This work demonstrates the advantage of carrying out silicon ion (Si+) implantation at high temperatures for forming controlled heavily doped regions in gallium oxide. Room temperature (RT, 25 °C) and high temperature (HT, 600 °C) Si implants were carried out into MBE grown (010) β-Ga2O3 films to form ∼350 nm deep Si-doped layers with average concentrations up to ∼1.2 × 1020 cm−3. For such high concentrations, the RT sample was too resistive for measurement, but the HT samples had 82.1% Si dopant activation efficiency with a high sheet electron concentration of 3.3 × 1015 cm−2 and an excellent mobility of 92.8 cm2/V·s at room temperature. X-ray diffraction measurements indicate that HT implantation prevents the formation of other Ga2O3 phases and results in reduced structural defects and lattice damage. These results are highly encouraging for achieving ultra-low resistance heavily doped Ga2O3 layers using ion implantation.

Ga2O3 is emerging as a promising wide bandgap semiconductor material and is attracting significant attention from the research community for power electronic applications.1–3 Among different phases, the monoclinic β phase is thermally the most stable with a reported bandgap of 4.6–4.9 eV,4 and a breakdown electric field of 8 MeV/cm,5 which is much higher compared to SiC and GaN. β-Ga2O3 has robust chemical resistance and radiation hardness, which are attractive for devices operating in harsh environments.5 In addition, the prospect of low-cost mass production due to a floating zone (FZ)6 and an edge-defined film-fed (EFG)6 growth method is also a key advantage for this material.

Ion implantation is attractive for device processing as it offers a way to selectively dope different regions of a material with precise control of dopant concentration and profile.5 Sasaki et al.7 performed Si implantation into β-Ga2O3 (010) at room temperature (RT) with concentrations ranging from 1019 to 1020 cm−3 followed by annealing ranging from 700 to 1100 °C. The highest carrier concentration and lowest resistivity were achieved for samples implanted with 5 × 1019 cm−3 and annealed at 1000 °C. The ratio of the free electron concentration to the total amount of Si in Ga2O3 (activation efficiency, η) was found to be 63% for 5 × 1019 cm−3 implant concentration. However, for a concentration of 1020 cm−3, a severe drop in η to ∼6% was reported, which resulted in a resistivity even higher than samples with the lowest implanted concentration of 1 × 1019 cm−3. These experiments seemed to define a maximum effective implanted concentration of 5 × 1019 cm−3 for room temperature implanted Si into β-Ga2O3. In the case of 4H-SiC, high temperature (HT) Al+ implantation results in a dramatic improvement in activation efficiency and conductivity over room temperature implantation due to lower defect densities and better recrystallization after annealing.8 Also, improvement in electrical activation by high temperature ion implantation in Si is a well-established phenomenon.9 In this work, it is demonstrated that ion implantation at elevated temperatures can be used to efficiently dope β-Ga2O3 as well.

