# Fileset

[NIMS_MDR.pdf](https://mdr.nims.go.jp/filesets/ad8fa049-392e-4689-8f0e-f588d2492ca0/download)

## Creator

[Ibrahima Gueye](https://orcid.org/0000-0001-5296-3894), [Shigenori Ueda](https://orcid.org/0000-0001-9425-0614), Atsushi Ogura, [Takahiro Nagata](https://orcid.org/0000-0002-8591-2943)

## Rights

This document is the Accepted Manuscript version of a Published Work that appeared in final form in ACS Applied Electronic Materials, copyright © 2024 American Chemical Society after peer review and technical editing by the publisher. To access the final edited and published work see https://doi.org/10.1021/acsaelm.4c00049[In Copyright](http://rightsstatements.org/vocab/InC/1.0/)

## Other metadata

[Direct Analysis of Stacked Au/Ti/In<sub>2</sub>O<sub>3</sub>/Al<sub>2</sub>O<sub>3</sub>/p<sup>+</sup>-Si Transport Mechanisms Using Operando Hard X-ray Photoelectron Spectroscopy](https://mdr.nims.go.jp/datasets/db391d8f-2d6d-483e-87df-1c7c8a49d800)

## Fulltext

Direct Analysis of Stacked Au/Ti/In2O3/Al2O3/p+-Si Transport Mechanisms Using operando HardX-ray Photoelectron SpectroscopyIbrahima Gueye*,1, 2 Shigenori Ueda,1, 3, 4 Atsushi Ogura,5, 6 and Takahiro Nagata*1, 61Research Center for Functional Materials, National Institute for Materials Science (NIMS), 1-1 Namiki, Tsukuba, Ibaraki 305-0044, Japan2Center for Synchrotron Radiation Research, Japan Synchrotron RadiationResearch Institute (JASRI), 1-1-1 Kouto, Sayo, Hyogo 679-5198, Japan.3Research Center for Advanced Measurement and Characterization,NIMS,1-2-1 Sengen, Tsukuba, Ibaraki 305-0047, Japan4Synchrotron X-ray Station at SPring-8, NIMS, 1-1-1 Sayo, Hyogo 379-5148, Japan5Meiji Renewable Energy Laboratory, Meiji University,1-1-1 Higashimita, Tama-ku, Kawasaki, Kanagawa 214-8571, Japan.6Graduate School of Science and Technology, Meiji University,1-1-1 Higashimita, Tama-ku, Kawasaki, Kanagawa 214-8571, Japan.Corresponding Authors: ibrahima.gueye@spring8.or.jp; NAGATA.Takahiro@nims.go.jpABSTRACT: Oxygen transport mechanisms for two different Au/Ti/In2O3/Al2O3/p+-Si samples were ex-perimentally evaluated with hard X-ray photoelectron spectroscopy (HAXPES). The deposition temperature foratomic layer deposition (ALD) of In2O3, as well as the bias voltages applied on the entire stacked structures,were the main parameters used in the work. Chemical analyses of the In2O3 layers deposited at 150 ◦C and200 ◦C for the samples named T_150 and T_200, respectively, revealed a decreased carbon impurity content inthe host In2O3 used as a dopant. Ex-situ interfacial analysis of In2O3/Al2O3 also indicated oxygen transportfrom Al2O3 to In2O3. Moreover, we observed that the Ti adhesion metal attracted oxygen and carbon fromthe In2O3 to form TiO2 and TiC conductive interlayers. Furthermore, operando-HAXPES under an appliedbias voltage also revealed that In2O3 underwent phase separation, likely due to variations in the space charge(carriers) around the In2O3/Al2O3 interface for the T_150 sample. Finally, our results emphasize the prominentroles of migration for the ionic oxygen/carbon species and the uncompensated interfacial charge formed by thebias voltage for the metal-semiconductor-oxide stacked structure.Keywords: Operando, HAXPES, Buried-interfaces, In2O3, Al2O3, Band bending, Chemical-structuresI. INTRODUCTIONThin film transistors (TFTs) have undergone extensive evo-lution, development, and refinement1,2. Recently, amorphous oxide semiconductors (AOS) have been proposed as propi-tious alternatives to extend the limits of the amorphous sil-icon (a-Si) semiconductors predominantly used as channel layers in TFTs3. AOS materials involving cations with the (n − 1)d10ns0(n ≥ 4) electronic configuration provide perfor-mance enhancements thanks to their high conductivities, high transparencies, and excellent mechanical properties4. Indeed, charge carriers for AOS migrate through the TFTs at rates 20 to 40 times faster than those for a-Si TFTs. In addition, the ad-vantages of the AOS materials over the current a-Si technol-ogy include superior surface planarity and easy manufactur-ing on plastic substrates at low servicing temperatures, which leads to flexible electronics without compromising their ex-cellent properties5.From the perspective of TFT applications, the AOS chan-nel materials require reasonably low carrier densities and high electron mobilities while also providing good ohmic con-tact to source/drain electrode materials. Among a plethora of amorphous oxide semiconductors (AOS), indium ox-ide (In2O3) emerges as a compelling and promising can-didate for transparent and flexible electronics. Its ap-peal stems from its wide band gap, high carrier mobil-ity (∼ 40 cm2/Vs), and compatibility with diverse appli-cations necessitating transparency and low-temperature deposition6–8. However, it is notable that consensus re-garding the precise value of the In2O3 band gap remains elusive9–12. It is imperative to acknowledge that band gap values can exhibit variability within a defined range ow-ing to factors such as experimental techniques, sample preparation methods, measurement conditions, or theo-retical calculations. In 2013, Irmscher et al.11 conducted optical absorption measurements on In2O3 single crystals with very low defect density, reporting an optically mea-sured gap of approximately 2.7 eV, consistent with the elec-tronic or fundamental gap. Conversely, Muydinov et al.12 posited that pure In2O3 cannot possess an indirect gap due to the parabolic nature of the conduction band. Addition-ally, they asserted that the minimum electronic or funda-mental band gap, symmetrically forbidden, hovers around 2.9 eV, with the first allowed transition originating from a level roughly 0.8 eV below the valence band top, yielding the commonly observed optical band gap value of approx-imately 3.7 eV. These recent findings align with previous calculations by Karazhanov et al.10.In2O3 unfortunately exhibits instability over time, pos-ing a significant hurdle to its future commercialization. This instability is largely attributed to the presence of an uncontrollable excess of oxygen vacancies (VO2+)13. In-deed, oxygen migration outwards from amorphous In2O3 at room temperature under air or vacuum conditions due to the relatively low dissociation energy of the In-O bond has been previously reported14,15. Fuh et al.14 observed a decrease in threshold voltage (Vth) values of amorphous IGZO-TFTs with prolonged exposure to the environment,2where the average threshold voltage shifted from 10.2 V to 5.8 V after a 9-day duration. They linked the insta-bility of passivation-free amorphous IGZO-TFT channels to the exposure of the a-IGZO layer to ambient condi-tions. The authors suggested that the weak bond of oxy-gen atoms (BDE: 346 kJ/mol16) lacked sufficient energy for structural relaxation during annealing, leading to easy desorption of oxygen atoms from the a-IGZO film and the formation of oxygen vacancies, thereby providing free car-riers. The precise mechanism (vacancy-mediated diffu-sion, interstitial diffusion, grain boundary diffusion, sur-face diffusion, etc.) by which oxygen is transported out of the amorphous In2O3 thin film remains unclear to date. In addition, the degree of similarity between the structure and properties of point defects in amorphous solids and those in the crystalline phase remains a subject of ongoing debate and exceeds the scope of this report17,18. However, the presence of excess VO2+ defects, each offering two elec-trons, inevitably leads to a transition from semiconduct-ing to metallic behavior19–21. Additionally, the VO2+ sites in the TFTs have been identified as the cause of negative Vth shifts in the transfer characteristics22,23. These sites are also suspected of influencing other TFT properties, in-cluding carrier concentration/mobility, current on/off ra-tio, subthreshold swing, and contact resistance. To mit-igate the issues associated with VO2+ defects, numerous research groups have pursued various optimization strate-gies.Simple chemical doping process represents one of the most common methods to regulate the excess carrier of In2O3 with precise control over the doping process. Adding elements such as Ti8,16, W8,16, La24, Ga25, and Si8 are also expected to stabilize the existing VO2+. In addition, the defect