In this work, we demonstrate a hydrogen-terminated diamond (H-diamond) field-effect transistor (FET) with Al2O3/CeB6 gate materials. The CeB6 and Al2O3 films have been deposited by electron beam evaporation technique, sequentially. For the 4/8/12/15 μm gate length (LG) devices, the whole devices demonstrate distinct p-type normally off characteristics, and all the threshold voltage are negative; all the absolute values of leakage current density are 10−4 A/cm2 at a VGS of −11 V, exhibiting a relatively low leakage current density compared with CeB6 FETs, and this further demonstrates the feasibility of the introduction of Al2O3 to reduce the leakage current density; the maximum drain–source current density is −114.6, −96.0, −80.9, and −73.7 mA/mm, which may be benefited from the well-protected channel. For the 12 μm LG devices, the saturation carrier mobility is 593.6 cm2/V s, demonstrating a good channel transport characteristic. This work may provide a promising strategy for the application of normally off H-diamond FETs significantly.

As an ultra-wide bandgap semiconductor, diamond exhibits the advantages of ultrahigh breakdown field strength, high carrier mobility, exceptionally high thermal conductivity, good corrosion and radiation resistance, etc.1–3 These excellent characteristics enable diamond-based electronics devices to work safely and stably in the extreme environment, such as high frequency, high pressure, high temperature, and strong radiation. Nevertheless, the development of diamond-based electronic devices has been greatly hampered by the traditional doping technique due to the high activation energies of the dopants (boron of 370 meV and phosphorus of 650 meV).4 Fortunately, under the condition of C–H bond and negative adsorbates [such as O2, O2(H2O)n, etc.], hydrogen-terminated diamond (H-diamond) with two-dimensional hole gas (2DHG) accumulation layer comes into view, demonstrating a carrier density of 1012–1014 cm−2 and a carrier mobility of 10–300 cm2/V s.5–8 To date, researchers have made significant progress on H-diamond field-effect transistors (FETs), such as carrier mobility of 680 cm2/V s,1 current density of 1.3 A/mm,9 power density of 4.2 W/mm,10 maximum oscillation frequency of 120 GHz,11 cut-off frequency of 70 GHz,12 and breakdown voltage of V (Ref. 13).

For the application of H-diamond FETs, energy saving and safety protection require the normally off operation. As presented in our previous article, a low work function material is considered to be an effective technique to realize normally off H-diamond FETs.14 In addition, the preliminary work has fully verified the feasibility of normally off H-diamond FETs with CeB6 material, which exhibits the advantages of low work function, good chemical stability, high melting point, etc.15 The device demonstrates good electrical characteristics, yet the absolute value of the leakage current density (J) is large. In order to reduce the J, Al2O3 is a good candidate, which demonstrates a large valence band offset with H-diamond.16–19 Thus, the Al2O3/CeB6 gate material has been utilized in this work. To the best of the author's knowledge, few reports have been made on H-diamond FET with Al2O3/CeB6 gate material.

In this work, the fabrication of the Al/Al2O3/CeB6 H-diamond FET has been performed, and its electrical characteristics have been evaluated.

The fabrication process, cross-sectional schematic, band diagram of gate voltage (VGS) = 0 V, band diagram of |VGS| > |threshold voltage (VTH)|, and device top view of the Al/Al2O3/CeB6 H-diamond FET are demonstrated in Fig. 1. In this experiment, a (100) high pressure high temperature (HPHT) single crystal diamond with dimensions of 3 × 3 × 0.5 mm3 was utilized as the substrate. First, before the epitaxial layer growth, the substrate was cleaned with mixed acid solutions of H2SO4:HNO3 = 1:1 at 250 °C for 1h to remove the non-diamond contaminates, and then, the substrate was immersed in acetone, alcohol, and de-ionized water sequentially for ultrasonic cleaning, which was dried by nitrogen flow. Second, a 200 nm undoped epitaxial layer was grown on the cleaned substrate by microwave plasma chemical vapor deposition (MPCVD) technique with a gas flow, CH4/H2 ratio, temperature, pressure, and power of 500 SCCM, 1%, 900 °C, 100 Torr, and 1 kW, respectively. Furthermore, the CH4 flow was set to zero, and the substrate was irradiated by hydrogen plasma for the formation of an H-diamond surface with 2DHG conduction channel for at least 20 min.15 Third, 150 nm Au was adopted as the source/drain ohmic contact electrodes fabricated by photo-lithography, electron beam evaporation (EB), and standard lift-off technique. Fourth, the nonactive region of the substrate was oxidized by 15 min ultraviolet ozone treatment with UV wavelengths of 254 and 185 nm. Finally, 150/30/30 nm Al/Al2O3/CeB6 gate electrodes were fabricated by photo-lithography, EB, and lift-off techniques. The dimensions of the devices are the gate length (LG) of 4/8/12/15 μm, gate width (WG) of 100 μm, and source–drain length (LSD) of 20 μm. The characteristics were characterized by a semiconductor analyzer Agilent B1505 A at room temperature.

