A Review of Ga2O3 Materials, Processing, and Devices

A review of Ga2O3 materials, processing, and devices
S. J. Pearton, Jiancheng Yang, Patrick H. Cary, F. Ren, Jihyun Kim, Marko J. Tadjer, and Michael A. Mastro
Citation: Applied Physics Reviews 5, 011301 (2018);

View online: https://doi.org/10.1063/1.5006941
View Table of Contents: http://aip.scitation.org/toc/are/5/1
Published by the American Institute of Physics
APPLIED PHYSICS REVIEWS 5, 011301 (2018)
APPLIED PHYSICS REVIEWS
A review of Ga2O3 materials, processing, and devices

S. J. Pearton,1,a) Jiancheng Yang,2 Patrick H. Cary IV,2 F. Ren,2 Jihyun Kim,3,a)
Marko J. Tadjer,4 and Michael A. Mastro4
1
Department of Materials Science and Engineering, University of Florida, Gainesville, Florida 32611, USA
2
Department of Chemical Engineering, University of Florida, Gainesville, Florida 32611, USA
3
Department of Chemical and Biological Engineering, Korea University, Seoul 02841, South Korea
4
US Naval Research Laboratory, Washington, DC 20375, USA
(Received 28 September 2017; accepted 30 October 2017; published online 11 January 2018)
Gallium oxide (Ga2O3) is emerging as a viable candidate for certain classes of power electronics,
solar blind UV photodetectors, solar cells, and sensors with capabilities beyond existing
technologies due to its large bandgap. It is usually reported that there are five different polymorphs
of Ga2O3, namely, the monoclinic (b-Ga2O3), rhombohedral (a), defective spinel (c), cubic (d), or
orthorhombic (e) structures. Of these, the b-polymorph is the stable form under normal conditions
and has been the most widely studied and utilized. Since melt growth techniques can be used to
grow bulk crystals of b-GaO3, the cost of producing larger area, uniform substrates is potentially
lower compared to the vapor growth techniques used to manufacture bulk crystals of GaN and SiC.
The performance of technologically important high voltage rectifiers and enhancement-mode
Metal-Oxide Field Effect Transistors benefit from the larger critical electric field of b-Ga2O3 rela-
tive to either SiC or GaN. However, the absence of clear demonstrations of p-type doping in
Ga2O3, which may be a fundamental issue resulting from the band structure, makes it very difficult
to simultaneously achieve low turn-on voltages and ultra-high breakdown. The purpose of this
review is to summarize recent advances in the growth, processing, and device performance of the
most widely studied polymorph, b-Ga2O3. The role of defects and impurities on the transport and
optical properties of bulk, epitaxial, and nanostructures material, the difficulty in p-type doping,
and the development of processing techniques like etching, contact formation, dielectrics for gate
formation, and passivation are discussed. Areas where continued development is needed to fully
exploit the properties of Ga2O3 are identified. Published by AIP Publishing.