To this end, 300 nm thick, unintentionally doped β-Ga2O3 films are grown by molecular beam epitaxy (MBE) on top of the Fe doped semi-insulating substrate using the process described in the supplementary material. Based on previous MBE runs, the doping and mobility of the film are estimated to be ∼2 × 1016 cm−3 and 100 cm2/V·s, respectively, giving an approximate sheet resistance of 1 × 105 Ω/□. Samples were implanted with 275 and 425 keV Si+ ions targeting a box profile approximately 300 nm thick as shown in the SRIM profile in Fig. 1. The SRIM10 simulated profile for Si is shown in Fig. 1 that gave idea about two energies of implantation based on the targeted implant profile. The ion flux for the lower and higher energy was 4.61 × 1011 and 6.46 × 1011 ions/cm2/s, respectively, maintaining identical beam current (1 μA for both energies with a chamber pressure of 3.0 × 10−6 Torr) for all the implants with total fluences of 2.4 × 1015 or 4.8 × 1015 cm−2. To overcome the limitation of the lowest available Si+ beam energy in the pelletron accelerator that was used (the lowest possible energy is ∼100 KeV), 110 nm of Mo was used as an energy reducing layer to enable doping of the film with the target SRIM profile. A 30 nm thick Al2O3 layer was used between Mo and β-Ga2O3 to prevent any Mo knock-ons into β-Ga2O3. However, Al atoms get recoiled into the MBE film. Based on SRIM simulation (see supplementary material Fig. S1), this Al containing surface layer is ∼15 nm and likely to be thinner due to wet etching of Al2O3 prior to annealing. Therefore, this layer is not likely to significantly impact the bulk electron transport results of the much thicker Si doped region. The implants were done with the sample holder at 21.6 °C (RT implants) and at 600 °C (HT implants). The temperature was measured at the back of the sample holder using a thermocouple as shown in Fig S4 of the supplementary material. After implantation, Mo and Al2O3 were removed by etching in hydrogen peroxide and a commercial (1:7) buffered oxide etch, respectively. This was followed by annealing in flowing nitrogen at 970 °C for 30 min to anneal the implant damage and activate the dopants. Optical and atomic force microscopy was performed after implantation and annealing. The RMS roughness (Rq) was less than 1 nm for all samples. However, the higher fluence RT sample had long (∼mm) line-like features throughout the surface after annealing possibly due to surface reconstruction (see AFM and optical micrographs in the supplementary material) that require further investigation.

FIG. 1.

SIMS profiles of Si+ in β-Ga2O3 implanted with 275 and 425 keV Si+ ions with a total fluence of 4.8 × 1015 cm−2 for RT and HT implanted samples before and after annealing. The Si SRIM simulated profile and SIMS of Fe after annealing for RT and HT implanted samples are also shown.

FIG. 1.

SIMS profiles of Si+ in β-Ga2O3 implanted with 275 and 425 keV Si+ ions with a total fluence of 4.8 × 1015 cm−2 for RT and HT implanted samples before and after annealing. The Si SRIM simulated profile and SIMS of Fe after annealing for RT and HT implanted samples are also shown.

Close modal

Secondary ion mass spectrometry (SIMS) profiles of Si in samples implanted with a fluence of 4.8 × 1015 cm−2 at HT and RT before and after activation annealing are shown in Fig. 1. The SIMS profiles are deeper than the simulated profile possibly due to partial channeling of the ions into the crystal lattice since the samples were not tilted during implantation. The RT implanted sample prior to annealing had similar Si ion distribution to that of the as implanted HT sample. After annealing, however, the Si concentration increased in HT whereas it decreased for RT. More measurements are required to determine the reason for this or whether it is a real effect. Previously, it has been reported that ion induced damage at the film/substrate interface results in enhanced out-diffusion of Fe from the substrate into the film.11 This is observed here as well, but the Fe profiles are comparable for the RT and HT implantation after activation as shown in Fig. 1. The thickness of the conducting layer was considered as the depth at which the Si concentration falls to 1019 cm−3, where compensation effects of Fe would not affect the conductivity. The dip of the Fe profile in the MBE layer is likely due to the interplay between damage and diffusion11 during annealing, but further investigation is needed for more evidence.

Hall measurement was performed in the van der Pauw configuration in the temperature range of 27–325 °C. The contacts were made by depositing Ti/Au at four corners of the nominally square sample followed by annealing in an argon (Ar) atmosphere at 550 °C for 5 min. Multiple measurements were made, and the average values are reported here with the error bars representing the standard deviation of the measurements. Figure 2 shows that the sample implanted at high temperatures with the higher fluence has the lowest sheet resistance of 20.8 ± 0.68 Ω/□ at 27 °C. On the other hand, the higher fluence RT implanted sample was too resistive for measurement. For the lower fluence, the HT and RT samples had sheet resistances of 41.4 ± 8.24 and 51.3± 1.4, respectively, at 27 °C. At higher temperature, the overall resistance increased for all samples, but the HT implanted samples maintained advantage even at 325 °C.

FIG. 2.

Sheet resistance vs temperature for samples implanted at HT and RT with total Si+ fluences of 2.4 × 1015 and 4.8 × 1015 cm−2. Error bars represent variations over four measurements.

FIG. 2.