FIG. 1.

Al/Al2O3/CeB6 H-diamond FET: (a) fabrication process, (b) cross-sectional schematic, (c) band diagram of VGS = 0 V, (d) band diagram of |VGS| > |VTH|, (e) device top view.

FIG. 1.

Al/Al2O3/CeB6 H-diamond FET: (a) fabrication process, (b) cross-sectional schematic, (c) band diagram of VGS = 0 V, (d) band diagram of |VGS| > |VTH|, (e) device top view.

Close modal

The H-diamond has been characterized by x-ray diffraction (XRD) technique, as demonstrated in Fig. 2. The full width at half maximum (FWHM) is 33.4 arcsec, indicating a high quality of H-diamond.

FIG. 2.

XRD result of H-diamond.

FIG. 2.

XRD result of H-diamond.

Close modal

In addition, CeB6 and Al2O3 have been characterized by atomic force microscopy (AFM) and x-ray photoelectron spectroscopy (XPS) techniques, respectively. In Fig. 3, the root mean square roughness is 1.10 and 1.36 nm for the Al2O3 and CeB6 materials with a scanning size of 5 × 5 μm2, illustrating an adequate morphology for device fabrication. The XPS spectra of Al2O3 are demonstrated in Figs. 4(a) and 4(b). The Al 2p spectrum located at 73.9 eV corresponds to Al–O. And, the O 1s spectrum has been fitted with the Gauss function. The peak located at 530.59 eV belongs to C–O, and the peak located at 531.86 eV corresponds to Al–O. The XPS spectra of CeB6 are demonstrated in Figs. 4(c) and 4(d). There are two peaks located at 886.0 and 904.3 eV for the Ce 3d spectrum and also two peaks of 187.9 and 191.8 eV for B 1s spectrum.

FIG. 3.

AFM topography image with a scanning size of 5 × 5 μm2: (a) Al2O3 and (b) CeB6.

FIG. 3.

AFM topography image with a scanning size of 5 × 5 μm2: (a) Al2O3 and (b) CeB6.

Close modal
FIG. 4.

High resolution XPS spectra of (a) Al 2p and (b) O 1s obtained from the Al2O3, (c) Ce 3d, and (d) B 1s obtained from the CeB6.

FIG. 4.

High resolution XPS spectra of (a) Al 2p and (b) O 1s obtained from the Al2O3, (c) Ce 3d, and (d) B 1s obtained from the CeB6.

Close modal

Furthermore, as shown in Fig. 5, the interface property has been characterized by transmission electron microscopy (TEM) technique, and the composition of the cross section of the Al/Al2O3/CeB6/H-diamond has also been characterized by energy dispersive x-ray analysis (EDX) technique. The TEM result demonstrates a good interface and a uniform distribution of Al, Ce, and B.

FIG. 5.

TEM result of Al/Al2O3/CeB6 H-diamond: (a)–(c) cross-sectional image with sizes of 10, 2, and 20 nm, (d)–(f) EDX elemental mapping of Al, Ce, and B, respectively.

FIG. 5.

TEM result of Al/Al2O3/CeB6 H-diamond: (a)–(c) cross-sectional image with sizes of 10, 2, and 20 nm, (d)–(f) EDX elemental mapping of Al, Ce, and B, respectively.