https://doi.org/10.1063/1.5006941
TABLE OF CONTENTS PROPERTIES AND PROCESSING OF b-Ga2O3 . . . 13

DOPING AND DEFECTS IN Ga2O3 . . . . . . . . . . . . . . 16
OXIDE p-n JUNCTION HETEROSTRUCTURES. . . 19
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
THEORY OF DEFECTS IN Ga2O3 . . . . . . . . . . . . . . . 19
BASIC PROPERTIES AND APPLICATIONS OF
Ga2O3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 LOW AND HIGH FIELD TRANSPORT IN Ga2O3 . 20
BULK GROWTH TECHNOLOGY OF Ga2O3 . . . . . . 6 BAND STRUCTURE . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
PROGRESS IN METAL ORGANIC CHEMICAL EPR OF Ga2O3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
VAPOR DEPOSITION (MOCVD) GROWTH OF HYDROGEN IN Ga2O3 . . . . . . . . . . . . . . . . . . . . . . . . . 23
Ga2O3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 OHMIC CONTACTS TO Ga2O3 . . . . . . . . . . . . . . . . . . 24
HALIDE VAPOR PHASE EPITAXY (HVPE) . . . . . . 9 SCHOTTKY CONTACTS TO Ga2O3 . . . . . . . . . . . . . 26
PULSED LASER DEPOSITION (PLD) . . . . . . . . . . . . 9 WET ETCHING OF Ga2O3 . . . . . . . . . . . . . . . . . . . . . . 28
MIST-CVD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 DRY ETCHING OF Ga2O3 . . . . . . . . . . . . . . . . . . . . . . 30
MBE OF Ga2O3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 BAND ALIGNMENTS OF DIELECTRICS ON
ATOMIC LAYER EPITAXY (ALE) AND (201) Ga2O3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
ATOMIC LAYER DEPOSITION (ALD) Ga2O3 . . . . 11 RADIATION DAMAGE IN Ga2O3 . . . . . . . . . . . . . . . 34
Ga2O3 NANOSTRUCTURES. . . . . . . . . . . . . . . . . . . . . 12 THIN FILM SOLAR-BLIND UV DETECTORS . . . . 35
POWER RECTIFIERS . . . . . . . . . . . . . . . . . . . . . . . . . . 37
POWER MOSFETs AND MESFETs . . . . . . . . . . . . . . 40
a)
Authors to whom correspondence should be addressed: spear@mse.ufl.edu Ga2O3-BASED GAS SENSORS . . . . . . . . . . . . . . . . . . 42
and hyunhyun7@korea.ac.kr SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . 44
1931-9401/2018/5(1)/011301/56/$30.00 5, 011301-1 Published by AIP Publishing.

011301-2 Pearton et al. Appl. Phys. Rev. 5, 011301 (2018)
INTRODUCTION
Gallium oxide (Ga2O3) has a long history, and the phase
equilibria of the Al2O3-Ga2O3-H2O system was first reported
in 1952,1 in which the polymorphs (i.e., different forms or
crystal structures) of Ga2O3 and their regions of stability
were also identified. There are five commonly identified pol-
ymorphs of Ga2O3, labeled as a, b, c, d, and e.2–10 These are
known as corundum (a), monoclinic (b), defective spinel (c),
and orthorhombic (e), with the d phase commonly accepted as
being a form of the orthorhombic phase.1,2,7,11 Among these
different phases of Ga2O3, the orthorhombic b-gallia structure
(b-phase or b-Ga2O3) is the most stable crystal structure and
has attracted most of the recent attention. The different poly-
morphs can be either insulators or conductors, depending on
the growth conditions.7 The resulting crystals are layered
material, similar to the behavior of GaSe and GaTe. Table I
summarizes the lattice parameters, crystal structure, and some
of the bulk properties of the polymorphs of Ga2O3. The origi-
FIG. 1. Number of publications on gallium oxide since 1952. A total of
nal studies indicated that the other polymorphs of Ga2O3 con- 2492 papers which had either “Ga2O3” or “Gallium Oxide” in the title have
vert to the stable b-form with heat treatment.1 been published in the last 65 years (Data: Thomson Reuters).
In this review, we will primarily focus on the monoclinic
structured b-Ga2O3 polymorph, which is attracting interest for
power electronic devices, as well as solar-blind UV photodetec- current status of power electronic devices, solar-blind UV
tors, photocatalysts, gas sensors, solar cells, phosphors, and trans- photodetectors, and gas sensors will also be covered.
