Analysis of the Effect of Copper Doping on the Optoelectronic Properties of Indium Oxide Thin Films and the Thermoelectric Properties of an In₂O₃/Pt Thermocouple

Abstract

The detection and real-time monitoring of temperature parameters are important, and indium oxide-based thin film thermocouples can be integrated on the surface of heaters because they operate normally under harsh conditions and provide accurate online temperature monitoring. The higher stability and appropriate optical and electrical properties of In₂O₃ make it very suitable as an electrode material for thermocouple sensors. This work demonstrates that copper doping can alter the optical and electrical properties of In₂O₃ films and regulate the output performance of thermocouples. Copper-doped In₂O₃ thin films were prepared using the magnetron co-sputtering method. The doping concentration of Cu was controlled using direct current (DC) power. An In₂O₃/Pt thermocouple sensor was prepared, and the optoelectronic and thermocouple properties were adjusted by changing the copper doping content. The thickness valve of the thin film sample was 300 nm. The results of the X-ray diffraction suggested that the structure of the doped In₂O₃ thin films was cubic. The results of the energy-dispersive X-ray analysis revealed that Cu was doped into the In₂O₃ thin films. All deposited films were n-type semiconductor materials according to Hall effect testing. The 4.09 at% Cu-doped thin films possessed the highest resistivity (30.2 × 10⁻³ Ω·cm), a larger carrier concentration (3.72 × 10²⁰ cm⁻³), and the lowest carrier mobility (0.56 cm²V⁻¹s⁻¹). The optical band gap decreased from 3.76 to 2.71 eV with an increase in the doping concentration, and the transmittance of the film significantly reduced. When the DC power was increased, the variation range of Seebeck coefficient for the In₂O₃/Pt thermocouple was 152.1–170.5 μV/°C, and the range of thermal output value was 91.4–102.4 mV.

1. Introduction

Transparent conductive oxide (TCO) thin films have a significant research value and are widely used in many applications such as solar cells, display devices, photoelectronic devices, and touchscreens [1,2,3,4]. Metal oxides commonly used in TCO thin film systems include cadmium oxide (CdO), zinc oxide (ZnO), tin oxide (SnO₂), and indium oxide (In₂O₃) [5,6,7,8]. The demand for TCO thin films has increased over the last few years with the rapid development of various devices. In₂O₃ is an n-type semiconductor. Its band gap is 3.5–3.7 eV, and it is widely used in transparent conducting electrodes, solar cells, displays, and photoelectric sensors [9,10].

The electrical and optical properties of an In₂O₃ thin film can be improved by doping metal elements such as Sn, Cr, Mo, Ce, Sb, Li, Zn, and Cu [11,12,13,14,15,16,17,18,19,20,21,22]. Pramod et al. prepared Sb-doped In₂O₃ films [18]; the highest optical transmittance was obtained at 1.5 at.%, and the doping amount was able to change the optical band gap of the In₂O₃ films. Ahmed et al. prepared (In_1−xCr_x)₂O₃ thin films under different Cr doping concentrations [16]. As the Cr doping concentration increased, the films preferentially oriented along the (222) direction under different component conditions, and the optical band gap value increased from 3.21 eV to 3.46 eV. Shi et al. prepared Ce-doped In₂O₃ thin films under low-temperature conditions using radio frequency reactions [17]. According to an electrical performance test, the carrier concentration of the films reached 2.1 × 10²⁰ cm⁻³, and the highest mobility obtained was 153.7 cm²/Vs. Wang et al. prepared In₂O₃ thin films with a resistivity of 3.76 × 10⁻⁴ Ω·cm by doping Hf [18]. Their study revealed that the Hall mobility could reach 79.6 cm²/Vs, and the average transmittance of the thin film could reach 83% in the range of 300–1500 nm. Indium–gallium–zinc oxide (IGTO) thin films on flexible substrates were developed by Kang et al. using co-doping with Ga and Ti when the oxygen flow was zero [19]. The sheet resistance of IGTO was 39.3 Ω/sq, and the transmittance of the IGTO film reached 86.96%. Kaleemulla et al. prepared Mo-doped In₂O₃ thin films on glass substrates using the activated reactive evaporation method [20]. The films synthesized at a Mo doping level of 3% and a substrate temperature of 573 K. The electrical resistivity achieved a minimum value of 5.2 × 10⁻⁴ Ω·cm, the average optical transmittance of the films reached 90%, and the band gap was 3.68 eV. Fan Ye et al. prepared Cu-doped In₂O₃ using the reactive DC magnetron sputtering method [21]. The results revealed that Cu doping could significantly alter the electrical and optical properties of thin films. The resistivity of the doped In₂O₃ thin films reached 1.82 Ω·cm, and the minimum optical band gap could be reduced to 3.70 eV. Otto J. Gregory constructed Cu-In-O thin films [22], and found that in a composition of 40.0 atom % indium, thin films can transition from p-type to n-type conduction.