Close modal

The VTH and transconductance (Gm) characteristics of Al/Al2O3/CeB6 H-diamond FET are demonstrated in Fig. 6. For the 4/8/12/15 μm LG devices, the VTH are −0.4, −0.5, −0.4, and −0.5 V extracted from the curve of drain–source current (IDS) and VGS,18 demonstrating normally off characteristics for the whole devices. The experimental results further verify the effectiveness of the low work function material to realize normally off H-diamond FETs. In addition, Gm are slightly higher than our previous work with values of 11.4, 10.7, 10.1, and 8.1 mS/mm,4,14 indicating a comparatively good control of VGS on the IDS.

FIG. 6.

Transfer characteristics (|IDS|1/2/GmVGS) of the Al/Al2O3/CeB6 H-diamond FET (a)–(d) with LG of 4, 8, 12, and 15 μm, respectively.

FIG. 6.

Transfer characteristics (|IDS|1/2/GmVGS) of the Al/Al2O3/CeB6 H-diamond FET (a)–(d) with LG of 4, 8, 12, and 15 μm, respectively.

Close modal

As presented in Fig. 7, distinct saturation and pinch-off characteristics are observed, and |IDS| increases with the increased |VGS|, indicating a p-type channel with hole carriers under the Al/Al2O3/CeB6 materials. The maximum IDS (IDSmax) are −114.6, −96.0, −80.9, and −73.7 mA/mm at a drain–source voltage (VDS) of −20 V and VGS of −11 V for the 4/8/12/15 μm LG devices. Obviously, |IDSmax| decreases with an increase in LG. In addition, the relatively high IDSmax can be attributed to the well-protected conduction channel, which is not degraded remarkably by EB technique.

FIG. 7.

Output characteristics (IDSVDS) of the Al/Al2O3/CeB6 H-diamond FET (a)–(d) with LG of 4, 8, 12, and 15 μm, respectively.

FIG. 7.

Output characteristics (IDSVDS) of the Al/Al2O3/CeB6 H-diamond FET (a)–(d) with LG of 4, 8, 12, and 15 μm, respectively.

Close modal

Figure 8 presents the J of the Al/Al2O3/CeB6 H-diamond FET with LG of 4, 8, 12, and 15 μm, respectively. At the VGS of 0 V, all the J values are as low as 10−7 A/cm2. The J demonstrates a relatively low value of 1.2 × 10−4, 2.0 × 10−4, 2.2 × 10−4, and 4.6 × 10−4 A/cm2 at a VGS of −11 V, and this result shows that the leakage current density increases with the increased gate length, yet all the values are around 10−4 A/cm2. Accordingly, the Al2O3/CeB6 gate dielectric is not excellent, but it seems to be ok. In addition, the reason behind the increased leakage current density may be ascribed to the uneven Al2O3/CeB6 gate dielectric. Compared with our previous work, the J of the 8 μm LG CeB6 FET is 2.4 × 10−3 A/cm2 @VGS = −8 V,15 and the J of the 8 μm LG Al2O3/CeB6 FET is 3.3 × 10−5 A/cm2 @VGS = −8 V. Even when the VGS increases to −11 V, the J of the 8 μm LG Al2O3/CeB6 FET is 2.0 × 10−4 A/cm2. All this demonstrates that the introduction of Al2O3, indeed, have an effect on reducing the leakage current density.15 

FIG. 8.

J characteristics (JVGS) of the Al/Al2O3/CeB6 H-diamond FET (a)–(d) with LG of 4, 8, 12, and 15 μm, respectively.

FIG. 8.

J characteristics (JVGS) of the Al/Al2O3/CeB6 H-diamond FET (a)–(d) with LG of 4, 8, 12, and 15 μm, respectively.