parent conducting films for electrodes on a variety of
optoelectronic devices.2,3,12–20 Ga2O3 is optically transparent to BASIC PROPERTIES AND APPLICATIONS OF Ga2O3
250 nm and is electrically conducting, making it useful as a
window on some types of optical devices. In terms of device The number of publications on Ga2O3 has accelerated in
applications, thin films of polycrystalline b-Ga2O3 containing O recent years, as is evident from Fig. 1, due to the interest in
vacancies have long been known as sensors for a variety of gases electronic and photonic devices with capabilities beyond exist-
including H2, CH4, CO, and O2 which change the electrical con- ing technologies. There has been a healthy balance of experi-
ductivity upon adsorption.21–25 In addition, the conduction elec- mental and theoretical investigations of the properties of the
tron spins in this material produces a magnetic memory effect Ga2O3 system.5,6,8–11,28,29 The monoclinic phase of b-Ga2O3 is
that is stable to above room temperature (RT).26,27 the stable one under normal conditions of temperature and pres-
The review will cover basic properties of Ga2O3, the role sure, and can be converted into other phases at higher pressures
of defects and impurities, the status of doping studies, as well or temperatures.28,29 For example, it undergoes a transition to
as developments in device fabrication processes such as etch- the hexagonal a-Ga2O3 phase at a pressure of 4.4 GPa at
ing and contacting, gate dielectrics, and passivation films. The 1000 C.30 This high pressure phase can remain as a metastable
TABLE I. Summary of the properties of Ga2O3 polymorphs.
Lattice Optical Volume

parameters Refractive dielectric expansion Bulk modulus
Polymorph (Å) index, n constant at 1200 K (300 K, GPa) Comment References
a a, b ¼ 4.98–5.04, 1.74–1.95 3.03–3.80 0.035 185 Corundum, rhombohedral Yoshioka et al.,4 Stepanov
c ¼ 13.4–13.6 structure, space group R
3c, et al.,2 and He et al.5
bandgap larger than all other
polymorphs (5.2 eV)
b A ¼ 12.12–12.34, 1.68–1.89 2.82–3.57 0.024 150 Monoclinic structure, space Kohn et al.,10 Stepanov
b ¼ 3.03–3.04, group C2/m et al.,2 and He et al.6
c ¼ 5.80–5.87
c A ¼ 8.24–8.30 Defective spinel, cubic struc- Stepanov et al.2
ture, space group Fd3m
d A ¼ 9.4–10.0 0.04 160 Possibly bixbyite. Suggested Roy et al.1, Playford et al.8
to be a nanocrystalline form
of e-Ga2O3
e A ¼ 5.06–5.12, 1.6 0.028 160 Orthorhombic structure, Yoshioka et al.4 and Kroll
b ¼ 8.69–8.79, space group Pna21 et al.7
c ¼ 9.3–9.4
phase if quenched to room temperature.5,6 The b phase is the

only stable phase up to 1800 C, while the rhombohedral
corundum a phase is metastable but can exist under ambient
conditions. The b-phase can be transformed into the a-phase
under hydrostatic pressure at higher temperatures.
Due to the relative difficulty in isolating the different phases
in pure crystalline form,31–34 much of what we know about the
properties has come from theory, especially density functional
theory (DFT). For example, Yoshioka et al.4 calculated the lat-
tice parameters, space groups, and volume expansivity normal-
ized to room temperature for the different polymorphs. The
results of these calculations are shown in Fig. 2(a). The b-phase
has the lowest volume expansion and this increases in the order
b, e, a, and d. The expansivity of the a-phase is in good agree-
ment with the reported experimental data.35 The calculated bulk
moduli for the different phases are shown in Fig. 2(b)4 and these
increase in the order b, e, d, and a below 400 K.
He et al.6 reported the structural parameters, band struc-
ture, and Debye temperature of b-Ga2O3, also calculated
from density functional theory. The band structure showed
FIG. 3. Band structure of b-Ga2O3. The Fermi energy is aligned to zero. The
the conduction band minimum (CBM) is at the zone center, kk points are C ¼ (000), A ¼ (0012), Z ¼ (12120), M ¼ (121212),
with a fairly flat valence band, as shown in Fig. 3.6,7 The L ¼ (01212), and V ¼ (0120). Reprinted with permission from Appl. Phys.