The common methods used In the preparation of In₂O₃ thin films include sol–gel, pulsed laser deposition (PLD), metal–organic chemical vapor deposition (MOCVD), and magnetron sputtering [23,24,25,26,27,28]. Gupta et al. prepared Gd-doped In₂O₃ films on quartz substrates using the PLD method [23]. Under a vacuum and 600 °C growth conditions, high-conductivity indium oxide thin films were obtained with a carrier concentration of 1.74 × 10²⁰ cm⁻³, a resistivity of 2.80 × 10⁻⁴ Ω·cm, and a mobility of 1.28 × 10² cm²V⁻¹s⁻¹. Sanchez et al. prepared In₂O₃ thin films on the c-axis of sapphire substrates using a low-temperature sol–gel method [24]. They observed an increase in the carrier concentration of the films to 2.45 × 10¹⁹ cm⁻³, with a resistivity of 3.5 Ω·cm noted after heat treatment at 700 °C. Baqiah et al. prepared In₂O₃ thin films using the CVD method, with InCl₃·4H₂O as the initial raw material [25]. The resistivity of the films in their study reached 1.458 × 10⁻² V·cm at a deposition temperature of 720 K. According to Du et al., the resistivity of In₂O₃ thin films prepared using the MOCVD method could reach 1.458 × 10⁻² V·cm at a deposition temperature of 720 K [26]. Compared with other preparation methods, the magnetron sputtering method is widely used in certain areas because of its superiority in terms of uniformly deposited films, its high purity, and wide application to target materials. Cho used a radio frequency reaction method to prepare In₂O₃ thin films on target materials [27]. The films were preferentially oriented along the (222) and (400) directions, and the average transmittance reached 88.6% in the range of 370–1100 nm. Yuan et al. prepared In₂O₃ thin films on amorphous glass using the magnetron sputtering method [28]. A high field-effect mobility of thin film transistors was obtained by optimizing the annealing process.

Due to its excellent stability and significant thermoelectric potential output under high-temperature conditions, indium oxide is often used as an electrode material for high-temperature thermocouples [29,30,31]. Its thermoelectric output characteristics can be adjusted via doping with indium oxide thin films. Liu Dan et al. prepared indium oxide thin films using radio-frequency magnetron sputtering, and constructed La₀.₈Sr₀.₂CrO₃/In₂O₃ thermocouples [32]. The thermoelectric potential output reached 305.8 μV/°C at 1100 °C. By doping Sn to adjust the thermoelectric output characteristics of indium oxide thin films, Xiaohui Zhao et al. used magnetron sputtering to construct ITO/Pt thin film thermocouples on nickel alloy substrates. A high-temperature tolerance of up to 1100 °C was recorded [33]. To improve the stability of thermocouples, Zhongkai Zhang et al. constructed In₁.₃₅ZnO₂.₁₁/In₂O₃ thermocouples using screen-printing technology. This resulted in a thermoelectric potential output of 39.8 μV/°C at 1500 °C. The drift rate was 0.84 °C/h [34]. Zhao X.H et al. prepared ITON-InON ceramic thin film thermocouples [35]; the samples were heated at 1000 °C for 5 h in a nitrogen-rich atmosphere and 5 h in air, respectively, which improved the output stability of the thermocouple, and the average Seebeck reached 64.7 μV/°C. Liu et al. treated indium oxide thin films at a high temperature of 1150 °C for 10 h, and the In₂O₃/ITO thermocouple was able to output normally [36].