Close modal
As shown in Fig. 9, the on/off ratio, subthreshold swing (SS), the capacitance (CGS), the flatband voltage (VFB), the carrier density (ρ) and the saturation carrier mobility (μsat) characteristics of the 12 μm LG devices are evaluated. The on/off ratio reaches up to around 109, which is high enough for practical applications. The SS is extracted to be 120 mV/dec, demonstrating a relatively fast transition rate between on and off states. The CGS is 0.06 μF/cm2, indicating that the quality of the CeB6/Al2O3 material is not excellent, which may be improved in our future work. The VFB is calculated to be −1.69 V based on the relationship of d 2 C GS / d 2 V GS = 0.20 Furthermore, the CGSVGS curve shifts to the negative direction corresponding to the position of VGS = 0 V, indicating the existence of a fixed positive charge (Qf) in the CeB6/Al2O3 film.21 In addition, it is calculated to be 4.0 × 1011 cm−2 based on formula (1).21 Here, Δ W means the work function difference between H-diamond (4.9 eV) and Al (4.28 eV), and q is the electronic charge of 1.6 × 10−19 C. Furthermore, the ρ is deduced to be 4.1 × 1012 cm−2 obtained at a VGS of −10 V based on C d V GS.20 Moreover, μsat is also investigated to evaluate the channel transport characteristics based on expression (2).3 The μsat of the CeB6/Al2O3 H-diamond FET is 593.6 cm2/V s at a VGS of −3 V,
I DSmax = W G 2 L G C GS μ sat ( V GS V TH ) 2 ,
(1)
Q f = V FB Δ W / q q C GS .
(2)
FIG. 9.

Electrical properties of the Al/Al2O3/CeB6 H-diamond FET with a LG of 12 μm (a) |IDS| – VGS, (b) CGS – VGS, (c) ρ – VGS, (d) μsat – VGS.

FIG. 9.

Electrical properties of the Al/Al2O3/CeB6 H-diamond FET with a LG of 12 μm (a) |IDS| – VGS, (b) CGS – VGS, (c) ρ – VGS, (d) μsat – VGS.

Close modal
In addition, to evaluate the interface characteristics of the device, the interface state density (Dit) is calculated to be 5.9 × 1012 cm−2 eV−1 based on the following expression:22,
S S = ( ln 10 ) k T q C GS + C D + q D it C GS ,
(3)
where k is the Boltzmann constant and CD is the depletion capacitance that may be negligible since its value is much smaller than CGS.22 

Table I demonstrates the electrical properties’ comparison with the reported H-diamond FETs. The hBN FET demonstrates excellent performances with a μsat of 680 cm2/V s, yet the fabrication process is complicated.1 The J of the Al2O3 FET is 4.46 × 10−6 A/cm2 @VGS = −4 V,23 and for the Al2O3/CeB6 FET in this work, J is 3.6 × 10−6 A/cm2 @VGS = −4 V. So, the two devices demonstrate similar J at the same VGS. Compared with the CeB6 FET, the J of the Al2O3/CeB6 FET demonstrates competitive J. Dit for the hBN, CeB6, and CeB6/Al2O3 gate H-diamond FET are 6.8 × 1011, 1.93 × 1012, and 5.9 × 1012 cm2 eV1, respectively. The Dit of hBN is low compared with other FETs, and this is a key factor for the high performance with a μsat of 680 cm2/V s. Dit for the CeB6 and Al2O3/CeB6 FETs are all orders of magnitude 1012.

TABLE I

Electrical properties’ comparison with the reported H-diamond FETs.

Gate materialshBNAl2O3CeB6Al2O3/CeB6
LG (μm) 8.09 40 12 
VTH (V) −0.99 … −0.46 −0.4 
IDSmax (mA/mm) −200 … −83.8 −80.9 
J (A/cm23 × 10−7 4.46 × 10−6 2.4 × 10−3 2.2 × 10−4 
@VGS = −4 V @VGS = −8 V @VGS = −11 V 
ρ (cm−26.6 × 1012 9.4 × 1012 1.19 × 1013 4.1 × 1012 
μsat (cm2/V s) 680 94.2 260.5 593.6 
Dit (cm−2 eV−16.8 × 1011 … 1.93 × 1012 5.9 × 1012 
Ref. 1  23  15  This work 
Gate materialshBNAl2O3CeB6Al2O3/CeB6
LG (μm) 8.09 40 12 
VTH (V) −0.99 … −0.46 −0.4 
IDSmax (mA/mm) −200 … −83.8 −80.9 
J (A/cm23 × 10−7 4.46 × 10−6 2.4 × 10−3 2.2 × 10−4 
@VGS = −4 V @VGS = −8 V @VGS = −11 V 
ρ (cm−26.6 × 1012 9.4 × 1012 1.19 × 1013 4.1 × 1012 
μsat (cm2/V s) 680 94.2 260.5 593.6 
Dit (cm−2 eV−16.8 × 1011 … 1.93 × 1012 5.9 × 1012 
Ref. 1  23  15  This work 