stabi-lization process will cancel the out-diffusion of the oxygen atoms from the In2O3 matrix by raising the In-O BDE.Furthermore, a doping process by controlling the deposi-tion temperature has been lately reported as a practical way of improving In2O3 electrical performance and stability. In this later method, the variation of the deposition temperature can affect the decomposition and reaction of precursors and lead to changes in the film’s properties, including its conduc-tivity, bandgap, or structural characteristics. More recently, it has been reported that low-temperature deposited In2O3 films grown by an atomic layer deposition (ALD) method showed the presence of impurities, which mainly consisted of carbon26. This residual carbon was expected to mitigate the VO2+ issues specially observed at 200 ◦C deposition tem-peratures. Indeed, from these papers26, authors claim that the involvement of these carbon impurities at 150 ◦C reduces the VO2+ and allows transistor to reach a threshold voltage (Vth) of 3.2 V and high mobility of 20.4 cm2/Vs. Finally, from the transfer and output characteristics of the In2O3-based TFT, au-thors also reported a large negative Vth shift with higher de-position temperature while no Vth shift was observed at lower deposition temperature.Post-thermal annealing (PTA) under an oxygen atmosphere has been also used to partially compensate for the oxygen losses from In2O313,27. Ma and coworkers also reported thepositive influence of the PTA process on In2O3-TFTs28, which was also suggested earlier by Yeom et al.29. Note that post-process annealing has been also shown to improve the thresh-old voltage shift30, field-effect mobility31,32 and channel-to-source/drain metallization interface resistivity33–35.The effects of these aforementioned In2O3 thermal treat-ments on the chemical and electronic structures of the adjacent layers and buried interfaces must also be considered seriously during optimization of the indium oxide-based TFTs36–38. Generally, the subthreshold swings (SS: the minimum gate voltage required to switch the transistor from OFF to ON) or Vth of AOS-TFT devices are mainly determined by the qual-ity of the AOS-gate dielectric interface. As we expected, the thermal conditions used for TFT manufacturing will affect the AOS-gate dielectric interface.To make excellent ohmic contact at the AOS-source (AOS-drain) heterojunction, it is important to select metal electrodes with low specific resistivities. Indeed, high specfic contact resistance at the AOS-source (or drain) metallization interface is associated with decreases in the field effect mobility. The chemical states at the interfaces and the effects of common metal contacts on TFT stability should be also considered to improve the TFT performance.Therefore, it seems necessary to shed more light on the changes in the chemical structures occurring in the films and buried interfaces TFT devices with non-destructive analyses. Hard X-ray photoelectron spectroscopy (HAXPES)39–41 is a suitable method to study the chemical states of buried layers and interfaces in real device structures with a non-destructive way owing to a large probing depth of photoelectrons with several ten nanometers. Thus, with HAXPES analyses for In2O3-based TFTs, we will be able (1) to observe how oxy-gen species move and react inside stacked devices, (2) to see whether C doping has the expected effect on In2O3 such as a reduction in the VO2+ level as well as suppression of vacancy migration, and (3) to confirm whether there are further issues that must be addressed to improve the general performance of the device.Since the purpose of this work is limited to understand-ing the reactions (mainly oxygen) that occur as a result of the deposition as well as PTA temperatures and the applied external electric field, we only focus on the prototypical Au/Ti/In2O3/Al2O3/p+-Si stacked structure and not on a full TFT device.II. EXPERIMENTAL SECTIONA. Materials and PreparationAmorphous In2O3 films were grown on Al2O3/p+-Si sub-strates by using an AT-400 ALD System from Anric Tech-nologies. For low-temperature In2O3 deposition using met-alorganic precursors, a liquid ethylcyclopentadienyl indium (InEtCp) compound was used42. In reference to the In2O3 deposition temperature, the In2O3/Al2O3/p+-Si samples with In2O3 deposition at 150 ◦C and 200 ◦C were labelled T_150-bare and T_200-bare, respectively, as shown in Figure 1(a).3The fully stacked Au/Ti/In2O3/Al2O3/p+-Si samples with Au/Ti top electrodes were named T_150 and T_200, respec-tively, as depicted in Figure 1(b).For each sample, a 5-nm-thick In2O3 layer on a 5-nm-thick Al2O3 layer was deposited on a heavily doped p+-Si substrate (bottom electrode), as follows. First, the p+-Si substrate was cleaned with an organic solvent solution and deionized wa-ter, followed by cleaning with HF solution to remove a native SiO2 layer. Then, a trimethylaluminium ((CH3)3Al) precur-sor and O3 oxidant gas were used to deposit the amorphous Al2O3layer via ALD at 200 ◦C. The presence of unintention-ally formed SiO2 intermediate layer between the p+-Si and Al2O3 layers was already identified and reported in our previ-ous work43.Then, In2O3 layers were grown on the Al2O3 layer by ALD. The ALD process consisted of alternating exposure to the In-EtCp precursor and the H2O/O3 oxidant gases, as follows13. The temperature of the InEtCp precursor was fixed at 80 ◦C, and N2 was used as both the carrier and purge gas. Then, to improve the properties of the indium oxide, PTA was per-formed on the T_150-bare and T_200-bare samples for 90 min at 150 ◦C under an O3 atmosphere. Finally, to evaluate the de-vice properties under the bias voltages, Au/Ti were patterned on the T_150-bare and T_200-bare layers by electron beam evaporation to obtain T_150 and T_200 (Figure 1). The sur-face size associated with the top electrode stacked device was 3 mm × 2 mm.B. Stacked device characterizationThe HAXPES measurements at room temperature and with a stable base pressure (∼ 10−7 Pa) were conducted at BL15XU at the SPring-8 synchrotron radiation facility in Japan44. Since the angle between the incoming X-ray and the photoelectron analyzer was set to 90o, the X-ray beam impinged on samples with a grazing angle of 5o in our ex-perimental setup (Figure 1). This experimental geometry with a take-off angle (TOA: the angle between the sample surface and photoelectrons entering the analyzer) of 85o was chosen to increase as much as possible the probing depth through the stacked devices. With this TOA of 85o, an X-ray beam with an expanded footprint size of approximately 400 × 30 µ m2was projected on the surfaces. The excitation photon energy was set to 5.95 keV, and photoelectrons were analyzed with a hemispherical electron analyzer (VG-Scienta R4000). The total energy resolution was 0.24 eV determined by fitting the Au Fermi edge. The binding energy (BE) was calibrated by using the Fermi edge of Au.The inelastic mean free paths (IMFPs) of photoelectrons were calculated with QUASES-TPP2M software, and probing depth d is ∼3 times the IMFPs45. With binding energy of 99.3eV for Si 2p3/2, 1561.8 eV for Al 1s and 445.4 eV for In 3d5/2at the excitation photon energy of 5.95 keV, we estimated d as approximately 30.4 nm, 24.6 nm, and 21.0 nm, respectively, for a TOA of 85o. These latter d values were multiplied by factors of 0.77, 0.5, and 0.17 when the TOAs were fixed at 50o, 30o and 10o, respectively, according to the equation d ∼3× IMFP× sin(TOA).FIG. 1: Schematic illustration of (a) T_150-bare andT_200-bare samples for HAXPES measurements and (b) T_150 and T_200 samples for operando-HAXPES.For the operando analysis, an ADCMT 6241A DC Volt-age current source/monitor system interfaced with an external PC controller was used to apply different bias voltages to the samples, as depicted in Figure 1(b) during the HAXPES mea-surements. The duration of each measurement with fixed bias voltage polarization of the sample was approximately 1 hour. Core level spectra were recorded once a steady current den-sity was reached under the fixed bias voltage (except with the relatively unstable current at +5 V), as reported in the elec-tronic supplementary Information (ESI), p. S-6, Figure S10 for T_150 and in ESI, p. S-6, Figure S11 for T_200.As a reference