times shorter than that of the bare CuO nanorod sensor, while the response time was 2-3-times longer [34]. Abaker et al. [35] have reported that the response time for the CuO nanocubes sensor toward 5 × 10 −9 mol·L −1 4-nitrophenol was 10 s [35,36]. Based on a literature review, it is very difficult to directly compare the obtained results with the results presented by others without access to the actual raw experimental results and without considering the target application. Currently, the authors have focused on gas sensors for the portable breath analyzer where the response and recovery time(s) are as much important as sensitivity. Figure 15 shows the response and recovery step of the resistance of the CuO-based, Au:CuO-based and Cr:CuO-based sensors toward 1 ppm C3H8 at 250 °C (RH: 50%). Table 3 Table 3. The response and recovery time (s) obtained for 50-nm films deposited at 100 °C and annealed at 500 °C for 4 h in air under exposure to 1 ppm C3H8 at the temperature with the maximum sensor response (see Figure 11).

Long-Term Stability
There are two adverse effects that may appear when the sensor works for a long period of time, i.e., drift of the baseline signal (defined as the conductance in air or in a reference gas) and drift in the sensor response. The long-term stability measurements are very important for the practical use of a sensor. Therefore, it is necessary to perform many thermal treatments and cycle calibrations before fabrication. Figure 16 shows the baseline resistance variation of CuO-based, Au:CuO-based and Cr:CuO-based sensors for three days at 250 °C in ambient air. The measurement points were fitted by an exponential curve (y = A − B × exp(−kt), where A, B and k are the experimental coefficients) with high values of the coefficient of determination R 2 : 0.90-0.95. It seems that such sensors will be stable enough to conduct several dozens of measurements over a long period of time. Commercially available gas sensors usually require a precondition period from 48 h up to seven days. During the long-term stability measurements, the working temperature was set to 250 °C; relative humidity was stabilized at 50%, and the air flow rate through the gas sensor chamber was set to 500 sccm. Previous experiment results with CuO-based and M:CuO-based thin films have shown good stability [23].

Conclusions
The investigation of metal oxide gas sensors was started over 40 years ago, and now, it seems to be constantly improving. However, the main research goals in MOX (Metal Oxide) gas sensors have been realized, such as improving the sensitivity, selectivity and stability for developed gas sensors. Moreover, the novel potential applications, e.g., in a portable breath analyzer, require extremely short response and recovery time(s) with very good selectivity and sensitivity. Therefore, with such a big variety of materials that can be used, the selection of optimal sensing material becomes a key problem in both the design and manufacturing of gas sensors with the required operation parameters. In this study, the CuOand M:CuO-based sensors (M = Ag, Au, Cr, Pt, Pd, Sb, Si) were presented. The gas-sensing characteristics were discussed for films with the highest response toward 1 ppm of propane, which can be considered as one of the biomarkers of diabetes. The results suggest that the sensing properties of the cupric oxide films are improved by the addition of M-dopants (except Si), which act as catalysts in propane sensors. The Cr:CuO-based structure, annealed at 400 °C for 4 h in air, showed the highest sensor response, of the order of 2.7 at an operation temperature 250 °C. The response and recovery time(s) were: 10 s and 24 s, respectively. Furthermore, a detailed study of the sensing mechanism of the Cr:CuO-based sensor is needed, especially with an impact of various Cr concentrations. All measured sensors were obtained keeping the deposition parameters constant, due to the various dopant concentrations being used. The dopants are capable of improving the sensor properties by the formation and stabilization of smaller grains, by increasing the nanostructure porosity and by enhancement of the long-term stability.