Transparent conducting oxide (TCO) is a material used as a transparent electrode in light-based devices such as flat-panel displays and solar cells that has a high optical transmittance of more than 80 % in visible light and a resistivity of less than 10⁻³ Ω·cm [1–5]. Among TCOs, indium tin oxide (ITO) is the most widely used material at present because it has high optical transmittance and excellent electrical conductivity and is also chemically stable [6–10]. However, as high-efficiency devices such as solar cells, photodiode, and displays have emerged, it has become more important for TCOs to have high transmittance and low resistance simultaneously. Achieving this with the existing pre-made ITO alone is difficult. In general, when the thickness of ITO increases, its optical transmittance decreases but the electrical conductivity increases. As such, there is a limit to simultaneously improving the light transmittance and electrical conductivity, which is crucial for practical applications [11,12].

ITO, which is created by putting Sn in the In position in In₂O₃, has more free electrons inside due to Sn having one more free electron than In, resulting in higher electrical performance. In other words, it is obtained by mixing indium oxide with tin oxide, and is optimized by doping Sn⁺⁴, which has one extra free electron than In⁺³, thereby increasing the number of free electrons to lower its resistivity. By doping Sn to In₂O₃, a material having excellent optical and electrical properties can be produced, but if Sn is inserted in amounts beyond a certain level, a stable oxide such as In₄Sn₃O₁₂, Sn₂O_x, or the like, is formed to suppress the movement of free charge and this deteriorates its electrical performance. The proper Sn content is known to be about 5-10 % [13–19]. When ITO is manufactured, a circular or rectangular target is employed, and the ITO target is manufactured by a sintering method in which raw material powder is mixed in a predetermined ratio, molded and sintered using a dry method or a wet method, and finally ground by a machine [20,21]. Therefore, once manufactured as a target, it is difficult to change the concentration of Sn, and it is also difficult to adjust the concentration of Sn to its limit.

In thisstudy, we propose a strategy for Sn-doping of a commercial ITO film at a composition of 10 wt.% in the form of a bi-layer thin film of ITO/Sn treated by in-situ annealing. To prepare samples to be evaluated we used RF sputtering equipment and then conducted in-situ annealing under a high vacuum. The Sn-doped ITO thin films showed a significant (roughly two-fold) improvement in sheet resistance without the lowering optical transmittance compared to undoped ITO. We present the details of the Sn-doped ITO fabrication and observation of optical and electrical properties according to the growth conditions of different material combinations.

Figure 1 shows a schematic of the sputter system for thin film production [11]. For this study, we used a commercial ITO target prepared with a weight ratio of In₂O₃ and SnO₂ of 90:10 [10 wt.%, ANP], and a target of Sn with a purity of 99.99 % (iTASCO). The base pressure of the sputter chamber was maintained below 6.0 × 10⁻⁶ Torr through overnight pumping. A plasma gas of argon (Ar, 99.999 %) was used at a flow rate of 30 sccm and the sputter process pressure was 10 mTorr. The sputter guns for ITO and Sn were operated by respective RF power supplies operated at 13.56 MHz. The RF power was 100 and 30 W and the target-to-substrate distance was 68 and 75 mm for the 10 wt.% ITO target and Sn target, respectively. In the RF sputtering system, a halogen lamp heater is located at the position of the ITO target. When necessary for annealing, the samples were set under the heater and exposed to the desired annealing temperatures, T_A of 260 and 400 °C for a time duration of 30 min. Table I summarizes the deposition parameters of RF sputtering for both ITO and Sn.

Table 1 . Summary of the growth parameters of 10 wt.% ITO and Sn by RF sputtering..

Parameters	ITO target	Sn target
RF power	100 W	30 W
Target-to-substrate distance	68 mm	75 mm
Growth rate	28.3 nm/min.	0.5 nm/pass
Substrate temperature	260 °C	-
Base pressure	≤ 6.0 × 10−6 Torr	≤ 6.0 × 10−6 Torr
Working pressure	10 mTorr	10 mTorr

Figure 1. Schematic of RF sputter system with in-situ annealing.

Figure 2 schematically illustrates the process of Sn doping into the ITO film performed in this experiment. Prior to thin film deposition, a 20 mm × 20 mm × 0.7 mm soda lime glass (SLG) substrate was ultrasonically cleaned with acetone, ethanol, and de-ionized water for 10 min successively and then dried with N₂ gas. First, a Sn film was deposited at ambient temperature on an SLG substrate, and then an ITO thin film was fabricated on it successively, and finally the bi-layer ITO/Sn film was heat-treated to finish the Sn-doping process through an in-situ annealing process.

Figure 2. Process sequence of Sn doping into ITO thin film.