data, Au 4f core level spectra from T_150 and T_200 are reported in the ESI, p. S-6, Figure S12. Spec-tra were acquired using different bias voltages applied to the bottom p+-Si. As expected for the grounded Au layer, there is no significant binding energy difference (energy shift) when the applied bias voltage was modified. We also note that the Au 4f peak intensities and shapes remained identical during all operando analyses. The Au 4f7/2 binding energy of 84.0 eV was consistent with metallic gold (Au0)43. The overall data analyses of core level spectra were performed with CasaXPS software46 with a Shirley algorithm used to subtract the back-ground and a pseudo-Voigt peak shape (Gaussian/Lorentzian with 30 % Lorentzian character) for the major peak fittings.III. RESULTS AND DISCUSSIONA. Influence of the In2O3 deposition and PTA temperatures onthe oxygen transfer process in the In2O3/Al2O3/p+-Si stackAfter the survey scans reported on ESI, p. S-2, Figure S1,the In 3d, Al 1s, O 1s core levels and the valence band HAX-PES spectra of the T_150-bare and T_200-bare samples were recorded and shown in Figure 2 (a)-(d), respectively. The components and corresponding binding energies obtained by the peak fitting for each core level are listed in Table I. From Figure 2 (a), the In 3d deconvolution yielded three contri-butions attributable to different chemical states. Based on the report of Qi et al.47, we attributed that the green (444.84FIG. 2: Core level HAXPES spectra of T_150-bare andT_200-bare: (a) In 3d, (b) Al 1 s, and (c) O 1s binding energies and (d) the valence band spectra for T_150-bare and T_200-bare.eV) and blue (445.4 eV) peaks were related to stochiomet-ric In2O348–50 and carbon-doped In2O3 (C-In2O3−x), respec-tively. From a combination of surface-sensitive X-ray pho-toelectron spectroscopy (XPS, hν = 1.486 keV) experimental data and density functional theory (DFT) calculations, Qi et al. explained that the In 3d moved to a higher binding energy compared to that for In2O3.From the peak fitting result shown in Figure 2 (a), we see that the C-In2O3−x signal for T_150-bare is larger than for T_200-bare. We observed the presence of the C-In2O3−x com-pound not only in T_150-bare but also in T_200-bare, con-trary to what had been previously reported in some research26. We ascribed this contradiction to the fact that XPS used in26 reports is surface-sensitive, while in this current work, bulk-sensitive HAXPES was used. In other words, we may assume that in the case of the T_200-bare sample, we have a gradi-ent in the carbon content from the surface of the In2O3−x film to the buried interface (In2O3−x/Al2O3−x). We associated this relatively high concentration of the blue peak with incomplete pyrolysis and dihydroxylation when the ALD growth temper-ature decreased. Thus, the variations in the C-In2O3−x sig-nal indicated that the C doping level may be tuned during the ALD process. Furthermore, from AFM (ESI, p. S-2, Figure S2) a decrease in the root mean square (RMS) surface rough-ness value from 0.73 and 0.16 nm is observed for T_150-bare and T_200-bare, respectively.The Al 1s core level spectra in Figure 2 (b) indicate the presence of stoichiometric Al2O3 indicated by the grey peaks and oxygen-deficient Al2O3 (AlOx) indicated by the orange peaks at 1562.5 eV and 1561.8 eV, respectively51. This in-dexation of components is consistent with the depth probing from the naked Al2O3 (-bare) reported in ESI, p. S-3, FigureTABLE I: Summary of the fitting parameters used for the In3d5/2, Al 1s, O 1s and Ti 2p3/2 peaks: the colours,components and binding energies in eV are listed.Elements Colors Components Binding energy (eV)In 3d5/2 Red In0 (In-Ti) 443.3Green In2O3 444.8Blue C-In2O3−x 445.4Al 1s Orange AlOx 1561.8Gray Al2O3 1562.5O 1s Green In2O3 530.3 - 530.4Blue C-In2O3−x 531.1Gray Al2O3 532.0Orange AlOx 532.9Ti 2p3/2 Cyan Ti0 - TiC 453.9 - 454.2Magenta TiO 455.3Yellow Ti2O3 457.4Olive TiO2 459.1S.3. By decreasing the information from 85o (deep probing) to10o (shallow probing), we see the presence of contribution atlower binding energy from both Al 1s (ESI, p. S-3, Fig-ureS.3 (a)) and Al 2p (ESI, p. S-3, Figure S.3 (b)). We do notconsider surface potential formation at the naked surfacebecause the VBM (not shown here) of Al2O3 was found be-low Fermi energy (EF )52. From ESI, p. S-3, Figure S.3 (c), weassume a mixed-valence state of Al, such as the coexis-tenceof stoichiometric Al2O3 and non-stoichiometric Al2O3 (AlOx) due to a lack of oxygen species (presence of VO2+). Then, from Figure 2 (b), we also highlight that the relative amount of AlOx increased from the T_150-bare to the T_200-bare samples. As reported by Qi et al.47, it is challenging to distinguish the VO2+-poor In2O3, VO2+-rich In2O3−x how-ever, by increasing the temperature of In2O3 deposition, more VO2+ inside the In2O3 layer is likely generated. This result is also consistent with the fact a lack of oxygen in In2O3 will trigger an oxygen transport from Al2O3 to In2O3 and there-fore we observe more AlOx in T_200-bare than T_150-bare.The O 1s peaks in Figure 2 (c) were deconvoluted into four components representing four different states associated to the Al2O3 and In2O3 layers. The green peak at approximately 530.3 - 530.4 eV was attributed to the oxygen in stochiometric In2O353. Then the blue peak at 531.1 eV is associated with the C-In2O3−x53. The gray peak at 532.0 eV was assigned to theoxygen in stochiometric Al2O354 and the orange peak centeredat 532.9 eV was from the AlOx54 due to the presence of VO2+.Complementary depth profiling analyses along the T_200-bare stacked structure were carried out by changing TOA from 85o (deep probing) to 20o (shallow probing). While the In 3d energy position (not shown here) was not affected by TOA,5FIG. 3: Depth-dependent analyses: HAXPES spectra for the Al 1s state of Al2O3-bare at TOA of 85o and T_200-bare samples at TOAs of 20 and 85o.the Al 1s position seemed to differ slightly with a TOA of 20o as shown in Figure 3. In contrast to Al2O3-bare (Al2O3/p+-Si), the T_200-bare results indicated the presence of substan-tial AlOx contribution when using a similar TOA of 85o. The AlOxratio was more important around the In2O3/Al2O3 inter-face as indicated by the TOA of 20o. We attribute this oxy-gen gradient to a redistribution of oxygen between In2O3 and Al2O3 to mitigate the chemical potential mismatch as previ-ously observed in the case of for TiO2/VO2.56. Indeed, a dif-ference in oxygen chemical potential between two materials may give rise to oxygen ion transfer to bring the system into equilibrium. In this scenario, In2O3−x took oxygen from sto-chiometric Al2O3 to produce VO2+-like defects in the AlOx. Mobile oxygen species were transferred to adjacent materials until the chemical potentials of the layers matched.Furthermore, when comparing the AlOx amount between the two samples in Figure 2 (c), we observe that the ALD growth temperature enhanced the oxygen transport. The rel-ative quantification show an increase of AlOx amount from∼ 40% to ∼ 55% for T_150-bare and T_200-bare, respec-tively. This oxygen loss from Al2O3 layer by diffusion willreduced the VO2+ amount in the In2O3 layer. Bayer et al. ob-served similar phenomena during the deposition of Al2O3 onan ITO substrate57. They showed that the oxygen from theITO substrate was the source of the additional reactant duringthe deposition of the Al2O3 layer. They assumed that the largegrowth per cycle observed during ALD process of the Al2O3layer on the ITO substrate was only possible if an additionaloxygen reactant was present for the formation of Al2O3. Inthe Bayer report, oxygen species were released at the surfaceand reacted with the trimethylaluminium (TMA) precursor toform Al2O3. The authors also reported that a sufficient con-centration of interstitial oxygen was liberated from the ITOsubstrate to provide enough oxygen for the growth of a 1-2nm Al2O3 layer. Finally, we add that although the oxygenspecies were sufficiently mobile at 200 ◦C58, oxygen releasewas also observed at room temperature during the deposition of organic molecules onto clean ITO59. Based on the above discussion, we expected an identical process for our stacked In2O3/Al2O3/p+-Si.In Figure 4, we sketched models to explain the reduction process of the