Flexible transparent heteroepitaxial conducting oxide with mobility exceeding 100 cm2 V−1 s−1 at room temperature

Flexible and transparent applications have become an emerging technology and have shifted to the forefront of materials science research in recent years. Transparent conductive oxide films have been applied for flat panel displays, solar cells, and transparent glass coatings. However, none of them can fulfill the requirements for advanced transparent flexible devices, such as high-frequency applications. Here, we present a promising technique for transparent flexible conducting oxide heteroepitaxial films: the direct fabrication of epitaxial molybdenum-doped indium oxide (IMO) thin films on a transparent flexible muscovite substrate. An n-type epitaxial IMO film is demonstrated with a mobility of 109 cm2 V−1 s−1, a figure of merit of 0.0976 Ω−1, a resistivity of 4.5 × 10−5 Ω cm and an average optical transmittance of 81.8% in the visible regime. This heteroepitaxial system not only exhibits excellent electrical and optical performance but also shows excellent mechanical durability. Our results illustrate that this is an outstanding way to fabricate transparent and flexible conducting elements for the evolution and expansion of next-generation smart devices. A thin, flexible material that is transparent and can conduct an electrical current has been developed by scientists in Taiwan. Materials with flexibility, transparency and conductivity are useful for wearable optoelectronics, but are rare. Indium tin oxide is a widely used transparent conductor, but when used in thin layers its electrical properties are not good enough for efficient thin-film transistors, thin-film solar cells and high-frequency devices. Ying-Hao Chu and co-workers from the National Chiao Tung University in Hsinchu used a technique called radio-frequency magnetron sputtering to deposit a thin film of indium oxide containing small amounts of molybdenum on a substrate. Measurements showed that the films had high charge mobility at room temperature, low electrical resistivity, and optical transparency, indicating that the material could be used in the next generation of smart devices. We present a promising technique for transparent flexible conducting oxide heteroepitaxial films: the direct fabrication of epitaxial molybdenum-doped indium oxide (IMO) thin films on a transparent flexible muscovite substrate. An n-type epitaxial IMO film is demonstrated with a mobility of 109 cm2 V−1 s−1, a figure of merit of 0.0976 Ω−1, a resistivity of 4.5 × 10−5 Ω cm and an average optical transmittance of 81.8% in the visible regime. IMO heterostructure not only exhibits excellent performance but also shows excellent mechanical durability. This study demonstrated an extraordinary achievement for the evolution and expansion of next-generation smart devices.


Introduction
Transparent conductive oxides (TCOs) exhibit impressive properties, including excellent electrical conductivity and high optical transmittance in the visible light range 1,2 . They have attracted great interest due to their potentially disruptive application in optoelectronics, including flat panel displays, light-emitting diodes, thinfilm transistors, solar cells and a variety of other applications that use both their electronic and transmittance features [3][4][5][6] . With their dramatically advanced properties for the integration of additional functionalities, flexible electronics-and numerous-related applications have become important research directions for soft technologies and wearable electronics 7,8 . In this research field, transparent flexible conducting components with excellent performance play a key role in almost all optoelectronics. Thus, the development of high-quality transparent flexible conducting elements has attracted a substantial amount of attention and become a promising research direction for next-generation consumer electronics.
Among the well-known TCOs, indium tin oxide (ITO) is the most extensively studied and widely used in lightemitting diodes, liquid crystal displays, and photovoltaic applications, for example 9,10 . ITO exhibits the desirable combination of a high electrical conductivity and high optical transparency. High-quality epitaxial ITO thin films possess a low resistivity of 2.3 × 10 −4 Ω cm with a carrier concentration of 4.8 × 10 20 cm −3 and mobility of 56 cm 2 V −1 s −111 . However, the characteristics of ITO films are still not acceptable enough to fulfill the requirements for advanced optoelectronics, especially thin-film transistors, thin-film solar cells and high-frequency devices, due to their requirements of a super high charge mobility for improving their performance. For instance, in the study of optically transparent patch antennas, a large enhancement in the microstrip efficiency with resonant frequencies between 100 MHz and 10 GHz was obtained with an increase in the electron mobility from 50 cm 2 V −1 s −1 to 100 cm 2 V −1 s −112 . Moreover, even with a high crystalline quality, the mobility of ITO is reduced for films fabricated on flexible substrates 13,14 . Therefore, it is productive to search for new candidates that break this barrier.