Sheet resistance vs temperature for samples implanted at HT and RT with total Si+ fluences of 2.4 × 1015 and 4.8 × 1015 cm−2. Error bars represent variations over four measurements.

Close modal

The average sheet electron concentrations (ns) are shown in Fig. 3. In Table I, the activation efficiencies η for all the samples are listed. Here, η was determined as the ratio of ns and the total amount of Si obtained by integrating the SIMS profiles. For the lower fluence, RT and HT implanted samples, η are equal to about 68.2% and 76.3%, respectively, a modest advantage in favor of the HT sample. Remarkably, for the higher fluence HT sample, η is ∼82.1%, which results in the increased electron sheet concentration. The likely reason for maintaining the high η is the reduction of ion induced compensating defects such as vacancies, interstitials, and/or their complexes by high temperature implantation. Figure 4 shows the electron Hall mobility as a function of temperature. For the lower fluence, mobility of HT and RT samples are 80.2 ± 1.1 and 82.8 ± 2.14 cm2/V·s, respectively, at 27 °C. For the higher fluence, the mobility increases to 92.8 ± 5.64 cm2/V·s (at 27 °C). This unusual trend of increasing mobility with higher electron concentration could be due to enhanced screening of the electron phonon-scattering by the increased carrier density12 and requires further investigation.

FIG. 3.

Sheet charge concentration vs temperature for samples implanted at HT and RT with total Si+ fluences of 2.4 × 1015 and 4.8 × 1015 cm−2. Error bars represent variations over four measurements.

FIG. 3.

Sheet charge concentration vs temperature for samples implanted at HT and RT with total Si+ fluences of 2.4 × 1015 and 4.8 × 1015 cm−2. Error bars represent variations over four measurements.

Close modal
TABLE I.

Activation efficiency (η) comparison.

Nominal fluence (/cm2)Implantation temperatureActivation efficiency (η) (%)
2.4 × 1015 RT (27 °C) 68.2 
2.4 × 1015 HT (600 °C) 76.3 
4.8 × 1015 RT (27 °C) ∼0 
4.8 × 1015 HT (600 °C) 82.1 
Nominal fluence (/cm2)Implantation temperatureActivation efficiency (η) (%)
2.4 × 1015 RT (27 °C) 68.2 
2.4 × 1015 HT (600 °C) 76.3 
4.8 × 1015 RT (27 °C) ∼0 
4.8 × 1015 HT (600 °C) 82.1 
FIG. 4.

Electron mobility vs temperature for HT and RT implanted samples' total fluences of Si+ ions of 2.4 × 1015 and 4.8 × 1015 cm−2. Error bars represent variation over four measurements.

FIG. 4.

Electron mobility vs temperature for HT and RT implanted samples' total fluences of Si+ ions of 2.4 × 1015 and 4.8 × 1015 cm−2. Error bars represent variation over four measurements.

Close modal

To study the crystal quality and strain, x-ray diffraction was performed using a Rigaku SmartLab XRD system using a Ge (220)×2 monochromator. Figure 5 shows HRXRD data for the higher fluence samples at the three different stages: as received, after Si implantation, and after activation annealing. The excellent crystal quality of the as received homoepitaxial samples is evident in the FWHM value of 52.5 arc sec corresponding to the (020) peak. After RT Si implantation [red in Fig. 5(a)], the FWHM increased to 122 arc sec consistent with ion implantation induced defect creation and strain. An additional broader peak with a FWHM of 196.5 arc sec also appeared centered at 63.3°, which indicates the formation of a different phase; it is identified in the literature as κ-Ga2O313,14 or γ-Ga2O3.15 On the other hand, no additional phase is observed for HT implantation [red in Fig. 5(b)], but broadening and distortion of the (020) peak is evident. After annealing at 970 °C, the broader peak in RT samples [blue in Fig. 5(a)] disappears, but an additional side shoulder appears at an angle higher than the original peak position along with a shift of the (020) peak, which can be attributed to the accretion of compressive strain in previous studies,13,16–18 although due to the strong anisotropy of Ga2O3, it does not necessarily hold for other orientations. The strain peak is significantly smaller in the HT samples after annealing [blue in Fig. 5(b)]. However, the ion damage is not completely recovered, possibly due to the residual strain19 suggesting optimization of the process is further needed. For the lower dose samples, similar effects were observed, and the results are included in the supplementary material.