Al2O3 layer during the growth of In2O3. Fig-ure 4 (a) describes the first stage of the interaction between the Al2O3-bare and the InEtCp precursor. After completion of the first cycle of In2O3 growth, as shown in Figure 4 (b), VO2+ was present at the surface and in the bulk In2O3 layer. Now, because the In atoms from the InEtCp precursor specifically bond to oxygen at the surface upon exposure to the InEtCp precursor, a decrease in the In2O3 growth rate was observed. To partially compensate for this lack of oxygen, the diffusion of oxygen atoms from the Al2O3 substrate takes place, as il-lustrated by the red arrows in Figure 4 (b). This oxygen migra-tion process led to the creation of new VO2+ inside the Al2O3, as shown in Figure 4 (c).From 150 ◦C to 200 ◦C, more significant oxygen transport occurred in Figure 4 (e) than in Figure 4 (d). At 200 ◦C, more oxygen atoms are extracted from the Al2O3 layer. As de-scribed theoretically with the DFT calculations60, the oxygen atoms easily moved inside the In2O3 reached the surface and reacted with the InEtCp precursor to form a new In2O3 layer. This explanation is consistent with the higher AlOx amount for T_200-bare compared to that for T_150-bare, as shown in Figure 2(b).Finally, the effect of PTA on the T_150-bare and T_200-bare samples was also studied. Results from In 3d and Al 1s core levels before and after the PTA process are reported in the ESI, p. S-3, Figure S4. For both samples, no significant changes were observed for the normalized In 3d level after the PTA step. However, one noticed a slight re-oxidation of the covered Al2O3 layer after PTA. The stochiometric Al2O3 signal increased while that for the AlOx material decreased. From the decrease of AlOx, we argue that during the PTA process, oxygen species went through the In2O3 layer and par-tially compensated for the lack of oxygen in the Al2O3 layer.As a summary of the ALD growth temperature and PTA effects, it is clear that oxygen transport process into the In2O3/Al2O3/p+-Si stack may have sharply affected the de-vice performance. We observe that the ALD growth temper-ature for In2O3 controlled both the C-doping level and the VO2+ contents of the In2O3 and Al2O3 layers. We confirmed that the PTA process lowered the VO2+ sites inside the stacked In2O3/Al2O3 structure.B. Impact of the patterned Au/Ti electrodes on theAu/Ti/In2O3/Al2O3/p+-Si stackThe impact of Ti metal from Au/Ti electrode on In2O3 with a PTA process was studied. Chemical analysis of Ti/In2O3 interfaces from the T_150 and T_200 samples are shown in Figure 5. From the In 3d/Ti 2p core level spectra, we dis-cern the oxidation states of titanium from Ti0 to Ti2+ (ma-genta), Ti3+ (yellow), and Ti4+ (olive)61,62. Additionally, a small red component at 443.3 eV, indicative of the reduc-6FIG. 4: Illustration of the oxygen migrationdriven by diffusion on In2O3/Al2O3 stackobtained by ALD. (a) Formation of the interfacewith the first layer of In2O3 on Al2O3-bare. (b)Oxygen migration from the Al2O3 substratetowards the interface with the fully developedinitial layer of In2O3. (c) Emergence of oxygenvacancies within the Al2O3 layer showing theintricate interplay between the two materials.(d-e) Comparison of oxygen vacancies ratiowithin the Al2O3 layer resulting from In2O3deposition at 150 ◦C and 200 ◦C respectively.tion of indium with the In0 component, was attributed to the metallic contribution of indium In0 from the binary Ti-In system63. The binary Ti-In system is characterized by the mixture of indium and titanium at the Ti/In2O3 in-terface. Notably, the binding energy of In0 (443.3 eV) ob-served in Figure 5 is lower than the expected binding en-ergy of pure indium (443.8 eV) from a reference source, as reported in the ESI, p. S-4, Figure S5 (c). This devia-tion towards the negative direction (decreased binding en-ergy) of In0 can be attributed to changes in the electronic structure, bonding interactions, or chemical environment surrounding the indium atoms within the Ti-In alloy. The binding energies and the corresponding components are re-ported in Table I. Regarding the In2O3 reduction process, a complementary depth profiling analysis was initially carried out on the T_200 reference sample (with as-deposited In2O3 and without PTA). First, results revealed that In0 species were mainly located at the Ti/In2O3 interface. As seen in the ESI, p. S-4, Figure S6 (a) and (b), we observe a reduction of In2O3 (Green arrow) relative to In0 (red arrow) when the TOA goes down (from deep or bulk to shallow or interfacial probing). This result is consistent with the reduction of stochiometric In2O3 and an increase of In0 around the Ti/In2O3 junction. Then, for the partial oxidation of Ti, the results also indicated that the Ti oxidation process was triggered by Ti deposition and did not require any PTA. Indeed, based on the reactivity of Ti and its strong affinity for oxygen, the Ti layer should take up oxygen from the In2O3 layer. In addition, oxygen transport towards the Ti layer may be facilitated by the relatively weak In-O bond in the case of In2O3 structure. In fact, with a un-stable Ti/In2O3 interface, oxygen atoms could preferentially migrate from In2O3 to form the more stable TiO2 interlayer. To determine which tie-line is theoretically favored at room temperature deposition (25 ◦C), the two possible tie-line reac-tions are written in the form of balanced chemical reactions:Ti(s) + 2In2O3(s) = 4In(s) + 3TiO2(s);∆Groxn = 3GoTiO2 − 2GIon2O3 ∼ −988.5; (T = 25oC)From the standard 25 ◦C thermodynamic values, the standardGibbs free energy (∆Groxn) was found by summing the forma-tion energies of the oxides multiplied by their stoichiometrymetric coefficients. Since ∆Groxn was negative, the forward re-action was favoured, and we conclude that the stable tie-linefor this system connected In metal to TiO2 and that the Ti in contact with In2O3 formed an unstable interface. Finally, we add that the TOA analysis of Al 1s from ESI, p. S-4, Figure S6 (c) does not support the existence of dropping electrostatic potential (gradient) across the Al2O3 layer after the deposi-tion and partial oxidation of Ti top layer on In2O3 layer. This explanation is consistent with the fact regardless of the angle of measurement the Al 1s remain unchanged (identical shape and position). No shift or change relative to a band bending.As for the PTA impact, we have observed an important oxi-dation of Ti to TiO2 with PTA as reported in Figure 5. For both T_150 and T_200 samples, the total amount of Ti0 was more than halved by the PTA, which led to the creation of a thick (≥ 2.5 nm) titanium oxide intermediate layer (as supported by the ESI, p. S-5, Figure S7). We assumed that the In2O3 acted as an oxygen tank during PTA and then released some of the stored oxygen into the Ti layer. We also expect that the significant amount of TiO2 will increase the VO2+ concentra-tion in the In2O3 layer. This gradual increase in VO2+ would then provoke a second oxygen migration (compensation pro-cess), which would take place between the In2O3 and Al2O3 layers. A shift towards a lower Al 1s binding energy with in-creasing titanium oxide concentration was confirmed (green curve in ESI, p. S-5, Figure S8). This energy shift was asso-ciated with an increase in the amount of AlOx remaining after oxygen species left the Al2O3 layer.Moreover, we underline that the binding energies of Ti metal in Figure 5 is slightly shifted towards the higher energy relative to the Ti0 position when In2O3 underwent a PTA pro-cess, as seen in Figure 5. The Ti0 peak at 453.9 eV exhibited a BE shift of 0.3 eV towards higher energy due to the presence of titanium carbide (454.2 eV64). The presence of carbon on the Ti layer also suggested residual carbon transport from the In2O3 layer to the Ti film during the PTA (a heat treatment performed below 150 ◦C). We note that, unlike the wide band gap for TiO2, which is not a good electrical conductor, TiC is a good electrical conductor65,66 and can likely play a central role in determining the channel-to-source (or drain) interface resistivity in case of actual TFT.7In light of these results, we can expect that the oxygen de-ficiency at the In2O3/TiO2 junction as well as the presence ofTiC will act as a highly doped n+ contact (n+ channel/source(drain) region