The large room-temperature mobility of molybdenumdoped indium oxide (IMO) films coupled with a low resistivity, high carrier concentration and optical transparency makes them a candidate for transparent flexible conducting elements to replace ITO films 15,16 . An IMO film was reported to show a resistivity of 1.7 × 10 −4 Ω cm, a mobility over 100 cm 2 V −1 s −1 and an average optical transmittance over 80% in the visible range 17 . Because molybdenum ions have a low affinity for oxygen ions and do not incorporate excess oxygen ions in their lattice, doped molybdenum can reduce the scattering centers, resulting in high-mobility IMO. In addition, when the valence difference between the dopant and substituted ion is 3, TCO can have a high carrier concentration, as shown by these defect chemical equations: Both a high-mobility and high carrier concentration make IMO an extraordinary TCO with excellent performance. Therefore, it is worth considering the challenges and fabricating transparent flexible conducting IMO thin films for advanced transparent flexible applications.
To produce a high-mobility conducting element featuring optical transparency as well as mechanical flexibility, two necessary components must be combined.
(1) There must be a high-quality epitaxial IMO thin film with low resistivity, high-mobility and high carrier concentration. (2) There must be an ideal substrate with a crystalline structure, high optical transparency and robustness against mechanical constraints. Recently, muscovite mica has been studied and identified as a favorable transparent flexible substrate for heteroepitaxial oxide films 18,19 . Numerous high-quality flexible oxide heterostructures were demonstrated that had excellent performance, providing new candidates for flexible applications. In this work, we demonstrate the fabrication of a high-mobility transparent flexible conducting heterostructure composed of heteroepitaxial IMO/mica. In our study, the heteroepitaxial IMO/mica system not only exhibits high electron conductivity and high optical transparency but also shows excellent stability against mechanical bending. Our results achieve an important step in the evolution of progressive optoelectronics.

Materials and methods
Growth of the IMO/AZO/mica Heterostructure The epitaxial IMO/AZO/mica heterostructure was deposited via RF magnetron sputtering using commercial IMO (96% In 2 O 3 and 4% MoO 3 ) and AZO (98% ZnO and 2% Al 2 O 3 ) targets with a diameter of 3 inches. An exfoliated artificial muscovite mica (1 × 1 cm) without treatment was used in this study. The vacuum chamber was pumped to a base pressure of 10 −6 Torr prior to deposition. During growth of the IMO and AZO, the chamber was maintained at 3 mTorr with an Ar flow of 20 sccm and an oxygen flow of 0.1 sccm. The deposition processes for both IMO and AZO were conducted while the substrate was heated to 550°C, and the RF power was 100 W. After deposition, the heterostructures were annealed in situ at 750°C for 1 h in the same atmosphere as during growth to improve the electrical conductivity.
Structural analysis X-ray diffraction (XRD) θ−2θ scans along the normal direction and Φ scans were performed to obtain structural information in a Bruker D8 high-resolution X-ray diffractometer using a monochromatic Cu K α1 radiation source. Cross-sectional TEM was used to study the microstructure of the heterostructure. The TEM specimens were prepared by a focused ion beam system along the normal direction. X-ray photoelectron spectroscopy spectra were obtained in a ULVAC-PHI PHI 5000 Versaprobe II.

Conductivity and transparency properties
The resistivity and Hall measurements were conducted in a physical property measurement system. The optical transmittance spectra were gathered in a PerkinElmer Lambda-900 spectrometer.

