In this letter, the reduction and removal of surface native oxide from as-received InGaAs surface by using dimethylaluminumhydride-derived aluminum oxynitride (AlON) passivation layer prior to HfTiO deposition is proposed to solve Fermi level pinning issue. It has been revealed that complete consumption of native oxides of AsOx and GaOx at the InGaAs surface, but no effect to InOx, has been realized after metalorganic chemical vapor deposition AlON at 300 °C. X-ray photoelectron spectroscopy observations of HfTiO/InGaAs gate stacks demonstrate that introducing AlON layer can suppress the regrowth of native oxide at the interface. In addition, the dependence of the valence band spectra of HfTiO/InGaAs gate stacks on AlON passivation layer has been discussed in detail.

The increasing need for higher speed and lower power consumptions has pushed the Si-based devices to their performance limit. Currently, more attention has been paid to the investigation of channel materials with higher electron mobility and appropriate effective mass, such as Ge and GaAs.1–3 Among these channel materials, III-V compound semiconductor such as InGaAs, with its lattice matched to InP, has been used as a backbone for almost all the electronic devices with high speed, for example, high electron mobility transistor (HEMT) with a high cutoff frequency ∼562 GHz and high performance metal-oxide-semiconductor field-effect transistors (MOSFET).4,5 However, in spite of this continuous demand for the MOSFET application, for the successful realization of InGaAs-based MOSFETs, there still exist a number of fundamental issues which should be solved. One of the most critical challenge is the poor interfaces with high interface trap density (Dit),6 which causes the Fermi level pinning, degradation of the drive current, and the sub-threshold swing. Thus, the high-k/InGaAs interface control is strongly needed for realizing low Dit. Fortunately, recent progress made in the combination of the surface passivation technology and atomic layer deposition (ALD) of Al2O3 layer has led to the demonstration of InGaAs MOSFET with high performance.7–12 

Another challenge for high-k/InGaAs gate stack is the reduction in equivalent oxide thickness (EOT) of high-k gate dielectric. Although the Al2O3/InGaAs interfaces have relatively low Dit, the Al2O3 dielectric constant (∼9) is not enough for the EOT scaling. Due to the high dielectric constant and good thermal stability, HfO2 has been regarded as a promising high-k gate dielectric for scaled InGaAs MOSFETs.13,14 However, it has been reported that the HfO2/InGaAs interfaces have the inferior MOS interface properties to the Al2O3/InGaAs.15 Therefore, the improvement of the HfO2/InGaAs interface properties is strongly needed to achieve low EOT and low gate leakage current with maintaining the excellent interface properties. Currently, Ti-doped Hf-based high-k gate dielectrics have demonstrated an adjustable dielectric constant, remarkable thermal stability, and improved electrical properties in MOS device.16–19 However, direct deposition of Hf-based high-k gate dielectrics on InGaAs exhibits anomalous characteristics with larger frequency dispersion, hysteresis, and also low effective mobility, originating from the oxides-induced interface pining.13,20 Hence, the necessity of surface passivation prior to the high-k gate dielectric deposition to minimize the oxides formation and eliminate the Fermi level pinning effect has been addressed by some researchers. Examples include in situ and ex situ deposition of amorphous silicon/germanium interlayers or nitridation pretreatment before high-k gate deposition.21–23 However, it can be noted that Si and Ge are amphoteric dopants for GaAs, a thin layer of Si or Ge between HfO2 and GaAs may alter the doping concentration or even induce the counter doping of the GaAs substrate, causing the instability of the threshold voltage.24 The development of alternative passivation process for InGaAs surface is important and needed.

Recent studies have also demonstrated a self-cleaning effect on the reduction and removal of surface oxides from GaAs substrate by metalorganic chemical vapor deposition (MOCVD) of Al2O3 using dimethylaluminumhydride [(CH3)2AlH, DMAH] and O2.3,25 The exact mechanism of this self-cleaning effect is still unclear, though much progress has been made. In this letter, we report the demonstration of DMAH-derived aluminum oxynitride (AlON) interfacial passivation layer for p-type InGaAs and its application in HfTiO-InGaAs gate stacks grown by sputtering. Although there exist some reports on the electrical properties of HfO2/GaAs gate stacks with sputtering-derived AlON interfacial passivation layer,3,26 few investigation on the evolution of the interface chemistry of HfTiO/InGaAs gate stacks, as well as the band alignment of the HfTiO/InGaAs, originating from the effective passivation of InGaAs surface by DMAH-derived AlON layer has not been fully identified yet. Additionally, based on previous publication that chemical reactions at an InGaAs surface during MOCVD process can be affected by the MOCVD temperature, we also investigated the impact of MOCVD temperature on the AlON/InGaAs interface to understand the factors determining the interface properties and band alignment of HfTiO/InGaAs gate stacks.