For fine control of Sn deposition, we rotated the substrate through Sn plasma such that Sn deposition took place only when passing over the target at a constant speed and multiple passes resulted in a thick layer. When an SLG substrate was passed 100 times, we obtained a Sn film of 47.8 nm, for which the Sn growth rate per pass (or rotation) was estimated as 0.5 nm, as determined from the cross-sectional SEM result shown in Fig. 3(a). Figure 3(b) shows a surface morphology consisting of Sn grains densely and completely covering the surface of the glass substrate without pin holes. In this experiment, the amount of Sn sputtering in an ambient temperature was adjusted as follows: t_Sn of 0.0 (no pass over Sn), 1.0 (two passes), and 2.0 (four passes) nm. On the other hand, the ITO deposition was performed inside plasma exposed for a certain time without moving the Sn/SLG substrate. The ITO growth rate was 28.3 nm/min at 100 W and a growth temperature T_G of 260 °C was used for all ITO deposition. Figures 3(c) and 3(d) exhibit that ITO films of 196 and 390 nm respectively prepared with 7 and 14 min deposition.

Figure 3. SEM measurements of cross-section and surface; Sn films after 100 rotations (or passes) (a), (b) for Sn, and ITO thin films (c) for 7 min deposition and (d) for 14 min deposition.

In-situ annealing was performed for the diffusion of Sn into the sample, where a small amount of Sn was laid on the ITO film. The annealing temperature was selected over the melting point of Sn 232 °C, and the optical and electrical characteristics according to the annealing temperature were measured at T_A = 260 and 400 °C.

As for sample characterization, the film growth morphology was analyzed using a field emission scanning electron microscope (SEM) (Jeol, JSM-7610F). We also examined the crystallographic properties of the films by using high-angle X-ray diffraction (XRD, Rigaku diffractometer) with CuKα radiation (λ = 1.54 Å). The optical transmittance was investigated in a wavelength of 300 to 1400 nm using UV-visible spectroscopy (Hitachi, UH4150; Shimadzu). From the optical spectra, the average visible transmittance (AVT) was evaluated in a wavelength range of 380 to 780 nm. The electrical properties of the Sn-doped ITO thin film were measured using a 4-point probe (Ecopia, HMS-3000).

In order to determine the thickness effect for ITO crystal growth, films of ITO without adding extra Sn were investigated by using XRD. The XRD patterns of the ITO thin films with T_ITO of 200 and 400 nm grown at T_G = 260 °C are shown in Fig. 4. Figure 4(a) presents the XRD spectrum for the ITO film of 400 nm in which the crystal orientations of the (211), (222), (400), (411), (431), (440), (611), and (622) planes are well defined, reflecting the polycrystalline structure of ITO. The dominant crystal growth was observed in the direction of (222) planes, as observed in a previous study [22,23]. As the layer thickness was reduced to 200 nm, however, the crystal growth of ITO exhibited a slight indication of preferred crystal growth in the atomic (222) plane but dominantly showed an amorphous phase without clear x-ray reflections, as shown in Figure 4(b). The XRD feature of ITO of 200 nm was similar to that of the bare glass amorphous sample, as seen in Fig. 4(c).

Figure 4. X-ray diffraction spectra for the samples of as-received 10 wt.% ITO film; (a) T_ITO = 400 nm, (b) T_ITO = 200 nm, and (c) bare SLG substrate.

When a certain amount or more of Sn is doped into a film prepared from an as-received 10 wt.% ITO target, the oxide would interrupt the flow of electrons, thereby lowering the conductivity. In consideration of this, optical and electrical characteristics in accordance with the thicknesses of the Sn thin film and the ITO thin film were also compared by doping Sn in ITO of both 170 and 400 nm, respectively.

A TCO film such as ITO should have high light transmittance. Although the sheet resistance would be lowered through Sn doping, it is necessary to investigate the effect of Sn doping on the transmittance of thin films. Figure 5 presents the optical transmittance spectra of ITO films depending on the annealing condition of layer thickness T_ITO and in-situ annealing temperature T_A for a fixed time period of 30 min.: (a) for type A sample T_ITO = 170 nm, T_A = 260 °C, (b) for type B sample T_ITO = 170 nm, T_A = 400 °C, and (c) for type C sample T_ITO = 400 nm, T_A = 400 °C. During preparation of the thin films, the substrate temperature for all the samples was kept at the growth temperature, T_G = 260 °C. The transmittance value of Sn-doped ITO was deconvoluted from the effect of the SLG substrate; i.e., only transmittance through the ITO film was taken into account. As the layer thickness increases, all the samples show an undulation in the transmittance spectra due to the Fabry-Perot resonance of the thin film. Unlike the type A (T_ITO = 170 nm, T_A = 260 °C) sample for annealing, the leading edge of transmittance exhibits a blue-shift with T_A = 400 °C for both T_ITO = 170 and 400 nm, as shown in Figs. 5(b) and 5(c). This result indicates that in-situ annealing at high temperature initiates Sn doping into ITO more actively for T_A > 260 °C. In addition, the Sn doping resulted in a reduction of transmittance for the wavelength λ ≥ ~900 nm. As Sn is added from no Sn to 2.0 nm, the transmittance exhibited a significant reduction, as seen in Fig. 5(c) for Sn (2 nm).