FIG. 5.

HRXRD data for Ga2O3 implanted with 4.8 × 1015 cm−2 Si+ at (a) RT and (b) HT before and after annealing compared to the as-received MBE sample spectrum.

FIG. 5.

HRXRD data for Ga2O3 implanted with 4.8 × 1015 cm−2 Si+ at (a) RT and (b) HT before and after annealing compared to the as-received MBE sample spectrum.

Close modal

In conclusion, we have demonstrated the advantage of HT ion implantation for forming heavily doped (>1020 cm−3) n-type regions in β-Ga2O3 devices and resistivity as low as 0.68 mΩ cm. Using Si+ as the dopant, we observe a substantial increase in the sheet electron concentration and, consequently, a significantly lower sheet resistance compared to room temperature implantation. In addition, HRXRD shows that HT implantation causes reduced structural defects and strain in the implanted layer by preventing formation of any other phase. Being able to maintain the high mobility for carrier concentrations above 1020 cm−3 due to the impressive dopant activation efficiency makes it pertinent to further investigate the limits of high temperature ion implantation in β-Ga2O3.

See the supplementary material for the details of the MBE growth process, AFM, and XRD for lower fluence samples.

The authors thank Mr. Max Cichon, Accelerator Laboratory, Department of Physics, Auburn University for support with these experiments. HRXRD was performed with a Rigaku SmartLab instrument purchased with support from the National Science Foundation Major Research Instrumentation Program through Grant No. NSF-DMR-2018794. R.B.C. gratefully acknowledges funding from the Air Force Office of Scientific Research under Award No. FA9550-20-1-0034. EAG, Inc. provided SIMS services.

The authors have no conflicts to disclose.

Arka Sardar: Investigation (equal); Writing – original draft (equal). Tamara Isaacs-Smith: Investigation (equal); Writing – review & editing (equal). Jacob Lawson: Investigation (equal); Writing – review & editing (equal). Thaddeus J. Asel: Investigation (equal); Writing – review & editing (equal). Ryan Comes: Investigation (equal); Writing – review & editing (equal). Joseph Neil Merrett: Conceptualization (equal); Resources (equal); Writing – review & editing (equal). Sarit Dhar: Investigation (equal); Writing – review & editing (equal).