In this study, we prepared copper-doped thin films using the co-sputtering method, achieved indium oxide thin films with adjustable transmittance and electrical properties, and obtained thin film thermocouples with an adjustable output thermal voltage. We investigated the electrical and optical properties of undoped In₂O₃ and Cu-doped In₂O₃ thin films, the effect of the Cu ion concentration on the microstructure, and the electrical and optical properties of In₂O₃. The thermoelectric properties of the In₂O₃/Pt thermocouple were systematically characterized.

2. Materials and Methods

In₂O₃ was created on quartz substrates with different concentrations of copper doping using magnetron co-sputtering equipment (SP3-80C, Chuangshiweina Technology Co., Beijing China). We altered the direct current (DC) sputtering power during the preparation process to obtain the different Cu-doping concentrations. The samples of In₂O₃ with different Cu-doping concentrations (sputtered by DC power) were prepared at room temperature using DC sputtering power supplies of 10 W, 15 W, 20 W, and 25 W. The diameter of all In₂O₃ and Cu targets was 80 mm, and the purity of both the In₂O₃ and Cu targets was 99.99%. The In₂O₃ target was set at a sputtering power of 100 W using a radio-frequency power supply throughout the entire sputtering process. Before sputtering, substrates with a size of 10 mm × 10 mm and a thickness of 0.7 mm were cleaned for 15 min in an ultrasonic bath using alcohol and acetone separately. These were then repeatedly rinsed using deionized water and dried with a high-pressure nitrogen gun. The base pressure of the chamber was 6.0 × 10⁻⁴ Pa, and the sputtering pressure was 1.5 Pa. The substrates were rotated at 14 revolutions per minute (rpm) during the deposition process. The oxygen-to-argon ratio was set at 1:6, the argon gas flow rate was maintained at 60 standard cubic centimeters per minute (sccm), and the sputtering time was set at 3 h. After the sputtering process, the thin films were annealed in a high-temperature atmosphere furnace (OTF-1200X; MTI, Hefei, China) at 500 °C in an air atmosphere for 1 h. The parameters of the indium oxide and copper targets during co-sputtering are presented in

Table 1. Parameters of indium oxide and copper targets during co-sputtering.

Target	Sputtering Method	Sputtering Power (W)	Sputtering Time (h)	Argon Gas Flow Rate (sccm)	Oxygen Flow Rate (sccm)
In2O3	RF	100	3	60	10
Cu	DC	0–25	3

, and the magnetron co-sputtering of In₂O₃ doped with Cu is diagrammatically represented in Figure 1.

The pure and doped In₂O₃ thin films were tested using an X-ray diffractometer (XRD) (Ultima IV, Rigaku, Hitachi, Tokyo, Japan). The surface images and EDX patterns of the thin film samples were determined using an SEM (Quanta 250, FEI, Hillsboro, OR, USA) equipped with an energy-dispersive spectrometer. The surface topography of the thin films was characterized by an AFM (Dimension Icon, Bruker, Billerica, MA, USA). The electrical and optical properties of the In₂O₃ thin films were investigated using a Hall effect testing system (CH300, CH-Magnetoelectricity Technology, Beijing, China) and an ultraviolet-visible spectrophotometer (Lamda950, PerkinElmer, Waltham, MA, USA), respectively.

To obtain the Seebeck coefficient of the pure and doped In₂O₃/Pt thermocouples, all of the films were fabricated (Figure 2). The length, width, and thickness of the quartz substrate were 10 cm, 10 mm, and 1.5 mm, respectively. After the deposition of In₂O₃ and Pt thin films, an In₂O₃ thin film with a width of 2.5 mm and a length of 9.5 cm was obtained. The Pt electrode was 2.5 mm, with a length of 9.5 cm. The overlapping area of the two electrodes was 2.5 mm × 2.5 mm. The preparation process is presented in Figure 3. First, pure and Cu-doped In₂O₃ thin films were fabricated using the co-sputtering method and a stainless steel metal mask to form an indium oxide electrode pattern. Second, the pure and Cu-doped In₂O₃ thin films were annealed in an air atmosphere at 500 °C for 1 h. Third, the Pt electrode was obtained using a sputtering system (JPG560, SKY Technology Development Co., Ltd., China) and a metal mask to form a Pt electrode pattern. The hot-end overlap area was 2.5 mm × 2.5 mm. The Pt target had a diameter of 4 inches, and the thickness was 4 mm. The instrumental settings of the sputtering process during the formation of the Pt electrode were a sputtering pressure of 0.5 Pa, a sputtering power of 100 W, and an argon flow rate of 80 sccm. The thickness of the obtained Pt thin film was approximately 300 nm.