in the case of an actual TFT device). The VO2+will serve as an electron source that forms a semiconductorand leads to a low-resistivity interface. With the existence ofthe upper n+ contact side, the region of In2O3 near the Al2O3layer can be considered the less conductive surface (dopedn−). Therefore, In2O3-TFT with a Ti electrode may have ann−n+/metal junction in the source (drain) region, leading tobetter ohmic behavior.FIG. 5: In 3d/Ti 2p core levels for T_200(Au/Ti/In2O3/Al2O3/p+-Si structure) with and without the PTA process under an O3 atmosphere for theIn2O3/Al2O3/p+-Si stack (T_200-bare).C. Chemical analysis of the Au/Ti/In2O3/Al2O3/p+-Sistructure under an applied bias voltageOperando-HAXPES analysis was carried out under an ap-plied bias voltage to investigate the electric field effects on T_150 and T_200 samples. Before the operando analysis, a standard current-voltage measurement was performed on samples as reported in the ESI, p. S-5, Figure S9.Figure 6 illustrates the binding energies and peak shapes of Al 1s and Si 2p for (Figure 6 (a-b)) and T_200 (Figure 6 (c-d)) stacked devices. Similar trends in energy shifts were observed for both samples. Regarding Si 2p (Figure 6 (b)-(d)), it is noted that, except for the -1 V bias voltage setting, the effective or actual applied bias voltage across the devices tends to be lower than the set values (1V, 5V). Specifically, the Si 2p shift is notably lower than the applied voltage when +1V or 5V are applied. Under normal circumstances, the Si2p positions are expected to shift concerning the Si 2p0V_forward reference by almost the exact applied voltageamount, particularly for heavily doped p+-Si substrates. Thisapproximation holds for a voltage of -1V. However, for +5 Vapplied on T_150 and T_200 samples, the Si 2p from p+-Sisubstrates experi-ences a shift ranging from +1.7 eV to +2.0eV, respectively. These findings suggest limited penetration of the electric field into the heavily doped p-type Si. Notably,the effec-tive bias voltage applied to heavily doped p-type Simaydeviate from the set bias voltage due to various factors, including contact potential, depletion region, series resis-tance, external circuit impedance, etc. Therefore, the ex-act cause of this lower effective bias voltage on the stacked devices cannot be definitively determined. However, issues related to the complex outer HAXPES circuit connection or current flow have been ruled out, as evidenced by the nearly -1 eV Si 2p shift observed at -1 V. Furthermore, de-spite thelower effective bias voltage compared to the ap-plied voltage,the impact of the bias voltage on Al2O3 and In2O3 functional layers, the main focus of this study, was successfully demonstrated. Lastly, no significant broaden-ing of the Si 2p peak was observed with applied voltage, with the full width at half maximum (FWHM) of the Si 2p3/2 remaining between 0.39 eV and 0.46 eV.Chemical alterations occurring on the Al2O3 layers in response to applied bias voltages on the p+-Si substrate are also elucidated in Figure 6 (a)-(c). It is crucial to emphasize that due to the time-consuming nature of the analysis, we only utilized a single take-off angle (TOA) to record chemical changes. Consequently, it becomes chal-lenging to discern binding energy shifts associated with variations in the built-in potential gradient and intensity changes based on depth analysis. The impact of the built-in potential on the spectrum with this single TOA can only be anticipated through the asymmetric broadening of core levels. In both samples, an asymmetrical energy shift along the energy axis relative to the 0 V curve, as well as band bending induced by the electrostatic poten-tial across the stacked devices, were observed. Notably, the presence of a global energy shift of the symmetric spectral shape along the energy axis led us to dismiss the possibility of potential chemical changes under an applied bias volt-age. Additionally, it is noteworthy that the Al 1s shifts are lower than the Si 2p core levels when the same bias volt-age is applied. This suggests that the electrostatic potential drop within the Al2O3 layer and the broadening of the Al 1s peak at +5 V are more likely induced by band bend-ing and/or an interfacial effect between Al2O3 and p+-Si(interface energy barrier).Figure 7 reports In 3d/Ti 2p core level spectra from T_150 and T_200 samples. For both samples, analyses were car-ried out with similar bias stresses (0V_forward, +1V, +5V, 0V_reverse, and -1V). As expected after PTA of the In2O3layer and the deposition of Ti, a significant TiO2 contribution, as well as a Ti0 shift due to the additional TiC signal, was observed. We also note that for both samples, variations in the applied bias voltages did not affect the chemistry of the Ti layer. This result was consistent with the fact that the Au/Ti was grounded.In Figure 7, we observe that the residual binding energy shift of the In 3d5/2 peak around 445 eV (red dot line) is asso-ciated with the In2O3 growth temperature but not to the bias voltages applied to the samples. This In 3d5/2 shift is related to the difference of In2O3 and C-In2O3−x amount between samples, as presented in Figure 2 (a). Furthermore, under ap-plied a progressive bias voltage, no significant modification (shape and intensity) was detected for the T_200 sample, in8FIG. 6: Core level spectra from Al 1s in (a)-(c) and Si 2p in (b)-(d) from T_150 and T_200 samples, respectively under anapplied bias voltage (0 V_forward, 1 V, 5 V, 0 V_reverse and -1 V. We note that the Al 1s and Si 2p core level spectra fromT_150 and T_200 displayed the same trends for the binding energy (in red) shifts and full width at half maximum (fwhm inblue) under the applied voltage.FIG. 7: In 3d/Ti 2p core level as a function of the externallyapplied voltage bias (0V_forward, 1V, 5V, 0V_reverse, and-1V).contrast to the T_150 sample. As emphasized by the arrows,T_150 shows a voltage-dependent phase (denoted as a split-ting phase (SP)) as well as clear variations in the maximumintensities at +5 V, 0_rev, and -1V, and contrarily 0_for and+1V.Figure 8 provides additional evidence on the SP phase. Aswe have seen from Figure 2 (a)-(c), the two first components(green and blue) and the two last peaks (gray and orange)of O 1s core levels are consistent with the peak fitting of In3d (In2O3) and Al 1s (Al2O3), respectively. This relationshipleads to the fact that any chemical change observed from In 3dor Al 1s will likely induce a change in the O 1s curves. There-fore, similar to In 3d (Figure 7) and Al 1s (Figure 6), fromO 1s core levels from Figure 8 (a) and (c), we noted that theshapes of O 1s peaks are also bias voltage-dependent. Withthe T_200 sample in Figure 8 (c), the change of O 1s shapestarts from 532 eV. As highlighted by the black arrow, theFIG. 8: O 1s and C 1s core levels determined for T_150 (a-b)and T_200 (c-d) with operando-HAXPES analyses under anapplied bias voltage (0V_forward, 1V, 5V, 0V_reverse and-1V).presence of a significant tail out to 536 eV is observed in thecase of +5 V. This feature induced by the band bending underthe bias voltage was ascribed to Al2O3 as already discussed inthe experimental section (B). Besides the Al2O3 band bendingcomponent, with the T_150 sample, an extra intensity varia-tion of around 531 eV is also observed as indicated by the redarrow. This latter intensity variation is related to the indiumoxide layer. This modification of O 1s intensity was tightlycorrelated with the emergence of the SP phase and was morevisible at +5 V. We note that a certain threshold bias voltagewas needed to generate the splitting phase for both the In 3dand O 1s core levels from In2O3.9FIG. 9: Deconvolution of the In 3d and Ti 2p core levels as afunction of the bias voltage (0V_forward, 1 V, 5 V, 0V_reverse and -1 V) applied on the T_150 sample during theoperando-HAXPES analyses. For 5 V, 0 V_reverse, and -1 V,the In 3d level showed a supplementary component ascribedto splitting (SP for splitting phase) of C-In2O3−x (blue) underthe bias stress. The chemical state of the Ti 2p contribution(Ti0/TiC, TiO, Ti2O3, and TiO2) was unaffected over thevoltage range with a major contribution from the TiO2 olivepeak.From C 1s data in Figure 8 (b) and Figure 8 (d), we do notobserve a direct link between the