The substrates used in this work were commercially available p-type (100) In0.53Ga0.47As/InP wafers. At first, the wafers were degreased using acetone and isopropnol to remove organic matters and other impurity ions, and remained native oxide layer adhered to the surface of the substrates. After surface treatment, each sample was followed by rinsing with de-ionized water and drying with N2 gas. Then, ex situ AlOxNy gate dielectric with thickness of 2 nm was deposited at 200 °C, 250 °C, and 300 °C by MOCVD, using DMAH and O2 as a precursor and oxidant. Propylamine, 1.0 vol.%, was added to DMAH to reduce its viscosity. Due to this additive, the deposited film induces 5.8% of nitrogen.3 HfTiO gate dielectrics with thickness of 15 nm have been deposited on InGaAs surface covered with AlON interlayer by cosputtering of Hf and Ti targets in a mixed ambient of O2 and Ar at room temperature. For comparison, HfTiO was also directly deposited on InGaAs wafer without AlON passivation layer with cosputtering of Hf and Ti targets at room temperature. The observed chemical ratio of the HfTiO film is HfTi1.05O4.98 by characterization from x-ray photoelectron spectroscopy (XPS) measurements. Additionally, XPS was performed to investigate the effect of the MOCVD temperature on chemical bonding states at the HfTiO/InGaAs interfaces and the physical origins of improvements the HfTiO/InGaAs MOS interface properties by introducing the AlON passivation layer. Here, the XPS ESCALAB MK (VG UK) system is equipped with monochromatic Mg Kα source under a base pressure of 2.1 × 10−9 Torr and a hemispherical analyzer with a pass energy of 20 eV. C 1s peak at 284.6 eV was taken as a reference for charge correction, and spectral deconvolution was performed by Shirley background subtraction using a Voigt function convoluting Gaussian and Lorentzian functions.

XPS analyses for the AlON/InGaAs interfaces were carried out to study the impact of the MOCVD temperature and AlON passivation layer on chemical bonding states at the MOS interfaces and the physical origins. Figure 1 shows the XPS spectra of As 3d, Ga 2p, and In 3d core-level spectra from an InGaAs surface after MOCVD AlON passivation layer depositions at different deposition temperatures compared to the InGaAs wafer without AlON deposition. It has been found that InGaAs surface is covered native oxide only after precleaning. The study of interfacial native oxide is critical to understand the electrical behavior of InGaAs-based MOS devices. Especially, As–As dimmers and Ga oxide are the unstable species that cause oxide-induced Fermi level pinning.27,28 Based on Fig. 1(a), it can be noted that more AsOx exists at the AlON/InGaAs interface with lowering the MOCVD temperature. The amount of AsOx is higher than that of the reported Al2O3/InGaAs interfaces, indicating that As oxidation occurs during MOCVD of AlON and As oxides remain even at low temperature.13 Figure 1(b) shows Ga 2p core level spectra from the AlON/InGaAs interface deposited at various temperatures. It is found that GaOx at the AlON/InGaAs interfaces also tends to increase with lowering the MOCVD temperature. Similar self-cleaning effect phenomenon has been observed for HfO2/InGaAs and HfAlO/InGaAs gate stacks.13,14 Figure 1(c) shows XPS In 3d spectra from the AlON/InGaAs interfaces deposited at various temperature. Judging from the In 3d spectra in Fig. 1, although the amount of InOx demonstrates a slight reduction with increasing the MOCVD temperature, MOCVD of AlON cannot remove InOx component fully. Therefore, it can be inferred that any MOCVD temperature dependence of the AlON/InGaAs interfaces would not be attributable to the existence of InOx but to those of AsOx and GaOx. On the other hand, the reason why the InOx remains almost unchanged at the lower MOCVD temperature is still not clear by far. Further detailed studies are needed to provide the appropriate reason. Based on our analysis, it can be concluded that oxidation of InGaAs surface prior to HfTiO deposition is controlled effectively by using AlON interfacial passivation layer deposited at appropriate temperature. Chang et al. also reported the self-cleaning effect of As oxide reduction by ALD HfO2 using TEMAH source, while the existence of In2O3, In(OH)3, and Ga2O3 at the interface between the ALD HfO2 and n-type In0.53Ga0.47As channel layer was observed.29 Our experimental observation is not consistent with the report, which can originate from the difference in high-k deposition process including pre-deposition wet chemical cleaning process.