Bending tests
A homebuilt computer-controlled bending system was used to conduct all bending tests. While one end was fixed, the other end could be moved by a stepping motor. The heterostructure was bent by the bending system and observed by an additional microscope.

Results and discussion
In this study, epitaxial IMO thin films were grown on muscovite by RF sputtering. The crystal structure and phase identification of the heteroepitaxial films were then characterized by XRD. Moreover, as shown in Fig. 1a, an Al-doped ZnO (AZO) layer was inserted between the IMO and mica as a seed layer to enable the epitaxial growth of a high-quality IMO film. It is difficult to deposit an epitaxial IMO thin film on bare muscovite without a seed layer since the surface energies of IMO are similar for different orientations. Figure 1b shows an out-of-plane θ−2θ scan of the IMO/AZO/mica heteroepitaxial system. Only IMO (lll), AZO (00l) and muscovite (00l) diffraction peaks can be observed, indicating epitaxy of the heterostructure without undesirable secondary phases. The outof-plane d-spacing of the IMO (222) planes was calculated from the XRD peak position and was 2.912 Å, suggesting a slight compressive strain of 0.3%. This measured strain was caused by the defects created during the deposition process and not the interaction between the IMO and muscovite substrate attributed to the van der Waals epitaxy. Moreover, Φ scans of IMO, AZO and muscovite reflections were applied to identify the in-plane epitaxial relationships, as shown in Fig. 1c. A primary set and two secondary sets of muscovite peaks at 120°intervals with different intensities suggest that different stacking sequences existed in the muscovite. Cubic bixbyite ITO {400} and hexagonal AZO {101} exhibited six peaks at 60°i ntervals, which indicates the existence of multidomain IMO and single crystalline AZO films on the muscovite substrate. Based on the XRD results, the epitaxial relationship was determined to be (111) IMO //(001) AZO // (001) mica and (01-1) IMO //(010) AZO //(010) mica for the IMO/AZO/mica heterostructure. To gather critical information about the crystallinity, rocking curve measurements were carried out, and the full width at half maximum values of~1.28°and~0.85°were determined for the IMO (222) and AZO (002) peaks, respectively, as shown in Fig. 1d. Furthermore, to study the microstructure of the IMO/AZO/mica heterostructure as well as explore the heteroepitaxy, interfaces between the thin films and substrate were investigated by transmission electron microscopy (TEM). Figure 1e exhibits highresolution cross-sectional TEM images captured along the [010] mica direction, showing clear and defect-free IMO/ AZO/mica interfaces. The reciprocal lattices of the IMO, AZO and muscovite, obtained from fast Fourier transforms, were clearly indexed to distinguish the structural information regarding the IMO/AZO/mica heteroepitaxy. The consistency of the epitaxial relationships between XRD and TEM results was carefully confirmed. The clear interfaces without observable defects indicate the presence of good heteroepitaxy. From the XRD and TEM results, high-quality IMO/AZO/mica heteroepitaxy was fabricated, which is key to attaining excellent properties. To verify the valence state of the In and Mo ions in the heterostructure and confirm the chemical composition, X-ray photoelectron spectroscopy (XPS) measurements were carried out. As shown in Supplementary Fig. 1, a typical XPS spectrum was obtained from the surface of the IMO/AZO/mica heterostructure. The high-resolution spectra of the In 3d and Mo 3d core levels indicate that the valence state of In was +3 and that for Mo was +6 in the heterostructure. By analyzing the XPS spectrum, the ratio of Mo dopant can be confirmed as~4 wt % in the IMO films. It has been reported that IMO thin films exhibit a relatively high electron mobility as well as low resistivity at this doping level 17,[20][21][22] .