carbon species and the SPphase observed for the T_150 sample under a bias voltage.However, the presence of TiC interlayer associated with theC 1s component at 282.1 eV was confirmed in both samples.This TiC is induced by the migration of carbon species intothe Ti layer.In Figure 9, In 3d/Ti 2p curve fitting of T_150 sample as thefunction of the applied bias voltage is reported. The SP signal(wine peak) began to distinctly appear with a +5 V bias volt-age and was induced by the splitting of C-In2O3−x componentshown in blue in Figure 9. Relative to the C-In2O3−x peak at445.4 eV, the binding energy shifts of the SP under the +5 V,0_rev, and -1 V bias voltages are +1.2 eV, +0.1 eV, and -1.0eV, respectively. We also note that the SP shift followed thesame trends as the Si 2p and Al 1s core levels as reported inFigure 6. Since the Au/Ti was grounded, we argue that chem-ical splitting took place at the In2O3/Al2O3 interface and wastriggered by interfacial band bending of the In2O3 layer. Inaddition, we did not observe an energy shift from the TiOxcontributions. For the T_200 sample, peak fitting data are re-ported in ESI, p. S-7, Figure S13.FIG. 10: Schematic cross-sectional view of theAu/Ti/In2O3/Al2O3/SiOx/p+-Si stacked structure deposited at: (a) T_150 and (b) T_200 samples. The arrows indicate the directions for oxygen and carbon movement.With a penetration of the electric field into the In2O3 film, the appearance of the SP for T_150 upon voltage application can have various origins such as the interaction of the electri-cal field with the chemical impurities, defects, charge accumu-lation/depletion, and even interfacial chemical reactions. The presence of this splitting phase also suggests that the T_150 sample (higher C-In2O3 contribution) is much more resistive than the T_200 sample (lower C-In2O3 contribution). This decrease of conductivity of T_150 is consistent with the fact the presence of doping carbon (C-In2O3) reduces the excess of oxygen vacancies (by raising the oxygen bond dissociation energy and reducing the excess of electrons from the remain-ing oxygen vacancies), and decreases the very high and un-controlled n-type conduction. As splitting occurred at the in-terface under a certain threshold voltage, we cannot exclude a self-forming interfacial dipole67–70 or interfacial electrochem-ical polarization mechanism in the present study. This in-terfacial state could be associated with the steep shift of the band structure across the interfaces. Indeed, with the In2O3 semiconductor-like T_150 sample, the charge carriers at the In2O3/Al2O3 interface are not suppressed by the injection or release of electrons from the grounded Ti/Au electrodes. Thus, the presence of an increased charge potential due to the interfacial charge might result in the polarization of the bot-tom side of the In2O3 layer and then lead to the formation of a separation phase for T_150.In Figure 10, we graphically summarized the chemical models for T_150 and T_200 stacked structures after the bias voltage was applied. In this model, we intentionally included10the native SiO2 interlayer as already discussed in our previ-ous paper43. For both samples, the presence of AlOx near the In2O3/Al2O3 interface resulted from oxygen diffusion into the indium oxide layer and compensated for the oxygen defi-ciency. Based on peak fitting of the In 3d core levels, the indium oxide layers from T_150 and T_200 are labelled as C-In2O3−x and In2O3, respectively. The presence of reactive Ti metal at the indium oxide surface led to a substantial oxy-gen shortage in the indium oxide layer. Ti removed oxygen and carbon species from the indium oxide film and then scav-enged them for the Ti layer. These transported species led to the creation of interfacial In0/In2O3 δ on the side of the in-dium oxide top interface and an −intermediate TiC/TiOx (TiO, Ti2O3, TiO2) on the Ti bottom edge. In response to the bias voltage applied at the p+-Si substrate, both samples exhibited band bending along the Al2O3 layer. Furthermore, no addi-tional chemical change occurred at the Ti (TiC and TiOx) ad-hesion layer under electrical polarization. Within this model, we highlight that in contrast to the metal-like T_200 sample, the incorporation of carbon into the In2O3 matrix increased the semiconductor-like behavior (T_150), which formed a splitting phase at the In2O3/Al2O3 interface under the applied electric field. This splitting phase undergoes a band bending which follows a similar direction to that of the Al2O3 layer.By qualitatively analyzing the HAXPES data, we observe that the ALD growth temperature and the post-annealing strongly affect the In2O3 film and its adjacent layers. We hypothesize that the imprint towards the negative bias as re-ported in ESI, p. S-5, Figure S9 can be linked to the interfacial defects (presence of energy barrier) while the hysteresis can be related to the back-and-forth movement of ionic species across layers.IV. CONCLUSIONSIn this paper, we performed a non-destructive chemical analysis of an Au/Ti/In2O3/Al2O3/p+-Si stacked structure. The study casts light on the impact of the ALD growth tem-perature and the post-annealing treatment as well. HAXPES analysis revealed that the chemical composition of the In2O3 semiconductor was temperature dependent, and the deposition of In2O3 on the Al2O3 insulator layer triggered oxygen trans-port from the insulator to the semiconductor to counterbalance the intrinsic oxygen deficiency of In2O3.With the Au/Ti patterned samples, we observed the migra-tion of the oxygen and carbon species from the In2O3 layer tothe Ti layer. This C and O absorption generated TiO2 and TiCinterlayers. Using operando-HAXPES analyses under appliedvoltages, Al2O3 band bending leading to charge accumulationat the In2O3/Al2O3 interface was observed. Conversely to theIn2O3 metal-like in T_200 sample, we observe a presence ofthe splitting phase in the In2O3 semiconductor-like behaviorin T_150 sample. We argue that this splitting phase in T_150sample was induced by the position-dependent charge carriersaround the In2O3/Al2O3 interface. From this paper, we notethat the ALD growth temperature as well as the post-annealingtreatment impacts may have major repercussions for interpret-ing and optimizing actual TFT performance as suggested bythe analysis of the Au/Ti/In2O3/Al2O3/p+-Si stacked devices.SUPPORTING INFORMATIONHAXPES spectra of wide range survey scans of StackedAu/Ti/In2O3/Al2O3/p+-Si; Surface morphologies of theT_150-bare and T_200-bare samples; HAXPES depth profileanalysis of naked-Al2O3; In 3d5/2 and Al 1s core levels fromT_150-bare and T_200-bare samples; Al 1s binding energyshifts for T_200 as a function of annealing under O3; HAX-PES depth profiling analyses of the Au/Ti/In2O3/Al2O3/p+-Siwithout PTA on In2O3/Al2O3/p+-Si; In 3d/Ti 2p core levelsfrom stacked T_150 and T_200 with and without PTA; Cur-rent density-voltage (J-V) characteristic; Investigation of thestability of the T_150 sample over time; Investigation of thestability of the T_200 sample over time; Au 4f, Al 1s and Si2p core level spectra for T_150 and T_200 samples under anapplied bias voltage; Deconvolution of the In 3d and Ti 2pcore level spectra as a function of the applied bias voltage.NOTESThe authors declare no competing financial interest.ACKNOWLEDGEMENTSThis work was partially supported by JSPS KAKENHI(Grant No: 20H02188). The HAXPES measurements wereperformed under the approval of the NIMS Synchrotron X-ray Station (Proposal No. 2020A4602 and 2020A4651).[1] Anthopoulos,T. D.; Noh, Y.-Y.; Jurchescu, O. D.Emerging Thin-Film Transistor Technologies and Ap-plications, Adv. Funct. Mater. 2020, 30, 2001678,https://doi.org/10.1002/adfm.202001678[2] Paterson, A. F. Anthopoulos, T. D. Enabling thin-film transistortechnologies and the device metrics that matter, Nat Commun.2018 9, 5264, https://doi.org/10.1038/s41467-018-07424-2[3] Sheng, J., Han, K.-L.; Hong, T.; Choi, W.-H.; Park, J.-S. Re-view of recent progresses on flexible oxide semiconductor thinfilm transistors based on atomic layer deposition processes,J. Semicond. 39, 011008 (2018), https://doi.org/10.1088/1674-4926/39/1/011008[4] Fortunato, E.; Barquinha, P.; Martins, R. Oxidesemiconductor thin-film transistors: a review of re-cent advances, Adv. Mater. 