# Fileset

[ApplSurFScieNCKU_SI.pdf](https://mdr.nims.go.jp/filesets/a4eacd24-7e3f-4c91-844d-95e69a83ed07/download)

## Creator

Van Dung Nguyen, [Takahiro Nagata](https://orcid.org/0000-0002-8591-2943), Huu Duc Luong, [K.-S. Chang](https://orcid.org/0000-0002-2876-9255)

## Rights

[Creative Commons BY-NC-ND Attribution-NonCommercial-NoDerivs 4.0 International](https://creativecommons.org/licenses/by-nc-nd/4.0/)

## Other metadata

[Electrical and UV-sensing performance of N-doped-Ba(Ti,Zr,Mo,Hf,Ta)O3-dielectric and ZnSnO-channel-based flexible thin-film transistors](https://mdr.nims.go.jp/datasets/5afe3c63-012f-49d9-bf3f-6c17b3f56352)

## Fulltext

1  Supporting Information    Electrical and UV-Sensing Performance of N-doped-Ba(Ti,Zr,Mo,Hf,Ta)O3-Dielectric and ZnSnO-Channel-Based Flexible Thin-Film Transistors  Van Dung Nguyen1, Takahiro Nagata2, Huu Duc Luong3, and K.-S. Chang1,*  1 Department of Materials Science & Engineering, National Cheng Kung University. No.1, University Road, Tainan City 70101, Taiwan. 2 Research Center for Electronic and Optical Materials, National Institute for Materials Science. 1-1 Namiki, Tsukuba, Ibaraki 305-0044, JAPAN. 3Laboratory for Chemistry and Life Science (CLS), Institute of Innovative Research (IIR), Tokyo Institute of Technology.  R1-25, 4259 Nagatsuta-cho, Midori-ku, Yokohama, 226-8501, JAPAN. *e-mail: kschang@mail.ncku.edu.tw; Tel.: +886-6-2757575 ext. 62922.                     2  Figure S1  The constituent elements (at.%) across the N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 film library were quantitatively determined using scanning XRF, and the results are displayed in Figure S1. Oxygen and nitrogen are excluded because of their weak fluorescence signals. Three-dimensional (3D) surface mapping for each element was performed to elucidate their distributions. The Ba (Figure S1a) and Ti (Figure S1b) variations) decrease similarly, as indicated by the dashed arrow in the figures. There are also consistent decreasing trends for Hf (Figure S1c) and Ta (Figure S1d), and Zr (Figure S1e) and Mo (Figure S1f), as indicated by the dashed arrows. The observed elemental distributions are in accordance with our experimental design.       Figure S1. Scanning XRF results of the N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 film library.     Ba TiHf TaZr Mo(a) (b)(c) (d)(e) (f)  3  Figure S2  XPS of Ba(Ti,Zr,Mo,Hf,Ta)O3 with and without N doping were studied. The peaks of constitutive elements on the B site shift to higher binding energy after N doping (upper spectra, Figure S2) compared with those without N doping (lower spectra, Figure S2), which was resulted from the difference in electron density (lower electronegativity of N than O). The VO concentration reduces from ≈25.7 to 13.97 at.% after N doping, indicating a decrease in the defect of the dielectric layer. This is consistent with our results in the bias stress study (Figures 9−11).    Figure S2. XPS results of N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 (upper spectra in all figures) and Ba(Ti,Zr,Mo,Hf,Ta)O3 (lower spectra in all figures): (a) Ba 3d, (b) Ti 2p, (c) Ta 4f, (d) Hf 4f, (e) Zr 3d, and (f) Mo 3d and (g) O 1s.    Binding energy (eV)Intensity (a.u.) (c) Ta5+ 4f7/2Ta5+ 4f5/2Ta4+ Zr 4p23 25 27 29 31 33Binding energy (eV)Intensity (a.u.) (d) Hf4+ 4f7/2 Hf4+ 4f5/2Hfx+12 15 18 21Binding energy (eV)Intensity (a.u.) (e) Zrx+ 3d3/2Zr4+ 3d5/2Ba 4p3/2Zrx+ 3d5/2Zr4+3d3/2174 178 182 186Binding energy (eV)Intensity (a.u.)(a)Ba 3d5/2Ba 3d3/2 Ba2+ (L)Ba2+ (S)Binding energy (eV)Intensity (a.u.)776 782 788 794 800Binding energy (eV)Intensity (a.u.) (b)Ti 2p3/2Ti 2p1/2Ti3+Ti4+454 459 464 469Binding energy (eV)Intensity (a.u.) (f) Mo6+ 3d5/2Mo6+ 3d3/2Mo4+227 230 233 236 239(g) O 1sO-M M-N-OVoH2O13.97 at.%525 528 531 534M-OVo25.7 at.