electron effective mass was calculated as 0.34m0 (where m0 Lett. 88, 261904 (2006). Copyright 2006 American Institute of Physics.6
is the free electron mass) for b-Ga2O3 and 0.28 for the high
pressure hexagonal a-phase. The relative flatness of the
valence band for b-Ga2O3 in momentum space would indi-
cate a large effective mass for holes. In fact, it is possible
that the holes are not free to move but form localized polar-
ons where the hole is localized at a lattice distortion.3,36,37
The top of the valence and the bottom of the conduction band
are made up of the anionic (O 2p states with contributions
from Ga 3d and 4s orbitals) and cationic states (Ga 4s states),
respectively.3,6,32,38–43 Since the valence-band states are
derived mainly from the O 2p orbitals and are characterized
by small dispersion, the holes have large effective masses,
and the valence band has a high density of states. Holes tend
to form localized small polarons, i.e., localized holes trapped
by local lattice distortions. It is common for density functional
calculations to produce a slight underestimation of the band
gap. There is little difference in energy between the direct and
indirect gaps for b-Ga2O3, with the direct bandgap being
4.69 eV compared to 4.66 eV for indirect value. The calcula-
tions are in good agreement with aborption44 and angle-
resolved photoemission spectroscopy measurements.45–47
The room temperature photoluminescence (PL) spectra
from Ga2O3 are typically dominated by a broad set of transi-
tions centered near 399 nm which have previously been
ascribed to oxygen-vacancy related transitions.48,49 Dong
et al.49 reported four bands in this region, with the one at
380 nm in the UV region suggested to be caused by transition
levels between the oxygen vacancy and unintended N impu-
rities. They also reported peaks centered at 416 nm, 442 nm
(both in the violet region), and 464 nm (blue region), with
all three emission peaks suggested to originate from the
electron-hole recombination formed by oxygen vacancies, or
FIG. 2. Temperature dependence of volume expansivity (a) and bulk modu-
lus (b) for a-, b-, d-, and e-Ga2O3. Reprinted with permission from Yoshioka to the recombination of Ga-O vacancy pair.50–53 The forma-
et al., J. Phys.: Condens. Matter 19, 346211 (2007). Copyright 2007 IOP.4 tion energy of the oxygen vacancies of b-Ga2O3 has been
investigated in the past years, and the results can vary with
different functional and approximation methods.50–53 There
are three types of O sites in b-Ga2O3.49 As a result, three
types of neutral oxygen vacancies exist, denoted as VOI, VOII
(both are 3-fold coordinated), and VOIII (4-fold coordinated),
respectively, as discussed in the work of Dong et al.49 They
predict from density functional calculations that oxygen
vacancies can induce absorption peaks at 3.80 eV, 3.52 eV,
and 3.37 eV for VOI, VOII, and VOIII, respectively.
Historically, Ga2O3 has shown three different groups of
emission bands, in the UV (3.2–3.6 eV), blue (2.8–3.0 eV),
and green (2.4 eV) regions, but it is fair to suggest that the
specific origins of the transitions are not finalized.2,3 Most
photoluminescence studies do not show intrinsic emission in
the deep UV (265–278 nm) but only emission in the ultra-
violet A (UVA) to visible range (350–600 nm).20 An
exception is the report of luminescence at 265 nm and
278 nm in nanowires (NWs), which correspond to the
experimentally reported absorption edges.20 Several mecha-
nisms have been invoked to explain the UVA/visible lumi-
nescence of b-Ga2O3, including the influence of native
defects and self-trapped holes (polarons).
The unit cell of the stable phase, b-Ga2O3, contains two
crystallographically different Ga atoms in the asymmetric
unit, one with tetrahedral and the other with octahedral coor-
dination geometry.54–65 The unit cell is composed of two FIG. 4. (a) b-Ga2O3 crystal structure and (b) (010) and ð
201Þ surfaces.