As presented in Figure 4, the Seebeck coefficient and thermal voltage test system mainly consisted of one heating furnace, one voltmeter, two thermometers, and one cooling plate. A Nabertherm high-temperature heating furnace (LHT 2-17, Nabertherm, Lilienthal, Germany) was adapted to heat the hot end of the sample. The measurement range was set from room temperature to 600 °C, and the length of the K-type thermocouple was 20 cm. This was inserted into the high-temperature furnace to test the temperature of the hot end of the sample. Alumina foam transfer and high-temperature asbestos material were used between the hot and cold ends to insulate the treatment. The hot end of the film sample was pushed into the heating furnace to obtain the high-temperature difference between the two ends. An alumina foam brick was used as insulation material for the furnace door. A voltmeter was used to measure the output voltage of the cold end. Standard K-type thermocouples were used to measure the temperature of the hot and cold ends. The cold end was cooled by circulating water throughout the whole process, and the voltmeter was used to display the value of the thermal voltage.

3. Results

3.1. Microstructure and Other Physical Characteristics of Cu-Doped In₂O₃ Thin Films

The phase structure of the prepared thin film samples was determined using the XRD patterns obtained from 25° to 65°, with a 0.15418 nm wavelength of Cu K-alpha radiation. As demonstrated in Figure 5a, the cubic phase of In₂O₃ was observed in all thin film samples, and the space group was Ia$\overline{3}$. No other peaks were observed, which implied that Cu doping had not added impurities, and that Cu was well incorporated into the lattice of the cubic In₂O₃ and had replaced the In atom. However, two new crystal phases with peaks of (431) and (611) simultaneously appeared. Figure 5b demonstrates the gradual shift in the peak (622) orientation of the doped thin film sample with an increase in the doping concentration. A similar observation was reported by Ye et al. [21]. The atomic radius of Cu is smaller than that of an In crystal; the substitution of smaller Cu for larger In could have changed the phase structure of the crystal, thereby shifting the corresponding diffraction peaks [37].

Figure 6 presents the crystal size values of the Cu-doped In₂O₃ films obtained using different DC sputtering powers. The grain size was estimated according to the Scherrer formula [38] (Equation (1)):

$$D=0.9\lambda /\beta \cos \theta$$

(1)

where λ is the X-ray wavelength, θ is the Bragg diffraction angle, and β is the full width at half maximum of the (622) peak. The increase in the DC sputtering power increased the full width at half maximum (FWHM) of the Cu-doped In₂O₃ films. The size of the crystallite increased from 17.1 nm to 38.9 nm when the DC sputtering power increased from 0 W to 25 W. The maximum value was obtained at DC 25 W. This was mainly attributed to the increase in the sputtering power and the increased number of grain boundary defects.

Figure 7a shows the the surface topography of unannealed pure film, and Figure 7b–d present the surface topography of the Cu-doped In₂O₃ thin film samples under 0 W, 15 W, and 25 W of DC power, respectively. Figure 7b–d reveal that the surface of the film was uniform. A distinct variation in the grain size was absent with an increase in the doping concentration. As presented in Figure 7e, the thickness of the prepared thin film sample was 300 nm. Figure 8 presents the AFM images at different DC sputtering powers; the effects of the Cu-doping concentration on the structural and surface morphology properties of the Cu-doped In₂O₃ thin films can clearly be observed. The Cu concentration evidently influenced the surface morphology of the Cu-doped In₂O₃ thin films. As the DC sputtering power increased, the grain size increased. This was the same trend of change as calculated from the XRD results. Figure 9 presents the surface roughness. The root mean square (RMS) increased from 4.3 nm to 11.1 nm as the DC sputtering power increased from 0 W to 25 W.