%   4  Figure S3  The high-resolution XPS results of ZnSnO are displayed in Figure S3. The Zn 2p spectrum (Figure S3a) exhibits two distinct peaks at approximately 1021.7 and 1044.9 eV, ascribed to Zn2+ 2p3/2 and Zn2+ 2p1/2, respectively [S1]. In the Sn 3d spectrum (Figure S3b), two peaks at 486.05 and 494.52 eV are associated with Sn4+ 3d5/2 and Sn4+ 3d3/2, respectively [S1]. The O 1s spectrum (Figure S3c) reveals three deconvoluted subpeaks: lattice oxygen (OL) at approximately 530.51 eV, VO (≈9.67 at.%.) at approximately 531.8 eV, and H2O at approximately 532.4 eV [S1].    Figure S3. High-resolution XPS spectra of ZnSnO: (a) Zn 2p, (b) Sn 3d, and (c) O 1s.  [S1] Q. Zhang, C. Ruan, H. Gong, G. Xia, S. Wang, Low-temperature and high-performance ZnSnO thin film transistor activated by lightwave irradiation, Ceram. Int. 47 (2021) 20413-20421. https://doi.org/10.1016/j.ceramint.2021.04.050.  478 484 490 496 502Intensity (a.u.)Binding energy (eV)(b)Sn 3d5/2Sn 3d3/21016 1026 1036 1046Intensity (a.u.)Binding energy (eV)Zn 2p3/2Zn 2p1/2(a) (b)(c)526 529 532 535Intensity (a.u.)Binding energy (eV)O-MVoH2O(9.67 at.%)  5  Figure S4  Figure S4a shows four PL emission bands (≈360, 409.1, 550, and 705 nm) of N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3. The emission at 409.1 nm is due to shallow defects (e.g., VO). The emission at 550 nm is associated with the recombination of induced electron−hole pairs due to the surface state. The emission at 705 nm is ascribed to the presence of defects such as Ti3+ interstitials or Ti3+ vacancies. Figure S4b shows the PL spectrum of ZnSnO. The strong luminescence band at ≈335 nm is due to the recombination of free excitons in an energy level close to the band edges of ZnSnO. The emissions at 412 and 545 nm are ascribed to electron transitions from the shallow and deep Vo levels to EC, respectively. The emission at 700 nm is associated with the presence of defects such as Zn vacancies and/or oxygen interstitials.     Figure S4. PL spectra of (a) N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 and (b) ZnSnO.                   6  Figure S5  Figure S5 compares the dielectric properties of the MIM structure before and after FGA treatment. Without FGA, low Cox values are observed from 20 Hz to 1 MHz, and Cox at 1 MHz is substantially lower than that at other frequencies (Figure S5a). However, when the configuration was treated with FGA at 350 °C for 40 min, there is only marginal frequency-dependent dispersion from 20 Hz to 1 MHz and substantially enhanced Cox (Figure S5b), and both features reveal the effective minimization of existing Dit in the film. In the JL–V characteristics of Figure S5c (inset: the measurement setup), the JL value after FGA (blue color) is remarkably smaller than that before FGA (red color). Moreover, the JL value after FGA (≈4.3 × 10−11 A⋅cm−2 at 1 V) is three orders of magnitude lower than that exhibited by other high-entropy-oxide-based devices reported in the literature [30, 34, 35].     Figure S5. Dielectric properties for an MIM stack comprising favorable N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3. C−V curves from 20 Hz to 1 MHz (a) before and (b) after FGA. (c) JL–V curves before (red) and after (blue) FGA.        0123-5 -3 -1 1 3 5(b)Capacitance density(µF/cm2 )Bias (V)-14-10-6-2-5 -3 -1 1 3 5After FGABefore FGA(c)Bias (V)Log [JL(A/cm2 )]0123-5 -3 -1 1 3 5(a)Capacitance density(µF/cm2 )Bias (V)20 Hz 100 Hz500 Hz 1 kHz10 kHz 100 kHz1 MHz20 Hz 100 Hz500 Hz 1 kHz10 kHz 100 kHz1 MHz  7  Figure S6  To investigate the impact of each element in the N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 film on the dielectric properties, N-doped medium-entropy-oxide [MEO, 3−4 elements on the B site, including Ba(Ti,Zr,Hf)O3, Ba(Ti,Hf,Ta)O3, Ba(Ti,Zr,Ta)O3, and Ba(Ti,Zr,Hf,Ta)O3] and N-doped low-entropy-oxide [LEO, 1−2 elements on the B site, including Ba(Ti,Ta)O3, Ba(Ti,Zr)O3, Ba(Ti,Hf)O3, and BaTiO3] films were