In conclusion, the H-diamond FET with Al2O3/CeB6 gate materials has been fabricated and characterized. For the 4/8/12/15 μm LG devices, all the devices demonstrate good electrical characteristics with IDSmax of −114.6, −96.0, −80.9, and −73.7 mA/mm at a VDS of −20 V and a VGS of −11 V. VTH are extracted to be −0.4, −0.5, −0.4, and −0.5 V, and this further verifies the effectiveness of the low work function material to realize normally off H-diamond FETs. All the J values are as low as 10−7 A/cm2 at a VGS of 0 V and 10−4 A/cm2 at a VGS of −11 V. For the 12 μm LG devices, it exhibits a high on/off ratio of 109 and a large μsat of 593.6 cm2/V s. In future work, the electrical properties will be further enhanced by optimizing the fabrication process, and this may significantly promote the application of normally off H-diamond FETs.

This work was supported by the National Key Research and Development Program of China (No. 2022YFB3608603), the China Postdoctoral Science Foundation (No. 2022M712516), the Natural Science Basic Research Program of Shaanxi Province (No. 2023-JC-QN-0718), and the National Natural Science Foundation of China (NNSFC) (Nos. U21A2073, 62074127, 62304173, and 62304175).

The authors have no conflicts to disclose.

Zhang Minghui: Conceptualization (lead); Data curation (lead); Methodology (lead); Software (lead); Validation (lead); Writing – original draft (lead). Wang Wei: Supervision (equal); Writing – review & editing (equal). Chen Genqiang: Formal analysis (lead); Methodology (supporting). Wen Feng: Resources (lead). Lin Fang: Investigation (lead). Wang Yanfeng: Software (supporting). Zhang Pengfei: Data curation (supporting). Wang Fei: Validation (supporting). He Shi: Formal analysis (supporting). Liang Yuesong: Formal analysis (equal). Fan Shuwei: Supervision (supporting). Wang Kaiyue: Investigation (supporting). Yu Cui: Project administration (equal). Min Tai: Project administration (equal). Wang Hongxing: Funding acquisition (lead); Supervision (supporting); Writing – review & editing (supporting).