An energy-dispersive X-ray (EDX) analysis was used to further study the element types and contents of the prepared thin films. The EDX patterns of all samples prepared at DC sputtering powers from 0 W to 25 W are presented in Figure 10. The EDX data clearly demonstrate that the Cu doped into In₂O₃, and that there was a significant increase in the Cu content with an increase in the sputtering power from 0 W to 25 W. This could be attributed to the participation of Cu in an efficient substitution for indium (In) atoms. The calculated contents of Cu were 1.41%, 1.98%, 4.09%, and 4.26% for the thin film samples prepared at sputtering powers of 10 W, 15 W, 20 W, and 25 W, respectively (Figure 11).

3.2. Electrical Properties of Cu-Doped In₂O₃ Thin Films

The resistivity (ρ), carrier concentration (n), and mobility (μ) of all prepared thin film samples were determined using a Hall effect meter at room temperature, and are presented in Figure 12. The Hall effect measurement confirmed that the introduction of Cu did not change the n-type semiconducting behavior of the In₂O₃ films. When the sputtering power increased to 15 W, the resistivity also increased from 1.85 × 10⁻³ Ω·cm to 5.17 × 10⁻³ Ω·cm. The carrier concentration remained almost unchanged (15.3 × 10¹⁸ cm⁻³) between 0 W and 15 W DC sputtering powers. At a DC sputtering power of 20 W, relatively large changes in resistivity, carrier concentration, and mobility occurred; the values reached were 30.4 × 10⁻³ Ω·cm, 3.72 × 10²⁰ cm⁻³, and 0.56 cm²V⁻¹s⁻¹, respectively. The change in the carrier concentration could have occurred for two reasons. It may have been due to the decrease in the band gap of the thin films causing a greater migration of carriers to the conduction band, thereby increasing the carrier concentration. The other reason may have been that the doping of the acceptor Cu could have generated holes to compensate for a negative charge [39,40]. The mobility tended to significantly decrease with an increase in the doping concentration.

3.3. Optical Properties of Cu-Doped In₂O₃ Thin Films

The transmittance spectra of the Cu-doped In₂O₃ films are presented in Figure 13. A clear decrease in transmittance in the visible light region was observed as the Cu sputter power increased from 0 W to 25 W. The average transmittance revealed an overall downward trend. The pure In₂O₃ sample was 96% and the doping concentration under 10 W DC power was 85%, whereas the doping concentration decreased to 4.2% under a DC power of 25 W. Consequently, the doping concentration of Cu should not be too high during the preparation of Cu-doped In₂O₃ films.

The optical band gap is an important parameter when studying film properties. The optical band gap of the prepared thin film samples was determined to understand the influence of the doping concentration. The optical band gap was calculated using Equation (2) [41]:

$${\left(\alpha h\nu \right)}^2=A(h\nu -E_g)$$

(2)

where hv is the photon energy. The value of A depends on the direct or indirect band gap. α is the absorption coefficient, which can be determined using a formula that ignores the reflection of light on the film (Equation (3)) [42]:

$$\alpha =\ln \left(\frac{1}{T}\right)/d$$

(3)

where d is the thickness of the film (the thickness value was 300 nm) and T is the transmission of the film (as demonstrated in Figure 12). The plot of (αhv)² vs. hv can then be obtained (as presented in Figure 13) by extending the linear section of the curve to the x-axis. The optical band gap can then be determined from the intersection. As presented in Figure 14, the band gap (3.76 eV) obtained for the undoped sample was very close to the values reported in the literature [42] and the doped thin film samples revealed a decrease in the band gap with an increase in the Cu content. Several studies have observed that adding an element to an oxide can narrow the energy band gap of the doped material [43]. Our results also revealed a narrowing of the forbidden band of In₂O₃ from doping with Cu.