24, 2945-2986 (2012),11https://doi.org/10.1002/adma.201103228[5] Kamiya, T.; Hosono, H. Material characteristics andapplications of transparent amorphous oxide semi-conductors, NPG Asia Mater. 2010, 2(1) 15-22,https://doi.org/10.1038/asiamat.2010.5[6] Park, J. W.; Kang, B. H.; Kim, H. J. A Review of Low-Temperature Solution-Processed Metal Oxide Thin-Film Tran-sistors for Flexible Electronics, Adv. Funct. Mater. 2020, 30,1904632, https://doi.org/10.1002/adfm.201904632[7] Bierwagen, O. Indium oxide-a transparent, wide-band gapsemiconductor for (opto)electronic applications, Semicond.Sci. Technol. 30 (2015) 024001, https://doi.org/10.1088/0268-1242/30/2/024001[8] Aikawa, S.; Nabatame, T.; Tsukagoshi, K. Effects of dopantsin InOx-based amorphous oxide semiconductors for thin-filmtransistor applications, Appl. Phys. Lett. 103, 172105 (2013),https://doi.org/10.1063/1.4822175[9] Walsh, A.; Da Silva, J. L. F .; Wei, S.-H.; Körber, C.; Klein,A.; Piper, L. F. J.; DeMasi, A.; Smith, K. E. Panaccione, G.;Torelli, P.; Payne, D. J.; Bourlange, A.; Egdell, R. G. Natureof the Band Gap of In2O3 Revealed by First-Principles Cal-culations and X-Ray Spectroscopy, PRL 100, 167402 (2008),https://doi.org/10.1103/PhysRevLett.100.167402[10] Karazhanov, S. Zh.; Ravindran, P.; Vajeeston, P.; Ulyashin,A.; Finstad, T. G.; Fjellvåg, H. Phase stability, elec-tronic structure, and optical properties of indium ox-ide polytypes. Physical Review B, 76(7), 075129 (2007)https://doi:10.1103/physrevb.76.075129[11] Irmscher, K.; Naumann, M.; Pietsch, M.; Galazka, Z.;Uecker, R.; Schulz, T.; Schewski, R.; Albrecht, M.;Fornari, R. On the nature and temperature dependenceof the fundamental band gap of In <sub/>2</sub> O<sub/>3</sub>. physica status solidi (a), 211(1), 54-58 (2014).https://doi:10.1002/pssa.201330184[12] Muydinov, R.; Steigert, A.; Wollgarten, M.; Michaowski,P.; Bloeck, U.; Pflug, A.; Erfurt, D.; Klenk, R.;Körner, S.; Lauermann, I.; Szyszka, B. Crystallisation Phe-nomena of In2O3:H Films. Materials, 12(2), 266 (2019)https://doi:10.3390/ma12020266[13] Kim, H.Y.; Jung, E.A.; Mun, G.; Agbenyeke, R.E.; Park,J.S.; Son, S.U.; Jeon, D.J.; Park, S. H. K.; Chung, T.M.;Han, J.H. Low-temperature growth of indium oxide thinfilm by plasma enhanced atomic layer deposition usingliquid dimethyl(N-ethoxy-2,2-dimethylpropanamido)indiumfor high-mobility thin film transistor application, ACSAppl. Mater. Interfaces 8 (40) (2016) 26924-26931,https://doi.org/10.1021/acsami.6b07332[14] Fuh, C.-S.; Liu, P.-T.; Chou, Y.-T.; Teng L.-F.; Sze S. M. Roleof Oxygen in Amorphous In-Ga-Zn-O Thin Film Transistorfor Ambient Stability, ECS J. Solid State Sci. Technol. 2, Q1(2013). https://doi.org/10.1149/2.012301jss[15] Aikawa, S.; Nabatame, T.; Tsukagoshi, K. Si-incorporatedamorphous indium oxide thin-film transistors, Jpn. J. Appl.Phys. 58 090506 (2019). https://doi.org/10.7567/1347-4065/ab2b79[16] Kurishima, K.; Nabatame, T.; Kizu, T.; Mitoma, N.;Tsukagoshi, K.; Sawada, T.; Ohi, A.; Yamamoto, I.;Ohishi, T.; Chikyow, T.; Ogura, A. Prospectively ofcarbon-doped indium-tungsten-oxide channel TFT forbias stress instability. ECS Trans. 75, 149-156 (2016),https://doi.org/10.1149/07510.0149ecst[17] Strand, J.; Shluger, A. L. On the Structure of Oxygen Defi-cient Amorphous Oxide Films, Adv. Sci. 2024, 11, 2306243,https://doi.org/10.1002/advs.202306243[18] Buchholz, D. Bruce; Ma, Qing; Alducin, Diego; Ponce, Ar-turo; Jose-Yacaman, Miguel; Khanal, Rabi; Medvedeva, JuliaE.; Chang, Robert P. H. (2014). The Structure and Propertiesof Amorphous Indium Oxide. Chemistry of Materials, 26(18),5401-5411. doi:10.1021/cm502689x[19] Liu, L.; He, C. C.; Zeng, J.; Peng, Y. H.; Chen, W. Y.; Zhao,Y. J.; Yang, X. B. Theoretical Study of Oxygen-Vacancy Distri-bution in In2O 3, J. Phys. Chem. C 2021, 125, 13, 7077-7085,https://doi.org/10.1021/acs.jpcc.1c01462[20] Tiwari, N.; Nirmal, A.; Kulkarni, M.R.; John, R.A.;Mathews, N. Enabling high performance n-type metaloxide semiconductors at low temperatures for thin filmtransistors, Inorg. Chem. Front., 2020, 7, 1822-1844,https://doi.org/10.1039/d0qi00038h[21] Reunchan, P.; Zhou, X.; Limpijumnong, S.; Janotti, A.;Van de Walle, C. G. Vacancy defects in indium oxide: Anab-initio study, Curr. Appl Phys. 11, S296-S300 (2011),https://doi.org/10.1016/j.cap.2011.03.051[22] Kim, T.; Jang, B.; Lee, S.; Lee, W.; Jang, J. Improved NegativeBias Stress Stability of Sol-Gel-Processed Mg-Doped In2O3Thin Film Transistors, IEEE Electron Device Letters, 2018 39,12, 1872-1875, https://doi.org/10.1109/LED.2018.2873622.[23] Flewitt, A. J.; Powell, M. J. A thermalization energy analy-sis of the threshold voltage shift in amorphous indium galliumzinc oxide thin film transistors under simultaneous negativegate bias and illumination, J. Appl. Phys. 2014, 115, 134501,https://doi.org/10.1063/1.4870457[24] Wang, W.; He, G.; Yu, H.; Gao, Q.; Wang, L.; Xu, X.; Zhang,Y.; Wu, X.; He, B. Enhanced electrical performance and stabil-ity of La-doped indium oxide-based thin-film transistors andapplication explorations. Phys. Status Solidi A (2021), 219,2100590, https://doi.org/10.1002/pssa.202100590[25] Lee, S. H.; Kim, T.; Lee, J.; Avis, C.; Jang, J. Solution-processed gadolinium doped indium-oxide thin-film transistorswith oxide passivation, Appl. Phys. Lett. (2017), 110, 122102,https://doi.org/10.1063/1.4978932[26] Kobayashi, R.; Nabatame, T.; Onaya, T.; Ohi, A.; Ikeda,N.; Nagata, T.; Tsukagoshi, K.; Ogura, A. Comparison ofcharacteristics of thin-film transistor with In2O 3 and carbon-doped In2O 3 channels by atomic layer deposition and post-metallization annealing in O 3, Jpn. J. Appl. Phys. 60 030903(2021), https://doi.org/10.35848/1347-4065/abde54[27] Tang, T.; Dacha, P.; Haase, K.; Kreß, J.; Hänisch, C.;Perez, J.; Krupskaya, Y.; Tahn, A.; Pohl, D.; Schneider,S.; Talnack, F.; Hambsch, M.; Reineke, S.; Vaynzof, Y.;Mannsfeld, S. C. B. Analysis of the Annealing Budget ofMetal Oxide Thin-Film Transistors Prepared by an AqueousBlade-Coating Process, Adv. Funct. Mater.2023, 33, 2207966,https://doi.org/10.1002/adfm.202207966[28] Ma, Q.; Zheng, H. M.; Shao, Y.; Zhu, B.; Liu, W. J.; Ding,S. J.; Zhang, D. W. Atomic-Layer-Deposition of Indium OxideNano-films for Thin-Film Transistors, Nanoscale Res. Lett. 13,4 (2018), https://doi.org/10.1186/s11671-017-2414-0[29] Yeom, H. I.; Ko, J. B.; Mun, G.; Park, S. H. High mobilitypolycrystalline indium oxide thin-film transistors by means ofplasma-enhanced atomic layer deposition, J. Mater. Chem. C,2016, 4, 6873-6880, https://doi.org/10.1039/C6TC00580B[30] Nomura, K.; Kamiya, T.; Ohta, H.; Hirano, M.; Hosono, H.Defect passivation and homogenization of amorphous oxidethin-film transistor by wet O2 annealing. Appl. Phys. Lett. 93,192107 (2008), https://doi.org/10.1063/1.3020714[31] Lee, S.; Bierig, B.; Paine, D.C. Amorphous struc-ture and electrical performance of low-temperature an-nealed amorphous indium zinc oxide transparent thin12film transistors. Thin Solid Films 520, 3764 (2012),https://doi.org/10.1016/j.tsf.2011.06.082[32] Nomura, K.; Kamiya, T.; Hirano, M.; Hosono, H. Origins ofthreshold voltage shifts in room-temperature deposited and an-nealed α-InGaZnO thin-film transistors, Appl. Phys. Lett. 95,013502 (2009), https://doi.org/10.1063/1.3159831[33] Lim, W. T.; Norton, D. P.; Jang, J. H.; Craciun, V.;Pearton, S. J.; Ren, F. Carrier concentration dependence ofTi/Au specific contact resistance on n-type amorphous indiumzinc oxide thin films, Appl. Phys. Lett. 92, 122102 (2008),https://doi.org/10.1063/1.2902322[34] Shimura, Y.; Nomura, K.; Yanagi, H.; Kamiya, T.; Hi-rano, M.; Hosono, H. Specific contact resistances be-tween amorphous oxide semiconductor In-Ga-Zn-O andmetallic electrodes, Thin Solid Films 516, 5899 (2008),https://doi.org/10.1016/j.tsf.2007.10.051[35] Lee, S.; Park, H.; Paine, D.C. A study of the specific contactresistance and channel resistivity of amorphous IZO thin filmtransistors with IZO source-drain metallization, J. Appl. Phys.109, 063702 (2011), https://doi.org/10.1063/1.3549810[36] Kwon, J.-Y.; Lee, D.-J.; Kim, K.-B. Transparent amorphousoxide semiconductor thin film transistor, Electron. Mater. Lett.7, 1 (2011), https://doi.org/10.1007/s13391-011-0301-x[37] Nag, M.; Bhoolokam, A;. Steudel, S.; Genoe, J.; Groeseneken,G.; Heremans, P. Impact of the low temperature gate dielectricson device performance and bias-stress stabilities of a-IGZOthin-film transistors, ECS J. Solid State Sci. Technol. 