FIG. 1.

(a) As 3d, (b) Ga 2p, and (c) In 3d XPS spectra recorded from native oxides/InGaAs and MOCVD AlON/InGaAs at different deposition temperatures.

FIG. 1.

(a) As 3d, (b) Ga 2p, and (c) In 3d XPS spectra recorded from native oxides/InGaAs and MOCVD AlON/InGaAs at different deposition temperatures.

Close modal

To further identify the full removal of native oxide of GaOx and AsOx attributed to the self-cleaning effect of AlON passivation layer deposited at 300 °C more clearly, the depth profile of the Ga 2p and As 3d spectra were investigated by angular-resolved XPS (ARXPS), as shown in Fig. 2. As we know, for ARXPS measurement, as θ increases, the probing depth becomes shallow and the region close to the surface is probed. According to Fig. 2(a), it can be noted that no apparent angular dependence of the Ga 2p spectra has been observed, indicating no formation of Ga oxide for InGaAs surface during MOCVD AlON passivation layer deposition at 300 °C. For As 3d spectra shown in Fig. 2(b), the same trend has happened. After AlON passivation layer deposition, no angular dependence of As 3d spectra can be detected, which can be attributed to the fact that a self-cleaning effect on the reduction and removal of surface oxides from InGaAs surface by MOCVD of AlON using DMAH has been demonstrated.30,31 Therefore, it can be concluded that DMAH-derived AlON passivation layer can act as an oxygen reaction barrier and minimize the formation of oxides with low quality which could cause Fermi level pinning and degrade the device electrical performance.

FIG. 2.

Angular dependent Ga 2p (a) and As 3d (b) XPS spectra for AlON/InGaAs deposited at 300 °C.

FIG. 2.

Angular dependent Ga 2p (a) and As 3d (b) XPS spectra for AlON/InGaAs deposited at 300 °C.

Close modal

To investigate the effect of 300 °C-derived AlON passivation layer on the interface chemistry and bonding states of HfTO/InGaAs gate stacks, O 1s, Hf 4f, and In 3d core level spectra have been paid more attention, as shown in Fig. 3. For O 1s spectra of HfTiO/InGaAs system with AlON layer (Fig. 3(a)), the deconvoluted spectra only show four peaks, which are assigned as Hf–O, Ti–O, Al–O–N, and In–O bonding states, respectively. Only InOx native oxide has been observed at the interface. For the samples without AlON layer, the O 1s spectra demonstrate the existence of Ga–O and As–O bonding states, besides the previously mentioned chemical bonding states, suggesting that the native oxide of GaOx and AsOx in InGaAs surface has been effectively suppressed by introducing AlON passivation layer. To confirm our conclusion, Hf 4f spectra, demonstrated in Fig. 3(b), have been paid more investigation. For the samples with AlON layer, Hf 4f spectra only show Hf–O and In–Ga–As bonding states. However, additional peak attributed to Ga–O binding state has been observed for HfTiO/AlON/InGaAs gate stacks, indicating that the GaOx native oxide still exists at the interface of HfTiO/InGaAs. For the In 3d spectra shown in Fig. 3(c), all the deconvoluted spectra demonstrate two components. One comes from InGaAs substrate and the other is due to the existence of InOx at the interface of HfTiO/InGaAs. Based on our analysis, it can be noted that 300 °C-derived AlON layer only has effect on removing the native oxide of GaOx and AsOx from InGaAs surface, but no effect on removing InOx component. Meanwhile, the direct sputtering deposition of HfTiO cannot remove any native oxide of InGaAs surface. The further detailed reason for the existence of the native oxide at the interface HfTiO/InGaAs may be needed.

FIG. 3.

O 1s (a), Hf 4f (b), and In 3d (c) XPS spectra for HfTiO/InGaAs gate stacks with and without AlON passivation layer.

FIG. 3.