3.4. Thermoelectrical Properties of Cu-Doped In₂O₃ Thin Films

To measure the output characteristics of the In₂O₃/Pt thermocouple, all electrodes were connected using enameled copper wire with a diameter of 0.2 mm and a length of 10 cm. The wire was bonded with silver paste; the contact was between the wire and the two electrodes. Figure 15 presents the thermoelectrical output voltage of the Cu-doped In₂O₃ films under different DC powers from room temperature to 600 °C. The output value was 102.4 mV at 0 DC power and a temperature difference of 600 °C. The absolute value of the output thermal voltage of the sample under 20 W DC power was 91.4 mV. The Seebeck coefficients of the Cu-doped In₂O₃ thin films were tested as a function of the hot-end temperature. The Seebeck coefficient was obtained using Equation (4) [44]:

$$S=-\frac{\Delta V}{\Delta T}=-(\frac{\Delta V_a-\Delta V_b}{T_h-T_c})$$

(4)

where S is the Seebeck coefficient, ∆V is the voltage difference between the two different electrode materials, and ∆T is the temperature difference between the hot end (T_h) and the cold end (T_c). As In₂O₃ is an n-type semiconductor, its output is negative; thus, the absolute value of the Seebeck coefficient of the Pt electrode was much smaller than that of the indium oxide material [45]. Figure 16 presents the trends of the Seebeck coefficient with the temperature variation. The Seebeck coefficient of the Cu-doped In₂O₃ films increased with an increase in the temperature difference. When the temperature difference between the two ends of the thin film samples reached 600 °C, the Seebeck coefficient increased to 170.5 μV/°C under 0 DC power. The Seebeck coefficient of Cu-doped In₂O₃ thin films under 20 W DC power reached 152.1 μV/°C; it reached 155.6 μV/°C under 25 W DC power. The Seebeck coefficient of the In₂O₃ thin films could be expressed by Equation (5) [44]:

$$S(N_D)=-\frac{Ak}{e}-\frac{k}{e}(\frac{{(2\pi m_c^{*}kT)}^{\raisebox{1ex}{$2$}\!\left/ \!\raisebox{-1ex}{$3$}\right.}}{h^3{N}_D})$$

(5)

where S is the Seebeck coefficient, k is the Boltzmann constant, e is the electron charge, m_c is the effective mass, h is the Planck constant, and A is the transport constant. The carrier concentration of the doped In₂O₃ thin films clearly changed due to different doping levels. This was observed from the electrical performance results of the Cu-doped In₂O₃ thin films. Under a DC power of 20 W, the carrier concentration reached its maximum value, resulting in the smallest overall Seebeck coefficient of the corresponding sample. The carrier concentration was the main factor affecting the variations in the Seebeck coefficients in the temperature range from room temperature to 600 °C. Figure 17 presents the output voltage of the Cu-doped In₂O₃ films under 20 W DC power for three heating processes. In the third heating process, the output value reached 89.3 mV at a 600 °C temperature difference; the Seebeck coefficient reached 148.8 μV/°C. The measurement results of the three heating processes indicated that the thermocouple had good repeatability.

4. Conclusions

In summary, this paper presents the preparation as well as the morphological, electrical, and optical properties of pure and Cu-doped In₂O₃ thin films fabricated using the magnetron co-sputtering process. The doping of copper effectively adjusted the optoelectronic properties of the In₂O₃ thin films and significantly impacted the output performance of the thermocouple. When the Cu atomic content increased to 4.09% (under 20 W DC sputtering power), the Cu-doped thin films possessed the highest resistivity (30.2 × 10⁻³ Ω·cm) and carrier concentration (3.72 × 10⁻¹⁹ cm⁻³) as well as the lowest carrier mobility (0.56 cm²V⁻¹s⁻¹). Copper doping resulted in a significant reduction in the optical band gap. Consequently, the band gap of Cu-doped In₂O₃ was narrower than that of undoped In₂O₃. The optical band gap was reduced to 2.71 eV when the DC power was 20 W. The Seebeck coefficient of the undoped In₂O₃/Pt thermocouple reached a maximum of 170.5 μV/°C and an output value of 106.5 mV at a high-temperature range of 600 °C. When the DC sputtering power was 20 W, the output value of the sample was 91.4 mV and the Seebeck coefficient was 152.1 μV/°C.