4 (8) N99-N102 (2015), https://doi.org/10.1149/2.0121508jss[38] Yim, J.-R.; Jung, S.-Y.; Yeon, H.-W.; Kwon, J.-Y.; Lee,Y.-J.; Lee, J.-H.; Joo, Y.-C. Effects of Metal Electrodeon the Electrical Performance of Amorphous InGaZnO ThinFilm Transistor, Jpn. J. Appl. Phys. 51 011401 (2011),https://doi.org/10.1143/JJAP.51.011401[39] Gueye, I.; Le Rhun, G.; Renault, O.; Cooper, D.; Defay, E.;Barrett, N. Correlation of electrical characteristics with inter-face chemistry and structure in Pt/Ru/PbZr0.52Ti0.48O3/Pt ca-pacitors after post metallization annealing, Appl. Phys. Lett.113, 132901 (2018); https://doi.org/10.1063/1.5041767[40] Gueye, I.; Le Rhun, G.; Renault, O.; Cooper, D.; Ceolin,D.; Rueff, J.-P.; Barrett, N.; Operando hard X-ray photoelec-tron spectroscopy study of the Pt/Ru/PbZr0.52Ti0.48O3interface, Appl. Phys. Lett. 111, 032906 (2017);https://doi.org/10.1063/1.4993909[41] Woicik, J. C. Hard X-ray Photoelectron Spectroscopy (HAX-PES) (Springer Ser. Surface Sc, Springer, Berlin, 2016),https://doi.org/10.1007/978-3-319-24043-5[42] Mizutani, F.; Higashi, S.; Inoue, M.; Nabatame, T.Atomic layer deposition of stoichiometric In2O3 films us-ing liquid ethylcyclopentadienyl indium and combinationsof H2O and O2 plasma, AIP Advances 9, 045019 (2019),https://doi.org/10.1063/1.5081727[43] Gueye, I.; Kobayashia, R.; Ueda, S.; Nabatame, T.;Tsukagoshi, K.; Ogura, A.; Nagata, T. Operandohard X-ray photoelectron spectroscopy study of buriedinterface chemistry of Au/In1.16O0.04/Al2O3/p+-Sistacks, Applied Surface Science 593 (2022) 153272,https://doi.org/10.1016/j.apsusc.2022.153272[44] Ueda, S.; Katsuya, Y.; Tanaka, M.; Yoshikawa, H.; Yamashita,Y.; Ishimaru, S.; Matsushita, Y.;Kobayashi, K. Present Status ofthe NIMS Contract Beamline BL15XU at Spring-8, AIP Conf.Proc. 1234, 403-406 (2010), https://doi.org/10.1063/1.3463225[45] Tanuma, S.; Powell, C. J.; Penn, D. R. Calculations of electroninelastic mean free paths. V. Data for 14 organic compoundsover the 502000 eV range, Surf. Interface Anal. 1994, 21, 165-176, https://doi.org/10.1002/sia.740210302[46] Walton, J.; Wincott, P.; Fairley, N.; Carrick, A. Peak Fittingwith CasaXPS: A Casa Pocket Book. Knutsford, UK: AccolyteScience, 2010[47] Qi, Y.; Jiang, J.; Liang, X.; Ouyang, S.; Mi, W.; Ning, S. Zhao,L.; Ye, J. Fabrication of Black In2O3 with Dense Oxygen Va-cancy through Dual Functional Carbon Doping for EnhancingPhotothermal CO2 Hydrogenation, Adv. Funct. Mater.2021, 31,2100908, https://doi.org/10.1002/adfm.202100908[48] Kim, S.; Carpenter, P. D.; Jean, R. K.; Chen, H.; Zhou, C.;Ju, S.; Janes, D. B. Role of Self-Assembled Monolayer Pas-sivation in Electrical Transport Properties and Flicker Noiseof Nanowire Transistors, ACS Nano. 2012, 6, 8, 7352-7361,https://doi.org/10.1021/nn302484c[49] Henegar, A. J.; Gougousi, T. Native oxide transport andremoval during the atomic layer deposition of Ta2O5 onInAs(100) surfaces, J. Vac. Sci. Technol. A 2016, 34, 031101,https://doi.org/10.1116/1.4945115[50] Kyndiah,A.; Ablat, A.; Guyot-Reeb, S.; Schultz, T.; Zu, F.;Koch, N.; Amsalem, P.; Chiodini, S.; Alic, T. Y.; Topal,Y.; Kus, M.; Hirsch, L.; Fasquel, S.; Abbas, M. Mul-tifunctional Interlayer for Solution Processed High Perfor-mance Indium Oxide Transistors. Sci Rep. 2018, 8, 10946,https://doi.org/10.1038/s41598-018-29220-0[51] Mendes, M. K.; Martinez, E.; Marty, A.; Veillerot, M.;Yamashita, Y.; Gassilloud, R.; Bernard, M.; Renault, O.;Barrett, N. Forming mechanism of Te-based conductive-bridge memories. Appl. Surf. Sci. (2018) 432, 34-40,https://doi.org/10.1016/j.apsusc.2017.07.187[52] Ohashi, N.; Yoshikawa, H.; Yamashita, Y.; Ueda, S.; Li,J;. Okushi, H.; Kobayashi, K.; Haneda, H. Determinationof Schottky barrier profile at Pt/SrTiO3:Nb junction by x-ray photoemission, Appl. Phys. Lett. 101, 251911 (2012)https://doi.org/10.1063/1.4772628[53] Nayak, P. K.; Hedhili, M. N.; Cha, D.; Alshareef, H. N. Highperformance In2O3 thin film transistors using chemically de-rived aluminum oxide dielectric, Appl. Phys. Lett. 103, 033518(2013), https://doi.org/10.1063/1.4816060[54] Liu, Q.; Qin, H.; Boscoboinik, J. A.; Zhou, G. Com-parative Study of the Oxidation of NiAl(100) by Molecu-lar Oxygen and Water Vapor Using Ambient-Pressure XrayPhotoelectron Spectroscopy, Langmuir, 2016, 32, 11414,https://doi.org/10.1021/acs.langmuir.6b02752[55] Ma, H. P.; Yang, J. H.; Yang, J. G.; Zhu, L. Y.; Huang,W.; Yuan, G. J.; Feng, J. J.; Jen, T. C.; Lu, H. L. System-atic Study of the SiOx Film with Different Stoichiometry byPlasma-Enhanced Atomic Layer Deposition and Its Applica-tion in SiOx/SiO2 Super-Lattice, Nanomaterials, 2019, 9, 55,https://doi.org/10.3390/nano9010055[56] Park, Y.; Sim, H.; Jo, M.; Kim, G.-Y.; Yoon, D.; Han, H.;Kim, Y.; Song, K.; Lee, D.; Choi, S.-Y.; Son, J. Direc-tional ionic transport across the oxide interface enables low-temperature epitaxy of rutile TiO2. Nat Commun 11, 1401(2020). https://doi.org/10.1038/s41467-020-15142-x[57] Bayer, T. J. M.; Wachau, A.; Fuchs, A.; Deuermeier, J.; Klein,A. Atomic Layer Deposition of Al2O3 onto Sn-Doped In2O3:Absence of Self-Limited Adsorption during Initial Growth byOxygen Diffusion from the Substrate and Band Offset Modifi-cation by Fermi Level Pinning in Al2O3, Chem. Mater. 2012,24, 23, 4503-4510, https://doi.org/10.1021/cm301732t[58] Quaas, M.; Steffen, H.; Hippler, R.; Wulff, H. Investiga-tion of diffusion and crystallization processes in thin ITOfilms by temperature and time resolved grazing incidenceX-ray diffractometry, Surface Science 540 (2003) 337-342,13https://doi.org/10.1016/S0039-6028(03)00850-1[59] Lee, K. H.; Jang, H. W.; Kim, K.-B.; Tak, Y.-H.; Lee, J.-L.Mechanism for the increase of indium-tin-oxide work functionby O2 inductively coupled plasma treatment, J. Appl. Phys. 95,586 (2004); https://doi.org/10.1063/1.1633351[60] Ágoston,P.; Albe, K. Ab initio modeling of diffusionin indium oxide, Phys. Rev. B (2010) 81, 195205,http://dx.doi.org/10.1103/PhysRevB.81.195205[61] Shvab, R.; Hryha, E.; Nyborg, L. Surface chemistry ofthe titanium powder studied by XPS using internal stan-dard reference, Appl. Surf. Sci., 2014, 295, 180-188,https://doi.org/10.1080/00325899.2016.1271092[62] Biesinger, M. C.; Lau, L. W. M.; Gerson, A. R.; Smart,R. S. C. Resolving surface chemical states in XPS analy-sis of first row transition metals, oxides and hydroxides: Sc,Ti, V, Cu and Zn, Appl. Surface Sci., 2010 257, 887-898,https://doi.org/10.1016/j.apsusc.2010.07.086[63] Gulay, L. D.; Schuster, J. C. Investigation of the titanium-indium system. J. Alloy. Compd. 360, 137-142 (2003)https://doi.org/10.1016/S0925-8388(03)00319-0.[64] Park, J.; Back, T.; Mitchel, W. C.; Kim, S. S.; Elhamri,S.; Boeckl, J.; Fairchild, S. B.; Naik, R.; Voevodin, A. A.Approach to multifunctional device platform with epitaxialgraphene on transition metal oxide. Sci Rep. 2015, 5, 14374,https://doi.org/10.1038/srep14374[65] Yukun, P.; Hua, X.; Rasit, K. Investigation of Electrical Con-ductivity and Oxidation Behavior of TiC and TiN Based Cer-mets for SOFC Interconnect Application, ECS Transactions, 7(1) 2427-2435 (2007), https://doi.org/10.1149/1.2729365[66] Guo, Q.; Wu, T.; Liu, L.; Hou, H.; Chen, S.; Wang, L.Flexible and conductive titanium carbidecarbon nanofibers forthe simultaneous determination of ascorbic acid, dopamineand uric acid, J. Mater. Chem. B, 2018,6, 4610-4617,https://doi.org/10.1039/C8TB00938D[67] Schaefer, M.; Halpegamage, S.; Batzill, M.; Schlaf, R. In-vestigation of the dipole formation and growth behavior atIn2O3|TiO2 heterojunctions using photoemission spectroscopyand atomic force microscopy, J. Appl. Phys. 119, 065305(2016); https://doi.org/10.1063/1.4940954[68] Lang, O.; Pettenkofer, C.; Sanchez-Royo J. F.; Segura, A.;Kleina, A.; Jaegermann, W. Thin film growth and band lineupof In2O3 on the layered semiconductor InSe, J. Appl. Phys. 86,5687 (1999); https://doi.org/10.1063/1.371579[69] Chen, Q.; Wang, C.; Li, Y.; Chen, L.; Interfacial Dipole in Or-ganic and Perovskite Solar Cells, J. Am. Chem. Soc. 2020, 142,43, 18281-18292, https://doi.org/10.1021/jacs.0c07439[70] Chokawa, K.; Shiraishi, K. Self-forming and self-decomposing gallium oxide layers at the GaN/Al2O3interfaces, Applied Physics Express 12, 061008 (2019),https://doi.org/10.7567/1882-0786/ab21b4