Figure 5. Optical transmittance spectra vs. wavelength for Sn thicknesses at different annealing conditions: (a) type A sample (

From the optical transmittance spectra in Fig. 5, we evaluated the AVT in the range of 380 nm ≤ λ ≤ 780 nm. Figure 6 presents the result of AVT vs. Sn thickness, t_Sn = 0.0, 1.0, and 2.0 nm. As expected, the AVTs at ITO with low thickness of 170 nm show higher values of ≥ 92 % at all doping conditions. Even for the type C (T_ITO = 400 nm, t_Sn = 2.0 nm) sample, the AVT was as high as 89.4 % for Sn-doped ITOs, which is the smallest value in this experiment. On the other hand, as presented in Fig. 6, the AVT was found to be independent of the degree of Sn doping and annealing temperature. Rather, it showed only a dependence on the ITO thickness, T_ITO. When the ITO layer thickness was increased from 170 to 400 nm, the AVT dropped by approximately 3.6 % with respect to case of 170 nm with no Sn doping.

To investigate the electrical properties of Sn-doped ITO, we employed Hall measurement using a 4-point probe (Ecopia, HMS-3000). Table II is a summary of all the properties according to the ITO thickness and annealing temperature. In Fig. 7, the electrical properties of (a) carrier concentration, n, (b) electron mobility, µ and (c) resistivity, ρ are plotted as a function of Sn thickness, t_Sn. As shown in Fig. 7(a), the carrier concentration n is measured as (3.01~21) × 10²⁰ /cm³ and increased with Sn doping. For the type A (170 nm, 260 °C) and type C (400 nm, 400 °C) Sn-ITO samples, the concentration increased monotonically with Sn doping while the type B (170 nm, 400 °C) sample exhibited the largest value for doping of t_Sn = 1.0 nm.

Table 2 . Summary of the optical and electrical properties of Sn-doped ITO thin film. All the samples showed negative polarity for n-type material..

Type	tITO [nm]	TA[°C]	tSn [nm]	AVT [%]	n(×1020) [/cm3]	µ[cm2/V•s]	ρ(×10−4) [Ω.cm]	R□[Ω/□]
A	170	260	0.0	92.3	6.41	22.66	4.30	23.86
			1.0	93.1	15.0	19.91	2.09	12.27
			2.0	92.3	21.0	11.54	2.51	14.79
B	170	400	0.0	92.6	7.27	18.61	4.61	30.74
			1.0	93.8	18.9	18.96	1.74	10.26
			2.0	93.3	8.37	30.42	2.45	14.41
C	400	400	0.0	89.6	3.01	34.22	6.05	15.14
			1.0	90.6	5.20	36.06	3.33	8.32
			2.0	89.4	8.68	30.57	2.35	5.88

For the mobility, asshown in Fig. 7(b), the undoped ITO prepared from the as-received 10 wt.% ITO target showed different values of 22.66, 18.61, and 34.22 cm²/V·s for type A, type B, and type C, respectively. The highest value of mobility was observed in the type C sample prepared at (400 nm, 400 °C). As t_Sn varied for 1.0 and 2.0 nm, the mobility showed a decrease in the type A and type C samples of -49.1 and -10.7 %, respectively. On the other hand, the type B sample displayed an increase of as much as 63.5 %. We believe that the higher mobility in the type C sample resulted from the sufficient doping of Sn atoms into the undoped 10 wt.% ITO film.

The resistivity of undoped ITO (i.e., t_Sn = 0.0 nm) was observed as 4.30 × 10⁻⁴, 4.61 × 10⁻⁴, and 6.05 × 10⁻⁴ Ω·cm for type A, B, and C, respectively. As the Sn doping increased, all the samples exhibited a large reduction in the resistivity, particularly for the type C sample. After t_Sn = 2.0 nm doping, all three samples converged to almost the same value of ~2.45 × 10⁻⁴ Ω·cm, regardless of the condition of the undoped ITO thickness T_ITO and the annealing temperature T_A.