studied (Figure S6). Their capacitance densities at 1 kHz (Figure S6a) and leakage current densities (Figure S6b) were improved when more principal elements were incorporated because of enhanced configurational entropy. N-doped BaTiO3 exhibited the lowest capacitance densities (k ≈ 31) and highest leakage current densities. By contrast, N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 exhibited the most promising dielectric properties among them. To investigate the N-doping impact on the dielectric properties of Ba(Ti,Zr,Mo,Hf,Ta)O3, the samples fabricated using various amounts of N2 (0−4 sccm) with the same thickness were studied. These samples are named as N0, N1, N2, N3, and N4, respectively. From the C−V characteristics (Figure S6c), the enhanced capacitance density is observed with an increase in the amount of N2 to 3 sccm. However, the value slightly decreases at N2 = 4 sccm. The similar variation in the leakage current density is also observed (Figure S6d). The sample fabricated without any N2 (N0) exhibited the lowest capacitance density and highest leakage current density. Their k values are displayed in Figure S6e. Ba(Ti,Zr,Mo,Hf,Ta)O3 with k ≈ 147 is observed. The N3 sample exhibited the most promising dielectric properties among the samples.                         8    Figure S6. Dielectric properties exhibited by various HEO-, MEO-, and LEO-based MIM structures: (a) C−V curve (capacitance density measured 1 kHz) and (b) Log JL−V curve. Dielectric properties exhibited by various N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3-based MIM structures at various amounts of N2: (c) C−V curve (1 kHz), (d) Log JL−V curve, and (e) associated k values.       N-doped Ba(Ti,Hf)O3 (LEO)N-doped Ba(Ti,Zr)O3 (LEO)N-doped Ba(Ti,Ta)O3 (LEO)N-doped Ba(Ti,Zr,Hf)O3 (MEO)N-doped Ba(Ti,Hf,Ta)O3 (MEO)N-doped Ba(Ti,Zr,Ta)O3 (MEO)N-doped BaTiO3 (LEO)N-doped Ba(Ti,Zr,Hf,Ta)O3 (MEO)N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 (HEO)0123-2 -1 0 1 2Capacitance density(µF/cm2 )(a)Bias (V)1 kHz-14-10-6-2-5 -3 -1 1 3 5(b)Bias (V)Log [JL(A/cm2 )]0123-2 -1 0 1 2Capacitance density(µF/cm2 )(c)Bias (V)N0 N1 N2N3 N4-16-12-8-40-2 -1 0 1 2(d)Bias (V)Log [JL(A/cm2 )]N0 N1 N2N3 N4146.7186.1240.2322.2285.40100200300400N0 N1 N2 N3 N4(e)k values   9  Figure S7  The current on/off ratio and SS of all transfer characteristics for the device at various PGBS times without and with UV illumination are depicted in Figure S7a and S7b, respectively. Only minor changes are observed.   Figure S7. Current on/off ratio and SS variation at various PGBS times (a) without and (b) with UV illumination.     0.000.050.100.150.200.2501230 100 500 1000 2000 3600SS (Vdec-1)I ON/IOFF(×108 )Stress time (s)PGBS = 3 V0.000.050.100.150.200.2501230 100 500 1000 2000 3600SS (Vdec-1)I ON/IOFF(×105 )Stress time (s)PGBS = 3 V; UV treated(a)(b)  10  Figure S8  The energy band diagrams of the N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 and ZnSnO layers in the TFT without any stress were constructed (Figure S8). UV−Vis spectroscopy was employed to obtain their Eg values (≈3.97 eV in Figure S8a and ≈3.7 eV in Figure S8b, respectively) through Tauc plots [(αhν)2 against hν, where α is the absorption coefficient, h denotes Planck’s constant, and ν indicates the frequency]. The Φ values (≈4.4 eV in Figure S8c and ≈4.3 eV in Figure S8d, respectively) are obtained through UPS by subtracting the cutoff energy (Ecutoff) of the two films (upper right insets of Figure S8c and S8d, respectively) from the source energy (≈21.2 eV). The EVmax values of N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 and ZnSnO (≈6.5 and 7.53 eV, respectively) were then calculated through the relationship of EVmax = ΦS + EV, offset (valence band offset), assuming EV, offset ≈ 2.1 eV for N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 and 3.23 eV for ZnSnO, as indicated in the upper left insets of Figure S8c and S8d, respectively. The EC values of ≈2.53 eV for N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 and ≈3.83 eV for ZnSnO were then obtained through the equation of EC = EVmax – Eg. The resulting energy band diagrams of the two films before and after contact on a vacuum scale are displayed in the left and right sides of Figure S8e, respectively. The diagram after contact helps elucidate the charge transport and trapping and photodetection behavior of the device, as discussed in Figure 9 and 12, respectively.   