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
Y.
Sasama
,
T.
Kageura
,
M.
Imura
,
K.
Watanabe
,
T.
Taniguchi
,
T.
Uchihashi
, and
Y.
Takahide
,
Nat. Electron.
5
,
37
(
2022
).
2.
J.
Isberg
,
J.
Hammersberg
,
E.
Johansson
,
T.
Wikström
,
D. J.
Twitchen
,
A. J.
Whitehead
,
S. E.
Coe
, and
G. A.
Scarsbrook
,
Science
297
,
1670
(
2002
).
3.
M.
Liao
,
L.
Sang
,
T.
Shimaoka
,
M.
Imura
,
S.
Koizumi
, and
Y.
Koide
,
Adv. Electron. Mater.
5
, 1800832 (
2019
).
4.
Z.
Minghui
,
W.
Wei
,
W.
Feng
,
L.
Fang
,
C.
Genqiang
,
W.
Fei
,
H.
Shi
,
W.
Yanfeng
,
F.
Shuwei
,
B.
Renan
,
M.
Tai
,
Y.
Cui
, and
W.
Hongxing
,
Funct. Diamond
2
,
258
(
2022
).
6.
K.
Hirama
,
H.
Takayanagi
,
S.
Yamauchi
,
Y.
Jingu
,
H.
Umezawa
, and
H.
Kawarada
, in
2007 IEEE International Electron Devices Meeting
(
IEEE
,
New York
,
2007
), Vols. 1, 2, p.
873+
.
7.
K. G.
Crawford
,
I.
Maini
,
D. A.
Macdonald
, and
D. A. J.
Moran
,
Prog. Surf. Sci.
96
,
100613
(
2021
).
8.
Z.
Ren
,
S.
Ding
,
Z.
Liang
,
Q.
He
,
K.
Su
,
J.
Zhang
,
J.
Zhang
,
C.
Zhang
, and
Y.
Hao
,
Appl. Phys. Lett.
120
, 042104 (
2022
).
9.
K.
Hirama
,
H.
Sato
,
Y.
Harada
,
H.
Yamamoto
, and
M.
Kasu
,
Jpn. J. Appl. Phys.
51
,
090114
(
2012
).
10.
C.
Yu
,
C.
Zhou
,
J.
Guo
,
Z.
He
,
M.
Ma
,
H.
Yu
,
X.
Song
,
A.
Bu
, and
Z.
Feng
,
Funct. Diamond
2
,
64
(
2022
).
11.
M.
Kasu
,
K.
Ueda
,
H.
Kageshima
, and
Y.
Taniyasu
, in
Physics Status Solidi C—Currents Top Solid State Physics
, edited by
Y
.
Hirayama
and
T
.
Sogawa
(
Wiley-V C H Verlag Gmbh
,
Weinheim
,
2008
), Vol. 5, No. 9, pp.
3165
3168
.
12.
X.
Yu
,
J.
Zhou
,
C.
Qi
,
Z.
Cao
,
Y.
Kong
, and
T.
Chen
,
IEEE Electron Device Lett.
39
,
1373
(
2018
).
13.
N. C.
Saha
,
S. W.
Kim
,
K.
Koyama
,
T.
Oishi
, and
M.
Kasu
,
IEEE Electron Device Lett.
44
10
(
2022
).
14.
M.
Zhang
,
W.
Wang
,
G.
Chen
,
H. N.
Abbasi
,
F.
Lin
,
F.
Wen
,
K.
Wang
,
J.
Zhang
,
R.
Bu
, and
H.
Wang
,
Appl. Phys. Lett.
118
,
5
(
2021
).
15.
M.
Zhang
,
W.
Wang
,
G.
Chen
,
F.
Wen
,
F.
Lin
,
S.
He
,
Y.
Wang
,
L.
Zhang
,
S.
Fan
,
R.
Bu
,
T.
Min
,
C.
Yu
, and
H.
Wang
,
Carbon N. Y.
201
,
71
(
2023
).
16.
S.
He
,
Y. F.
Wang
,
G.
Chen
,
M.
Zhang
,
W.
Wang
,
X.
Chang
,
Q.
Li
,
Q.
Zhang
,
T.
Zhu
, and
H. X.
Wang
,
Diamond Relat. Mater.
120
,
6
(
2021
).
17.
R. G.
Banal
,
M.
Imura
,
J.
Liu
, and
Y.
Koide
,
J. Appl. Phys.
120
, 115307 (
2016
).
18.
J. W.
Liu
,
M. Y.
Liao
,
M.
Imura
, and
Y.
Koide
,
Appl. Phys. Lett.
103
, 092905 (
2013
).
19.
J.
Liu
,
M.
Liao
,
M.
Imura
,
A.
Tanaka
,
H.
Iwai
, and
Y.
Koide
,
Sci. Rep.
4
,
2
(
2014
).
20.
J. F.
Zhang
,
W. J.
Chen
,
Z. Y.
Ren
,
K.
Su
,
P. Z.
Yang
,
Z. Z.
Hu
,
J. C.
Zhang
, and
Y.
Hao
,
Phys. Status Solidi
217
,
1
(
2020
).
21.
J. W.
Liu
,
H.
Oosato
,
M. Y.
Liao
, and
Y.
Koide
,
Appl. Phys. Lett.
110
, 203502 (
2017
).
22.
T.
Matsumoto
,
H.
Kato
,
K.
Oyama
,
T.
Makino
,
M.
Ogura
,
D.
Takeuchi
,
T.
Inokuma
,
N.
Tokuda
, and
S.
Yamasaki
,
Sci. Rep.
6
, 31585 (
2016
).
23.
Z.
Ren
,
W.
Chen
,
J.
Zhang
,
J.
Zhang
,
C.
Zhang
,
G.
Yuan
,
K.
Su
,
Z.
Lin
, and
Y.
Hao
,
IEEE J. Electron Devices Soc.
7
,
88
(
2019
).