types of gallium ions (GaI and GaII) and three types of oxy-
gen ions (OI, OII, and OIII). This leads to an anisotropy of compression. They heated powdered samples to 2000 K at
physical, optical, and electrical,54,57,60,66–68 predicted both
30 GPa, and found that a-Ga2O3 is the most stable structure
theoretically and observed experimentally.66–68 For example,
at the high pressure.69 Furthermore, the structural transi-
the thermal conductivity in Ga2O3 shows a strong anisotropy,
tion from b-Ga2O3 to a-Ga2O3 was found to be
with the [010] direction showing a thermal conductivity 2.5
irreversible.69
times higher than that in the [100] direction,66–68 although
Hexagonal e-Ga2O3 is the second-most stable phase of
both theory and experiment suggest there is little anisotropy
Ga2O3.54,56,57,61,64 The bandgap of e-Ga2O3 is also approxi-
in electron effective mass.2,3,68 The highest thermal conduc-
mately 4.9 eV, and this phase is compatible with the common
tivity was 29 W/m K in the [010] direction. The thermal
hexagonal wide bandgap semiconductors GaN and SiC.64,65
conductivity of Sn- and Fe-doped samples was lower than
It is also predicted that e-Ga2O3 has a large spontaneous
undoped samples due to the enhanced phonon-impurity scat-
tering contribution, which reduces the thermal conductivity. polarization and therefore could produce high-density two
The OI and OIII ions lie in the (010) plane and the OII dimensional (2D) electron gases (2DEG) used for conducting
ions align along the b-axis. The formation free energies have channels in heterostructure field effect transistors.
the tendency b < e < a < d < c at low temperatures. There Experimental results using dynamic hysteresis measurements
are a number of orientations of b-Ga2O3 in common use, do indeed show the presence of ferroelectricity of e-Ga2O3,51
including the ð 201Þ, (010), and (001) planes. Figure 4(a) as well as pyroelectric properties and large polarization as
shows a schematic of the unit cell of b-Ga2O3, while Fig. calculated by the Berry-phase approach.62
4(b) shows two of the most commonly used crystal planes It is worth commenting on the practical aspects that
for device applications, namely, the (010) and ð 201Þ. The enable potential technological benefits. In material grown by
ð
201Þ and (010) surfaces differ significantly in terms of their the edge-defined film-fed growth method,70,71 the (010) ori-
dangling bond densities of oxygen, and this might be ented crystals can be cut to produce ð201Þ planes in the spe-
expected to have an effect on processes like wet etching or cific crystallographic direction. This produces wafers of
metal contact formation. usable size of a material with a large bandgap of approxi-
The a-phase has the same corundum crystal structure as mately 4.9 eV and associated large estimated critical electric
Al2O3 or sapphire, leading to the possibility of high quality field (EC) strength of 8 MV/cm.69,72–84 A useful tool for visu-
epitaxial layers of Ga2O3 on sapphire substrates. At higher alizing the different crystal structures of the polymorphs is
temperatures, the difference in free energy between the b- from Momma and Izumi.85 The fairly recent advances in
phase and the e-phase becomes smaller and the other poly- bulk crystal growth are one of the drivers of the interest
morphs find their region of stability. Yan-Mei et al.68 found in Ga2O3 for its potential in electronic power switches.86,87
that the phase transition from the monoclinic b-Ga2O3 to the The suitability of different semiconductors for this and
trigonal a-Ga2O3 occurs at around 19.2 GPa under cold related applications is assessed by the calculating various
figures-of-merit (FOM). Power switch metrics such as The main market segments for high-power, high-frequency
Baliga’s figure of merit (BFOM) estimating dc conduction transistors are defense and military applications (radar, jam-
losses and Huang’s material figure of merit (HMFOM) incor- ming, counter-measures, and guided weapons), wireless infra-
porating dynamic switching losses are functions of either structure (3G, 3Gþ, WiMAX/LTE (long term evolution) base
bandgap to the third power or bandgap to first order, respec- stations, and backhaul), and broadcast and communication sat-