0 5 10 15 20 25 30Intensity (a.u.)Binding energy (eV)15 16 17 18Intensity (a.u.)Binding energy (eV)Ecutoff = 16.8 eV1 2 3 4Intensity (a.u.)Binding nergy (eV)Ev, offset = 2.1 eV(c)0 5 10 15 20 25 30Intensity (a.u.)Binding energy (eV)16.5 17 17.5Intensity (a.u.)Binding energy (eV)Ecutoff = 16.9 eV1 2 3 4Intensity (a.u.)Binding nergy (eV)Ev, offset = 3.23 eV(d)3 3.3 3.6 3.9 4.2(𝛼ℎ𝜈)2 (a.u.)Eg = 3.97 eVℎ𝜈 (eV)(a)2.5 3 3.5 4 4.5(𝛼ℎ𝜈)2 (a.u.)Eg = 3.7 eVℎ𝜈 (eV)(b)N-doped Ba(Zr,Ta,Ti,Hf,Mo)O3ZnSnOZnSnO  11    Figure S8. Tauc plots of (a) N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 and (b) ZnSnO. UPS results of (c) N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 and (d) ZnSnO. The upper right and upper left insets show the Ecutoff and EV, offset of the two films, respectively. (e) Energy band diagrams of the two films before (left) and after (right) contact.     (e)Before contactVacuum-8-7-6-5-4-3-2-10After contactN-doped Ba(Ti,Zr,Mo,Hf,Ta)O3ITOZTO ECEFEF∆ = 1.3 eV∆ = 0.23 eVVacuumΦITO= 4.53 eV-8-7-6-5-4-3-2-10ΦITO= 4.53 eVΦHEON= 4.4 eVEc= 2.53 eVEg= 3.97 eV1.87 eV2.1 eVΦ = 4.3 eVEc= 3.83 eV3.23 eVEg= 3.7 eVITON-doped Ba(Ti,Zr,Mo,Hf,Ta)O3 ZTO   12  Figure S9  The measurement procedure for NGBSI is illustrated in the lower right inset of Figure S9a, which is similar to that for PGBSI except for the application of NGBS = −3 V. In general, all transfer characteristics at various NGBS times (0−3600 s) without UV illumination (Figure S9a) also exhibit trivial variations in the current on/off ratio and SS (Figure S10a) and negligible changes in IDS, max, which is further elucidated in the upper left inset (blue curve). These features indicate that no operation adjustment is needed for turning the device on and off, and there is no deterioration in the current flow. These curves also exhibit remarkable stability with only a small negative VT shift (ΔVT ≈ −0.058 V, orange curve in the upper left inset of Figure S9a) when the NGBS time increases from 0 to 3600 s. The associated energy band diagram is depicted in Figure S9b. Upward band bending of ZnSnO at the N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3/ZnSnO interface facilitates the trapping of holes induced by the negative electric field by neutral VO-related traps (donor-like defects) in the Eg of N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3. When the NGBS time increases, more holes are induced, trapped, and distributed in the deep levels of the gap or into the bulk N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3. These accumulated hole traps induce a mirror layer of electrons (high charge carrier concentrations) in the ZnSnO channel layer, leading to a minor decrease in VT (ΔVT ≈ −0.058 V) [67-69]. However, when the device is additionally subjected to UV illumination (Figure S9c), all transfer curves at various NGBS times (0−3600 s) exhibit noticeable changes in the current on/off ratio and SS (Figure S10b). A remarkable increase in IDS, max (blue curve in the inset of Figure S9c) is also observed. Moreover, relative to the curve without UV illumination (black curve in Figure S9c), all curves under UV illumination are shifted substantially to negative VGS (ΔVT ≈ −1.34 V, orange curve in the inset of Figure S9c), indicating a high concentration of induced charge carriers in the channel because of the generation of electron−hole pairs by the light absorption of ZnSnO. The associated energy band diagram is depicted in Figure S9d, illustrating the generation of extra illumination-induced charge carriers and hole trapping (inducing mirror layers of electrons). Donor-like defects (i.e., neutral VO) are evident in the XPS (Figure S3) and PL (Figure S4) spectra.   