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

1.
M.
Higashiwaki
and
G. H.
Jessen
,
Appl. Phys. Lett.
112
,
060401
(
2018
).
2.
S. J.
Pearton
,
J.
Yang
,
P. H.
Cary
,
F.
Ren
,
J.
Kim
,
M. J.
Tadjer
, and
M. A.
Mastro
,
Appl. Phys. Rev.
5
,
011301
(
2018
).
3.
A. J.
Green
,
J.
Speck
,
G.
Xing
,
P.
Moens
,
F.
Allerstam
,
K.
Gumaelius
,
T.
Neyer
,
A.
Arias-Purdue
,
V.
Mehrotra
,
A.
Kuramata
,
K.
Sasaki
,
S.
Watanabe
,
K.
Koshi
,
J.
Blevins
,
O.
Bierwagen
,
S.
Krishnamoorthy
,
K.
Leedy
,
A. R.
Arehart
,
A. T.
Neal
,
S.
Mou
,
S. A.
Ringel
,
A.
Kumar
,
A.
Sharma
,
K.
Ghosh
,
U.
Singisetti
,
W.
Li
,
K.
Chabak
,
K.
Liddy
,
A.
Islam
,
S.
Rajan
,
S.
Graham
,
S.
Choi
,
Z.
Cheng
, and
M.
Higashiwaki
,
APL Mater.
10
,
029201
(
2022
).
4.
M. J.
Tadjer
,
J. L.
Lyons
,
N.
Nepal
,
J. A.
Freitas
,
A. D.
Koehler
, and
G. M.
Foster
,
ECS J. Solid State Sci. Technol.
8
,
Q3187
(
2019
).
5.
A.
Nikolskaya
,
E.
Okulich
,
D.
Korolev
,
A.
Stepanov
,
D.
Nikolichev
,
A.
Mikhaylov
,
D.
Tetelbaum
,
A.
Almaev
,
C. A.
Bolzan
,
A.
Buaczik
,
R.
Giulian
,
P. L.
Grande
,
A.
Kumar
,
M.
Kumar
, and
D.
Gogova
,
J. Vacuum Sci. Technol. A
39
,
030802
(
2021
).
6.
H.
Aida
,
K.
Nishiguchi
,
H.
Takeda
,
N.
Aota
,
K.
Sunakawa
, and
Y.
Yaguchi
,
Jpn. J. Appl. Phys.
47
,
8506
(
2008
).
7.
K.
Sasaki
,
M.
Higashiwaki
,
A.
Kuramata
,
T.
Masui
, and
S.
Yamakoshi
,
Appl. Phys. Express
6
,
086502
(
2013
).
8.
F.
Roccaforte
,
F.
Giannazzo
, and
G.
Greco
,
Micro
2
,
23
(
2022
).
9.
G. A.
Kachurin
,
I. E.
Tyschenko
, and
L. I.
Fedina
,
Nucl. Instrum. Methods Phys. Res. Sect. B: Beam Interact. Mater. Atoms
68
,
323
(
1992
).
10.
J. F.
Ziegler
,
M. D.
Ziegler
, and
J. P.
Biersack
,
Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms
268
,
1818
(
2010
). Version code available from http://www.srim.org, 2013.
11.
M. H.
Wong
,
K.
Sasaki
,
A.
Kuramata
,
S.
Yamakoshi
, and
M.
Higashiwaki
,
Appl. Phys. Lett.
106
,
032105
(
2015
).
12.
A. K.
Rajapitamahuni
,
A. K.
Manjeshwar
,
A.
Kumar
,
A.
Datta
,
P.
Ranga
,
L. R.
Thoutam
,
S.
Krishnamoorthy
,
U.
Singisetti
, and
B.
Jalan
,
ACS Nano
16
,
8812
(
2022
).
13.
A.
Azarov
,
C.
Bazioti
,
V.
Venkatachalapathy
,
P.
Vajeeston
,
E.
Monakhov
, and
A.
Kuznetsov
,
Phys. Rev. Lett.
128
,
015704
(
2022
).
14.
I.
Cora
,
F.
Mezzadri
,
F.
Boschi
,
M.
Bosi
,
M.
Čaplovičová
,
G.
Calestani
,
I.
Dódony
,
B.
Pécz
, and
R.
Fornari
,
CrystEngComm
19
,
1509
(
2017
).
15.
C. S.
Chang
,
N.
Tanen
,
V.
Protasenko
,
T. J.
Asel
,
S.
Mou
,
H. G.
Xing
,
D.
Jena
, and
D. A.
Muller
,
APL Mater.
9
,
051119
(
2021
).
16.
A.
Debelle
and
A.
Declémy
,
Nucl. Instrum. Methods Phys. Res. Sect. B: Beam Interact. Mater. Atoms
268
,
1460
(
2010
).
17.
V.
Trushin
,
A.
Nikolskaya
,
D.
Korolev
,
A.
Mikhaylov
,
A.
Belov
,
E.
Pitirimova
,
D.
Pavlov
, and
D.
Tetelbaum
,
Mater. Lett.
319
,
132248
(
2022
).
18.
M. J.
Tadjer
,
C.
Fares
,
N. A.
Mahadik
,
J. A.
Freitas
,
D.
Smith
,
R.
Sharma
,
M. E.
Law
,
F.
Ren
,
S. J.
Pearton
, and
A.
Kuramata
,
ECS J. Solid State Sci. Technol.
8
,
Q3133
(
2019
).
19.
K.
Tetzner
,
A.
Thies
,
E.
Bahat Treidel
,
F.
Brunner
,
G.
Wagner
, and
J.
Würfl
,
Appl. Phys. Lett.
113
,
172104
(
2018
).

Supplementary Material