O 1s (a), Hf 4f (b), and In 3d (c) XPS spectra for HfTiO/InGaAs gate stacks with and without AlON passivation layer.

Close modal

To accomplish a good high-k gate stack, the high-k must have sufficient band offsets of over 1 eV to act a barrier for both electrons and holes as well as be stable in contact with the semiconductor. The valence band (VB) maximum (Ev) of each sample is determined by extrapolating the leading edge of valence band spectrum to the base line, the cross point is taken to be Ev. Figure 4 demonstrates the dependence of VB alignment of the HfTiO/InGaAs gate stack on the AlON passivation layer. Based on the corrected leading edge of valence-band spectra for as-cleaned InGaAs (0.6 eV) and as-deposited HfTiO, the valence-band offsets (ΔEv) between the as-cleaned InGaAs substrates without/with AlON passivation layer and HfTiO, as highlighted in Fig. 4, are found to be 2.98 and 2.82 eV, respectively. The difference in ΔEv can mainly be attributed to the evolution of the interfacial component. For HfTiO/InGaAs sample, the existence of the native oxide will affect the ΔEv considerably; however, in HfTiO/AlON/InGaAs sample, the value of ΔEv is mainly determined by the valence band spectra of AlON due to the lack of interfacial oxide. The increase in ΔEv can be understood by the effect of the oxides existing at the interface between HfTiO and InGaAs, which is in good agreement with the conclusion form Robertson that the band offset of high-k oxides/high mobility substrates can be increased due to the interfacial layer.32 Taking the measured energy-bandgap of MOCVD-derived AlON and sputtering-deposited HfTiO to be 7.20 and 4.60 eV,18,19,33 and together with the In0.53G0.47aAs energy-bandgap as 0.74 eV,14 the conduction-band offset (ΔEc) of 1.04 eV is deduced with AlON passivation layer, whereas 0.88 eV without the passivation, as illustrated in Figs. 5(a) and 5(b). Based on VB spectra analysis, it can be noted that the band offset of HfTiO/InGaAs gate stacks with AlON passivation layer is adequate for the gate dielectric operation with energy barriers over 1 eV. However, compared to the reported ΔEc of 1.97 eV for p-InGaAs/HfAlO, the lowering of the ΔEc in HfTiO/AlON/InGaAs gate stack may be due to the smaller bandgap of HfTiO dielectrics.14 Therefore, Ti component in the films should be carefully controlled to guarantee excellent properties of HfTiON/AlON/InGaAs gate stacks in future devices.

FIG. 4.

Valence-band spectra of InGaAs and HfTiO/InGaAs with/without AlON passivation layer.

FIG. 4.

Valence-band spectra of InGaAs and HfTiO/InGaAs with/without AlON passivation layer.

Close modal
FIG. 5.

Schematic energy band alignment of HfTiO/InGaAs gate stack with and without AlON passivation layer.

FIG. 5.

Schematic energy band alignment of HfTiO/InGaAs gate stack with and without AlON passivation layer.

Close modal

In conclusion, surface passivation of the as-received InGaAs by using DMAH-derived AlON with self-cleaning effect prior to HfTiO deposition has been carried out. Results demonstrate that 300 °C-derived AlON fully remove the native oxides of GaOx and AsOx at the InGaAs surface, but no effect to InOx. XPS analysis of HfTiO/InGaAs gate stacks reveals that AlON passivation layer can effectively suppress the regrowth of native oxides (GaOx, AsOx) at the interface between InGaAs and as-deposited HfTiO. Measurements of valence band spectra of HfTiO/AlON/InGaAs gate stacks show reduction in valence band offset and increase in conduction band offset compared to that of HfTiO/InGaAs. The appropriate band offset relative to InGaAs and excellent interface properties render HfTiO/AlON/InGaAs promising gate stacks in future III-V-based devices.34 

The authors acknowledge the support from Anhui Provincial Natural Science Foundation (1208085MF99), National Key Project of Fundamental Research (2013CB632705), National Natural Science Foundation of China (51072001, and 51272001), National Science Research Foundation for Scholars Return from Overseas of Chinese Ministry of Education, Provincial Natural Science Foundation of Anhui Higher Education Institution of China (KJ2012A023), Key Project of Chinese Ministry of Education (212082), and Outstanding Young Scientific Foundation of Anhui University (KJJQ1103) and “211 project” of Anhui University.