After the production of van der Waals epitaxy in the IMO/AZO/mica heterostructure, it is crucial to build a comprehensive and quantitative knowledge of the characteristics of this heteroepitaxy for advanced soft electronics. Electrical contacts were fabricated in the van der Pauw configuration. Temperature-dependent Hall measurements were carried out to characterize the charge carrier type and carrier concentration of the IMO heteroepitaxial structure. The resistivity, mobility and carrier concentration as a function of temperature of the heterostructure are shown in Fig. 2a. The IMO/AZO/mica heterostructure exhibited excellent conductivity (ρ < 1.2 × 10 −4 Ω cm) with a high mobility (μ = 109 cm 2 V −1 s −1 ) at room temperature. The resistivity decreased with temperature across the whole temperature range, exhibiting nearly metallic behavior. In addition, this behavior can be simply controlled by annealing the samples in different oxygen atmospheres. The IMO heteroepitaxial structure, annealed in a high oxygen pressure, exhibited insulating behavior and its resistivity increased as the temperature increased, as shown in Supplementary Fig. S2. The Hall measurement of the heterostructure confirmed the n-type conducting nature. The carrier concentration, n, was 1.27 × 10 21 cm −3 and showed almost no temperature dependence. The Hall electron mobility reached a value of 109 cm 2 V −1 s −1 at 300 K and 140 cm 2 V −1 s −1 at 10 K. The thickness-dependent resistivity, mobility and carrier concentration of the IMO heteroepitaxial structure are shown in Fig. 2b. The conductivity of the IMO/AZO/mica structure obviously increased as the thickness increased, which was related to an increase in the electron mobility for thicknesses from 100 to 400 nm. The resistivity was 2.62 × 10 −4 Ω cm and decreased to 1.17 × 10 −4 Ω cm as the thickness of IMO increased from 100 to 400 nm. The electron mobility and carrier concentration were 56.5 cm 2 V −1 s −1 and 4.90 × 10 20 cm −3 for a 100 nm thickness and 88 cm 2 V −1 s −1 and 6.06 × 10 20 cm −3 for a 400 nm thickness, respectively. It was obvious that as the thickness of the IMO film increased, an increase in the crystallinity was obtained, which reduced the scattering of the carrier and showed increased transport behavior. A comparison was made between the IMO/AZO/mica heteroepitaxial structure in this study and the currently reported IMO thin films in Table 1. It is clear from the table that the heterostructure in this work showed a very low resistivity among all the IMO thin films due to both its relatively high-mobility and carrier concentration, which were related to the high crystallinity. Based on our results, IMO/AZO/mica heteroepitaxial structures have excellent performance and can play a potential key role in advanced electronics applications.
Since IMO has exceptional transport behavior and high optical transparency, it is critical to characterize the optical transparency of the IMO/AZO/mica structure and discuss its conductivity. A photograph of a 2-inch wafer with a 400 nm IMO on AZO/mica, as shown in the inset of Fig. 3a, directly illustrates the high optical transparency of the heteroepitaxial structure. The optical transmittance spectra of IMO heterostructures with various thicknesses are shown in Fig. 3a. The pure muscovite substrate had a high transmittance without absorption over the whole measured range. Above the muscovite substrate, the IMO/AZO/mica heterostructure showed an average visible transmittance of 81-83% with a thickness from 100 to 400 nm. The observed oscillation in the spectra was caused by interference effects. The transmittance of the heteroepitaxial structure in the visible range slightly decreased with increasing thickness. A figure of merit (Φ TC ), which is an important parameter, was used to identify and compare the performance of all transparent conducting elements. The value of Φ TC was defined to be T 10 /R s (Ω −1 ), where R s is the sheet resistance and T is the average optical transmittance. Typically, a higher Φ TC indicates an increased performance for transparent conducting films. The Φ TC for the IMO/AZO/mica structure with various thicknesses, IMO films 17,20,21,23-34 and numerous TCO films on flexible substrates 14,[35][36][37][38][39][40][41][42][43][44] are shown in Fig. 3b. The ultrahigh figure of merit values for the IMO/AZO/mica heteroepitaxial structure can be observed due to its excellent characteristics. A very high Φ TC value of 0.0976 Ω −1 was achieved for the best IMO/ AZO/mica heterostructure, which exhibited high mobility (109 cm 2 V −1 s −1 ), low sheet resistance (1.37 Ω ▫ −1 ) and high optical transmittance (81.8%). As a result, the transparent conducting IMO/AZO/mica heterostructure exhibited remarkable performance beyond that for all reported flexible TCO films.