Figure 8 presents graphs of the sheet resistance R_□ vs. the ITO film condition for (a) a fixed ITO thickness of T_ITO = 170 nm and different annealing temperature of T_A = 260 °C (type A), 400 °C (type B) and (b) the same annealing temperature T_A = 400 °C and different ITO thickness of T_ITO = 170 nm (type B) and 400 nm (type C). In Fig. 8(a), R_□ showed similar behavior for different T_As at a fixed thickness of T_ITO = 170 nm. It reduced with Sn doping by approximately -57 % with respect to that of non-doped ITO and then showed a slight return to -50 % with further Sn doping of t_Sn = 2.0 nm. This represents that an additional Sn-doping into the 10 wt.% ITO film is very effective for reducing R_□ in association with annealing at high temperature, T_A = 260 and 400 °C. R_□ for T_ITO = 170 nm, T_A = 400 °C yielded 10.26 Ω/□ with an Sn of 1.0 nm and then slightly increased to 14.41 Ω/□ with an Sn of 2.0 nm.

In addition, we studied the change of R_□ at the same T_A = 400 °C with different ITO thicknesses, as presented in Fig. 8 (b): T_ITO = 170 nm (type B), 400 nm (type C). For the sample of T_ITO = 400 nm without Sn doping, we observed an initial value of R_□ = 15.14 Ω/□, which was a better result compared to that of T_ITO = 170 nm. The sample also showed reduced optical transmittance, as indicated in Fig. 5(c), resulting from the layer thickness of ITO. As the Sn doping to 10 wt.% ITO was increased to t_Sn = 1.0 and 2.0 nm, the change in R_□ with respect to that of non-doped ITO was reduced by -45 and -60 %, respectively. The smallest value of sheet resistance was R_□ = 5.88 Ω/□, obtained from a type C sample for T_ITO = 400 nm, T_A = 400 °C. From this experiment, we concluded that Sn doping of a 10 wt.% ITO film was a very effective way of greatly improving electrical properties (by ≤ -57 %) at the expense of a small reduction in the AVT of 3.6 %.

In this doping experiment, we observed that high annealing temperature accelerated the diffusion of added Sn into the ITO matrix material. Figure 8(a) shows the lowered sheet resistance for the samples of T_ITO = 170 nm at t_Sn = 1.0 nm when the annealing temperature T_A was increased from 260 to 400 °C. This indicates that the diffusion of Sn was promoted due to the increase in the annealing temperature. A slight increase in R_□ at t_Sn= 2.0 nm even at higher annealing temperature is attributed to the formation of a stable oxide due to an increase of the internal Sn concentration in the ITO film [13–19]. Therefore, any further increase in the Sn content could lead to a decrease in the sheet resistance, R_□. On the other hand, for a monotonic decrease in R_□ for a sample of T_ITO = 400 nm, as presented in Fig. 8(b), we believed that the diffused Sn may be accommodated completely as the ITO thickness is increased. In other words, when the thickness of the ITO film and the amount of Sn are optimized, the sheet resistance R_□ can be lowered further.

In this study, we fabricated bi-layer thin films of ITO/Sn using RF sputtering and performed an in-situ annealing process at high temperature to enhance the doping of Sn into ITO thin films. The optical and electrical characteristics of Sn-doped ITO thin films prepared with different thickness and annealing temperature were investigated and analyzed according to the change of the added Sn film thickness. The results of this experiment show that it is possible to adjust the Sn concentration inside ITO to the limit while growing the ITO thin film. In order to allow use as a transparent electrode for opto-electronic devices, the optical transmittance must be high and the sheet resistance must be low at the same time. For the sample of Sn-doping to 10 wt.% ITO with T_ITO = 400 nm, t_Sn = 2.0 nm annealed at 400 °C, the obtained average visible transmittance was as large as 89.4 % in a range of 380 to 780 nm. In addition, the change in R_□ with respect to that of non-doped ITO was reduced for t_Sn = 1.0 and 2.0 nm by -45 and -60 %, respectively. The smallest value of sheet resistance was R_□ = 5.88 Ω/□. The results are suitable for use in transparent electrodes for most opto-electronic devices. We conclude that Sn-doping to a 10 wt.% ITO film was a very effective way of greatly improving the electrical properties by ≤ -57 % at the expense of a small reduction in the AVT of 3.6 %.

This research was partially supported by Cheongju University Research Scholarship Grants in 2022. This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (NRF-RS-2022-00143178). This work was partly supported by a Korea Institute of Energy Technology Evaluation and Planning (KETEP) grant funded by the Korean government (MOTIE) (No. 20224000000070, Human Resource Training For Smart Energy New Industry Cluster).

The authors declare no conflicts of interest.