1.
T.
Tabata
,
K.
Nagashio
, and
A.
Toriumi
,
Appl. Phys. Exp.
5
,
091002
(
2012
).
2.
S. V.
Ovsyannikov
,
I. V.
Korobeinikov
,
N. V.
Morozova
,
A.
Misiuk
,
N. V.
Abrosimov
, and
V. V.
Shchennikov
,
Appl. Phys. Lett.
101
,
062107
(
2012
).
3.
G.
He
,
L. D.
Zhang
,
M.
Liu
, and
Z. Q.
Sun
,
Appl. Phys. Lett.
97
,
062908
(
2010
).
4.
Y.
Yamashita
,
A.
Endoh
,
K.
Shinohara
,
K.
Hikosaka
,
T.
Matsui
,
S.
Hiyamizu
, and
T.
Mimura
,
IEEE Electron Device Lett.
23
,
573
(
2002
).
5.
J. J. J.
Gu
,
H.
Wu
,
Y. Q.
Liu
,
A. T.
Neal
,
R. G.
Gordon
, and
P. D.
Ye
,
IEEE Electron Device Lett.
33
,
967
(
2012
).
6.
P. D.
Ye
,
J. Vac. Sci. Technol. A
26
,
697
(
2008
).
7.
E. J.
Kim
,
E.
Chagarov
,
J.
Cagnon
,
Y.
Yuan
,
A. C.
Kummel
,
P. M.
Asbeck
,
S.
Stemmer
,
K. C.
Saraswat
, and
P. C.
Mclntyre
,
J. Appl. Phys.
106
,
124508
(
2009
).
8.
C. L.
Hinkle
,
A. M.
Sonnet
,
E. M.
Vogel
,
S.
McDonnell
,
G. J.
Hughes
,
M.
Milojevic
,
B.
Lee
,
F. S.
Aguirre-Tostado
,
K. J.
Choi
,
H. C.
Kim
,
J.
Kim
, and
R. M.
Wallace
,
Appl. Phys. Lett.
92
,
071901
(
2008
).
9.
M.
Milojevic
,
F. S.
Aguirre-Tostado
,
C. L.
Hinkle
,
H. C.
Kim
,
E. M.
Vogel
,
J.
Kim
, and
R. M.
Wallece
,
Appl. Phys. Lett.
93
,
202902
(
2008
).
10.
R. S.
Wang
,
M.
Xu
,
P. D.
Ye
, and
R.
Huang
,
J. Vac. Sci. Technol. B
29
,
041206
(
2011
).
11.
M. L.
Hung
,
Y. C.
Chang
,
C. H.
Chang
,
Y. J.
Lee
,
P.
Chang
,
J.
Kwo
,
T. B.
Wu
, and
M.
Hong
,
Appl. Phys. Lett.
87
,
252104
(
2005
).
12.
C. L.
Hinkle
,
E. M.
Vogel
,
P. D.
Ye
, and
R. M.
Wallece
,
Curr. Opin. Solid State Mater. Sci.
15
,
188
(
2011
).
13.
R.
Suzuki
,
N.
Taoka
,
M.
Yokoyama
,
S. H.
Kim
,
T.
Hoshii
,
T.
Maeda
,
T.
Yasuda
,
O.
Ichikawa
,
N.
Fukuhara
,
M.
Hata
,
M.
Takenaka
, and
S.
Takagi
,
J. Appl. Phys.
112
,
084103
(
2012
).
14.
H. J.
Oh
,
J. Q.
Lin
,
S. J.
Lee
,
G. K.
Dalapati
,
A.
Sridhara
,
D. Z.
Chi
,
S. J.
Chua
,
G. Q.
Lo
, and
D. L.
Kwong
,
Appl. Phys. Lett.
93
,
062107
(
2008
).
15.
G. J.
Burek
,
Y.
Hwang
,
A. D.
Carter
,
V.
Chobpattana
,
J. J. M.
Law
,
W. J.
Mitchell
,
B.
Thibeault
,
S.
Stemmer
, and
M. J. W.
Rodwell
,
J. Vac. Sci. Technol. B
29
,
040603
(
2011
).
16.
C. X.
Li
,
X.
Zou
,
P. T.
Lai
,
J. P.
Xu
, and
C. L.
Chan
,
Microelectron. Reliab.
48
,
526
(
2008
).
17.
J. P.
Xu
,
X.
Zou
,
P. T.
Lai
,
C. X.
Li
, and
C. L.
Chan
,
Thin Solid Films
517
,
2892
(
2009
).
18.
G.
He
,
Z. Q.
Sun
,
M.
Liu
, and
L. D.
Zhang
,