tively.84–86 Several FOM are listed in Table II. The large ellites (SatCom).14 The relatively low thermal conductivity, k,
bandgap of Ga2O3 allows high temperature operation, and of Ga2O3 creates self-heating effects that must be mitigated in
the large critical field allows high voltage operation (relative order to utilize Ga2O3 in high-frequency devices.14
to maximum breakdown). This has been discussed in detail The bulk and surface properties of the different poly-
recently by Jessen et al.87 Many of the figures-of-merit morphs have been the focus of a significant amount of
involve this reverse breakdown field strength, which scales work.88–91 Bermudez88 used ab-initio theory to examine
as a power of the bandgap, typically a power in the range of the structure of the (100), (010), (001), and (101) faces of
2–2.5.87 The Johnson figure of merit (JFOM) represents the b-Ga2O3. The (101) surface, which exhibits a high energy
power-frequency product for RF amplification. The large when ideally terminated, was found to exhibit changes in
electron saturation velocity (vsat ¼ 2 107 cm/s) produces a bonding during relaxation leading to a lowering of the sur-
high current density, Imax (Imax qnsvsat, where face energy. b-Ga2O3 was found to cleave easily on the
q ¼ 1.6 1019 C, ns ¼ sheet charge density, vs ¼ electron (100) plane, consistent with experimental observations.89,90
saturation velocity), and potentially high operating frequency No intrinsic surface states near the valence band minimum
as ft ¼ vsat/Leff. To take several specific cases, the BFOM for were shown,88 although extrinsic states have been observed
Ga2O3 is a factor of 4 times larger than for GaN, while the by electron energy loss spectroscopy.91
HMFOM for Ga2O3 is comparable to GaN. Other properties of interest that have been reported
In addition to its high voltage capability, Ga2O3 is include the exciton binding energy band gap of the a-Ga2O3
attractive for its potential for low power loss during high fre- polymorph, which is a metastable phase at ambient condi-
quency switching frequency in the GHz regime.87 Of course, tions. The exciton binding energy was determined as
new technologies requiring extensive materials growth 110 meV, with direct band gaps of 5.61 and 6.44 eV.92 The
development lose out in a comparison of manufacturability direct gap of 5.61 eV is in good agreement with recent the-
and expense, as embodied by Huang’s chip area manufactur- ory.93 The large exciton binding energy is consistent with
ing FOM (HCAFOM), but even here, the still high cost of the presence of a large amount of electron-phonon-coupling
GaN and SiC substrates mean that Ga2O3 is not uncompeti- in this polar material.94 This data provide a complement to
tive. A clear disadvantage from Table II for Ga2O3 is the that for b-Ga2O3, where the temperature dependence of the
poor thermal conductivity that is embodied in Huang’s high exciton resonance95 as well as the complete dielectric func-
temperature figure of merit (HTFOM).87 tion96 has been measured. Similar to the work on a-Ga2O3,
The high-power/high-voltage market is currently pri- the assumption of strong lattice polarization on the b-Ga2O3
marily dominated by Si lightly-doped metal-oxide semicon- band structure produced a band gap in good agreement with
ductor (LDMOS) and SiC Schottky rectifiers. Ga2O3 is a experiment.97 The low frequency permittivity of b-Ga2O3
candidate to address the ultra-high power market (>1 kW). was determined to be er ¼ 10.2 þ 0.3, and this dielectric con-
The speed with which the cost of Ga2O3 substrate decreases stant was unchanging over the range 77–297 K, and was also
will determine whether Ga2O3 devices can compete with SiC independent of frequency over the range 5–500 kHz.98
Schottky and Si-LDMOS in the medium to high power Furthmuller and Bechstedt reported that the four poly-
market. morphs a-, b-, d-, and e-Ga2O3 exhibit a correlation between
TABLE II. Properties of b-Ga2O3 relative to other more commonly used semiconductors. We also show some of the common figures-of-merit used to judge
the suitability or potential of these materials for various high temperature, high voltage or power switching applications.
