Analysis of Ultrahigh Apparent Mobility in Oxide Field-Effect Transistors.

For newly developed semiconductors, obtaining high-performance transistors and identifying carrier mobility have been hot and important issues. Here, large-area fabrications and thorough analysis of InGaZnO transistors with enhanced current by simple encapsulations are reported. The enhancement in the drain current and on-off ratio is remarkable in the long-channel devices (e.g., 40 times in 200 µm long transistors) but becomes much less pronounced in short-channel devices (e.g., 2 times in 5 µm long transistors), which limits its application to the display industry. Combining gated four-probe measurements, scanning Kelvin-probe microscopy, secondary ion mass spectrometry, X-ray photoelectron spectroscopy, and device simulations, it is revealed that the enhanced apparent mobility up to several tens of times is attributed to the stabilized hydrogens in the middle area forming a degenerated channel area while that near the source-drain contacts are merely doped, which causes artifact in mobility extraction. The studies demonstrate the use of hydrogens to remarkably enhance performance of oxide transistors by inducing a new mode of device operation. Also, this study shows clearly that a thorough analysis is necessary to understand the origin of very high apparent mobilities in thin-film transistors or field-effect transistors with advanced semiconductors.

. Gate-voltage dependent of mobility in the saturated regime. a.
Differential saturated mobility as a function of gate voltage for IGZO-H TFTs. The gradual decreased saturated mobility from 454 cm 2 /(V·s) to 300 cm 2 /(V·s) could be due to surface or phonon scattering due to strong gate-field, as explained in the reports by Nathan et al. [1] b. Differential saturated as a function of gate voltage for IGZO TFTs. The drain voltage is 20 V.
To ensure the mobility values, we examined the extracted mobility by using them to simulate ID-VGS and ID-VD curves as shown below ( Figure S3 and S4). The

Part II. Reliability test.
Reliability tests were given in Figure S5. Summarizing Figure S5, IGZO-H films are more stable than IGZO film in stressing and show a long shelf life. The improved reliability of IGZO-H TFTs is attributed to the protection from ambience by encapsulation layer SiOX/SiNX and the lower ratio of VO which induces gradual trapping of mobile electrons [2] . Statistical data are shown in Figure S6.

Part III. Effects of deposition and annealing of the encapsulation layers
The fabrication of high performance TFTs critically relies on depositing and annealing SiNX/SiOX layers. The impacts of each layers were investigated as below and the conclusion would be given finally. Moreover, the influence of annealing temperature and time on device performance were studied and presented in this part.
(1) The effect of SiOX is as shown below. From (a) to (b), the results indicate that deposition of SiOX slightly enhanced current from 4x10 -5 A to 10 -4 A (saturated regime), but the mobility was lowered due to destruction of back     few devices with 300 nm SiNX maintain can be turned off in the scanning, but most of the devices are conductive as shown in (f). Therefore, it is confirmed the post-annealing rather than PECVD deposition is the most important part in the hydrogenation of IGZO, which leads to high mobility and good on-off ratio. Also, it is possible to control the on-off ratio and mobility by controlling the thickness of SiNX layer.  Summarizing the above figures, clearly single SiOX layer did not affect the current or mobility, whereas single SiNX enhanced both current and mobility but lead to poor sub-threshold properties with large SS values ( Figure S8). Upon deposition of SiNX/SiOX layers on IGZO, the devices did not show high mobility and on-off ratio until after post-annealing ( Figure S9). In the case that annealing after etching SiNX, the devices exhibit regular mobility ( Figure S10). These results suggest that post-annealing with SiNX layer is the critical step to obtain high mobility; in this process, probably the residual SiH moiety in the SiNX layer acts as a source of hydrogen atoms and the SiOX layer acts as a buffer-layer.
Finally, the thickness of SiNX film thickness should be carefully optimized (~100nm) and too thin or too thick SiNX films led to devices with regular mobility or very low on-off ratio, respectively ( Figure S11).  are not proportional to the spatial distance (Figure S15e-g). Hence, in IGZO-T TFTs, that the channel resistance is mostly decided by the resistance of short, less conducting regions near the contacts.  For device simulations, the sub-gap density is simulated using the following model: The density of states for the sub-gap states are shown in Figure S16 and the modelling results are plotted in Figure S17-18. The calculations were performed as described in [3] .

Part VI. DFT calculations of IGZO-H and IGZO.
The electronic density of states has been simulated based on anisimov-type rotational invariant GGA+U calculations [4] . By way of OPIUM code in generation with recent RRKJ optimization method [5] , we gave the norm-conserving pseudopotentials within the most popular KB (Kleinman-Bylander) projector framework [6] , and the non-linear partial core corrections [4a, 7] .
The above Eq and EH are the total energy of a relaxed defective lattice in charge state q and the energy of an ideal host lattice at the ground-state, respectively.
The ΔEF in Eq. (1) is the change of Fermi energy with respect to the VBM (setting EV=0), and nα is the number of atoms for constituent element α chosen as targeted defect sites, then 0 is the referenced chemical potential, based on the method used in the work of Zunger et al [8] . The thermal transition energy/level is the critical Fermi level position in the band gap where the charge state changes from q to q' as ∆EF changes in the band gap with the lowest-energy, which means the formation energy follows ∆H(q, EF)=∆H(q', EF) based on Eq. (1). The detailed forms of TTL have been discussed by Janotti et al [9] and Zunger et al [8,10] and it is basically calculated by DFT procedure by the following equation: Here ED(q) and ED(q') refer to the total energies of the defective structure at the charge state q and q' respectively.
The DFT calculations of IGZO and IGZO with interstitial H ions are given.   The temperature dependence of conductance = D D was measured at around room temperatures (270 K to 320 K) and at varied gate-field ( Figure   S24a). It reveals a thermally activated conduction mechanism, as it follows = 0 exp (− ) (k is Boltzmann constant and Ea is the activation energy measuring the energy difference between the fermi level EF and the conduction band edge CBM), somehow similar to poly-crystalline silicon [11] . The thermally activated conductivity is apparently related to the donor-like states induced by hydrogen. The values of Ea are plotted against corresponding VG ( Figure S24c) and yet the main difference between the two lies in the characteristic conductance G0, as shown in (Figure S24d). The data show that conductivity of IGZO-H TFTs is generally thermally activated. Ea or G0 as a function of gate-voltage.

Part VIII. Mobility comparison with former studies.
We summarize our results and compare them with representative results of high mobility TFTs based on IGZO or IGZO composites in Table S1. High mobility IGZO TFTs have been frequently obtained by using nano-material composites, including carbon nanotubes or metal nanowires [12] . Using IGZO itself to achieve high mobility has been accomplished by partially covering the capping layers in the channel, which requires large patterns and a long time (several days) before achieving stable operation [13] .. Our approach enables facile fabrication methods and is compatible with conventional fabrication processes.

Table S1
Reported values of high-mobility TFTs based on IGZO or IGZO composites, showing field-effect mobility in the linear regime ("lin") or saturated regime ("sat"). The data shown below are the highest numbers reported in the references and this study.