Furthermore, to extend the practical application of this IMO/AZO/mica heterostructure with mechanical flexibility, a range of cycling tests were carried out. The operational stability and mechanical durability of the flexible IMO/AZO/mica heteroepitaxial structures were tested under both compressive (flex-in) and tensile (flexout) bending states, as shown in the insets in Fig. 4a, b. Figure 4a, b shows that the resistance of the 300-nm thick heterostructure remained nearly constant with a bending radius down to 3.5 mm under both compressive and tensile deformations, respectively. Heterostructures with thicknesses of 100 and 200 nm also exhibited a steady conductivity with bending radii down to 5 mm. Thus, IMO heteroepitaxial structures were robust against mechanical constraints for flexible device applications. Due to only weak van der Waals interactions between the IMO/AZO and muscovite and small film-to-substrate  thickness ratio, it was clear that the strain applied by bending was not enough to affect the behaviors of the IMO. In addition, the fluctuation of the resistance for the IMO heterostructures with the same thickness in Fig. 4 was caused by the difference in the contact resistance and displacement of the contact, since the contact was remade when the sample switched between different bending states. The stability of the heterostructure was investigated by measuring the conducting behavior as a function of bending cycles and time in flex-in mode. The resistance of the IMO/AZO/mica structure was steady and increased 7% after 10,000 bending cycles under a bending radius of 3.5 mm, as shown in Fig. 4c. The thickness-dependent conduction stability of the IMO heterostructures is shown in the inset of Fig. 4c, which showed little change in the resistance of the heteroepitaxial structure, especially at a thickness of 300 nm after 10,000 bending cycles under a bending radius of 3.5 mm. Figure 4d shows that the conduction behavior of the IMO heterostructure was stable under compressive bending for a long duration under room-temperature environmental conditions. Moreover, the durability of the IMO/AZO/mica structure was studied under both mechanical bending and electrical bias. To simulate a practical condition, we applied a steady voltage of 5 V to the heterostructure with a bending radius of 3.5 mm, and an additional resistor was added into the circuit to control the current (0.2 A) flow through the IMO/AZO/mica heterostructure. A negligible change (<3%) was observed after one month. This indicates long-term conducting stability of the heterostructure under mechanical bending and electrical bias. Table 2 exhibits the conducting features of the flexible IMO/AZO/mica heteroepitaxial structure and recently reported TCOs on flexible substrates. Obviously, the heterostructure in this work had the best conducting performance among all flexible TCOs with a mobility over 100 cm 2 V −1 s −1 . Therefore, it is evident that the transparent flexible conducting IMO/AZO/mica heterostructure exhibited a performance superior to that of other flexible TCOs and an excellent stability, which are highly beneficial for advanced flexible optoelectronic applications.

Conclusion
In conclusion, a high-quality transparent flexible conducting IMO/AZO/mica heteroepitaxial structure was successfully fabricated with a mobility larger than 100 cm 2 V −1 s −1 , very high figure of merit (0.0976 Ω −1 ), relatively low resistivity (~4.5 × 10 −5 Ω cm), high carrier concentration (1.3 × 10 21 cm −3 ), and high optical transmittance (81.8%) at room temperature. In addition, this heterostructure retained the excellent performance with good flexibility. This study demonstrated an extraordinary achievement and can accelerate next-generation transparent flexible optoelectronics.