Appl. Phys. Lett.
97
,
192902
(
2010
).
19.
G.
He
,
Z. Q.
Sun
,
Y. Q.
Ma
,
M. Z.
Wu
,
Y. M.
Liu
,
S. W.
Shi
,
G.
Li
,
X. S.
Chen
,
L. D.
Zhang
, and
Z. B.
Fang
,
Semicond. Sci. Technol.
26
,
105019
(
2011
).
20.
P.
Chang
,
W. C.
Lee
,
M. L.
Huang
,
Y. J.
Lee
, and
M.
Hong
,
J. Vac. Sci. Technol. B.
28
,
C3A9
(
2010
).
21.
H.
Hasegawa
,
M.
Akazawa
,
A.
Domanowska
, and
B.
Adamowicz
,
Appl. Surf. Sci.
256
,
5698
(
2010
).
22.
H.
Hasegawa
and
M.
Akazawa
,
Appl. Surf. Sci.
255
,
628
(
2008
).
23.
H. S.
Kim
,
I.
Ok
,
M.
Zhang
,
F.
Zhu
,
S.
Park
,
J.
Yum
,
H.
Zhao
,
J.
Oh
,
P.
Majhi
, and
J. C.
Lee
,
Appl. Phys. Lett.
92
,
102906
(
2008
).
24.
H. S.
Kim
,
I.
Ok
,
M.
Zhang
,
F.
Zhu
,
S.
Park
,
J.
Yum
,
H.
Zhao
,
J. C.
Lee
,
J.
Oh
, and
P.
Majhi
,
Appl. Phys. Lett.
92
,
032907
(
2008
).
25.
H. L.
Lu
,
X. L.
Wang
,
M.
Sugiyama
, and
Y.
Shimogaki
,
Appl. Phys. Lett.
95
,
212102
(
2009
).
26.
M.
Zhu
,
C. H.
Tung
, and
Y. C.
Yeo
,
Appl. Phys. Lett.
89
,
202903
(
2006
).
27.
I.
Krylov
,
A.
Gavrilov
,
M.
Eizenberg
, and
D.
Ritter
,
Appl. Phys. Lett.
101
,
063504
(
2012
).
28.
L.
Lin
and
J.
Robertson
,
Appl. Phys. Lett.
98
,
082903
(
2011
).
29.
Y. C.
Chang
,
M. L.
Huang
,
K. Y.
Lee
,
Y. J.
Lee
,
T. D.
Lin
,
M.
Hong
,
J.
Kwo
,
T. S.
Lay
,
C. C.
Liao
, and
K. Y.
Cheng
,
Appl. Phys. Lett.
92
,
072901
(
2008
).
30.
H. D.
Lee
,
T.
Feng
,
L.
Yu
,
D.
Mastrogiovanni
,
A.
Wan
,
T.
Gustafsson
, and
E.
Garfunkel
,
Appl. Phys. Lett.
94
,
222108
(
2009
).
31.
B.
Shin
,
D.
Choi
,
J. S.
Harris
, and
P. C.
McIntyre
,
Appl. Phys. Lett.
93
,
052911
(
2008
).
32.
J.
Robertson
and
B.
Falabretti
,
J. Appl. Phys.
100
,
014111
(
2006
).
33.
G.
He
,
L. D.
Zhang
, and
M.
Liu
,
Appl. Phys. Lett.
95
,
112905
(
2009
).
34.
See supplementary material at http://dx.doi.org/10.1063/1.4808243 for the current paper reports the reduction and removal of surface native oxide from as-received InGaAs surface by using dimethylaluminumhydride-derived aluminum oxynitride (AlON) passivation layer prior to HfTiO deposition to solve Fermi level pinning issue. It has been found that complete consumption of native oxides of AsOx and GaOx at the InGaAs surface, but no effect to InOx, has been realized after MOCVD AlON at 300 °C. XPS observations of HfTiO/InGaAs gate stacks demonstrate that introducing AlON layer can suppress the regrowth of native oxide at the interface.

Supplementary Material