13   Figure S9. NGBSI study of the TFT at NGBS = −3 V for various times and VDS = 2 V. (a) Transfer curves without UV illumination. The lower right and upper left insets show the measurement procedure and detailed variations in IDS, max and VT, respectively. (b) Energy band diagram without UV illumination. (c) Transfer curves with UV illumination. The inset shows the detailed variations in IDS, max and VT. (d) Energy band diagram with UV illumination.                    ITOZTO ECEFhvEV(d)Vgs, stress < 0ITOZTO ECEFEV(b)Vgs, stress < 0N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3N-doped Ba(Ti,Zr,Mo,Hf,Ta)O3++++(h+)(e-)VO-related traps(e-)1E-131E-101E-071E-04-4 -2 0 2 4 6I DS(A)VGS (V)NGBS = -3 V-0.2-0.10.00.100.40.81.20100500100020003600VT(V)I DS, max (mA)Stress time (s)VDS = 2 VVT(a)I-V -3 V(100 s)Recovery(15 min)I-V0 s 100 s500 s 1000 s2000 s 3600 s1E-091E-071E-051E-03-4 -2 0 2 4I DS(A)VGS (V)-2.6-2.1-1.6-1.1-0.600.511.520100500100020003600VT(V)I DS, max (mA)Stress time (s)NGBS = -3 VVTVDS = 2 V(c)0 s100 s500 s1000 s2000 s3600 sUV treated  14  Figure S10   Figure S10. Current on/off ratio and SS variation of all transfer characteristics of the TFT at NGBS = −3 V for various times and VDS = 2 V (a) without and (b) with UV illumination.       15  Figure S11  EA measures the energy required for electrons to jump from EF to EC (EC − EF) in a thermal process, and a higher EA indicates more difficulty for electron hopping. In the subthreshold region of transfer characteristics for a TFT, IDS is described by Equation S1:[70]  IDS = ID0 × exp[−EA(VGS)/KT] ≈ α × (VGS − VT) × exp[−EA(VGS)/KT]                  (S1), where ID0 is a prefactor, EA(VGS) is the activation energy as a function of VGS, K is the Boltzmann constant, and T is the temperature. Taking the logarithm on both sides of Equation S1, we obtain Equation S2.         ln(IDS) = [−EA(VGS)/KT] + ln α × (VGS − VT)                                                  (S2).  Using Equation S2, EA(VGS) can be deduced from the plot of ln(IDS) against 1/KT. Figure S11a and b plot the results for the device treated with NGBS (−3 V for 3600 s) and both NGBS and UV illumination from room temperature to 180 °C, respectively. The NGBS effect is elucidated by the observation of EA ≈ 0.39 eV (VGS = 0.2 V) and a smaller slope (0.91 eV⋅V−1 (inset of Figure S11a) compared with that (0.94 eV⋅V−1) without any treatment (inset of Figure 11a). When the device was subjected to both NGBS and illumination, their respective impact on the induction of traps and carriers (accumulated holes inducing a mirror layer of electrons in ZnSnO) in the channel is reflected by a substantially smaller slope of 0.564 eV⋅V−1 and the considerably negatively shifted curve in VGS with slightly reduced EA ≈ 0.26 eV (VGS = −2.2 V, inset of Figure S11b).    Figure S11. Plots of ln(IDS) against 1/KT (room temperature to 180 °C) at various VGS, and EA(VGS) against VGS (insets) of the TFT: (a) with NGBS (−3 V for 3600 s) and without UV illumination, and (b) with NGBS (−3 V for 3600 s) and UV illumination.  1E-281E-221E-161E-101E-0424 27 30 33 36 39Ln(I DS) (A)1/KT (eV-1)(a) NGBS = -3 V for 3600 sVGS00.20.40.6-3 -2 -1 0 1 2 3 4EA(eV)VGS (V)Falling rate = 0.91 eV/V off stateOn state4V2V3V0.5V1V0.4V0.3V 0.2V0.1VLn(I DS) (A)1E-281E-221E-161E-101E-0424 27 30 33 36 39(b) NGBS = -3 V for 3600 s; UV treated1/KT (eV-1)VGS00.20.40.6-3 -2 -1 0 1 2 3 4EA(eV)VGS (V)On stateoff stateFalling rate = 0.564 eV/V 4V 3V2V0V1V-1V-1.5V -1.9V-2V -2.1V-2.2V