Materials parameters Si GaAs 4H-SiC GaN Diamond b-Ga2O3 Comments
Bandgap, Eg (eV) 1.1 1.43 3.25 3.4 5.5 4.85 Bandgap of Ga2O3 reported in range 4.6–4.9 eV
Dielectric constant, e 11.8 12.9 9.7 9 5.5 10
Breakdown field, EC (MV/cm) 0.3 0.4 2.5 3.3 10 8 Experimental values for Ga2O3 have reached
0.5 times the theoretical maximum
Electron mobility, l (cm2/Vs) 1480 8400 1000 1250 2000 300
Saturation velocity, vs (107 cm/s) 1 1.2 2 2.5 1 1.8-2 1.8 h0 0 1i and h0 1 0i, 2.0 h0 1 0i
Thermal conductivity k (W/cm K) 1.5 0.5 4.9 2.3 20 0.1–0.3 0.13 h1 0 0i, 0.23 h0 1 0i
Figures of merit relative to Si
Johnson ¼ Ec2Vs2/4p2 1 1.8 278 1089 1110 2844 Power-frequency capability
Baliga ¼ elEc3 1 14.7 317 846 24 660 3214 Specific on-resistance in (vertical) drift region
Combined ¼ kelVsEc2 1 3.7 248.6 353.8 9331 37 Combined power/frequency/voltage
Baliga high frequency ¼ lEc2 1 10.1 46.3 100.8 1501 142.2 Measure of switching losses
Keyes ¼ k[(cVs)/(4pe)]1/2 1 0.3 3.6 1.8 41.5 0.2 Thermal capability for power density/speed
Huang HCAFOM, el0.5 EC2 1 5 48 85 619 279 Huang chip area manufacturing FOM
atomic density and energetic stability.99 They calculated the an Ir crucible under N2/O2. The sources for achieving n-type
tensor of elastic constants, lattice parameters, atomic coordi- doping with either Sn or Si donors are SnO2 or SiO2, respec-
nates, electron effective mass (0.268 m0 for the b-polymorph), tively.69,134 Ga2O3 single-crystal bulk plates are pulled from
and static dielectric constants (3.55 for b-Ga2O3 and 3.80 for a seed crystal in contact with the molten Ga2O3 at a typical
a-Ga2O3), which are in good agreement with experimental growth rate of 15 mm h1. The major unintentional impuri-
values of 3.4–3.8.100–112 ties in EFG bulk crystals are Si and Ir,69,134 which come
There is also interest in Ga2O3 as a potential surface pas- from the initial Ga2O3 source material and the crucible,
sivation on Si,113 as a gate dielectric material114 and as a thin respectively. This background n-type conductivity can be
tunneling layer to improve open-circuit potential in dye- compensated to produce semi-insulating crystals by addition
sensitized solar cells.115 This same kind of tunneling oxide of deep acceptors such as Mg and Fe. The electrical activity
can be employed in non-volatile, floating-gate memory will of Ir in Ga2O3 is undetermined at this point, but it is likely to
become thinner.116 To prevent leakage of stored charge back have a low solubility.18
to the substrate and therefore a loss of stored bits, new The CZ approach is also very attractive because of its
approaches are needed,116–120 such as ferroelectric random proven capability in growing large diameter, high quality
access memory,116 magnetic random access memory, phase single crystals. In fact, 2 in. diameter, cylindrical Ga2O3 sin-
change random access memory, and resistance random gle crystals grown by CZ have been recently reported,59,127
access memory (RRAM).116 RRAM has advantages of sim- while EFG has also demonstrated large diameters (4 in.) of
ple structure, low energy consumption, and higher operating commercialized crystals.69 Figure 5 shows state-of-the-art
speed.121–123 Studies on the bipolar switching behavior of Pt/ bulk b-Ga2O3 crystals of 2 or 5 cm in diameter grown at
GaOx/TiN show that it originates from conduction associ- either low [(a) and (b)] and high [(c) and (d)] oxygen concen-
ated with oxygen vacancies.116 tration in Ir crucibles.59 Control of the melt thermodynamics
Finally, Ga2O3 has attracted attention as a phosphor host is critical in avoiding decomposition of the oxide and
material for emissive display thin-film electroluminescent increasing crystal size in CZ.127 It is important to note that
(TFEL) displays.20 TFEL devices with emitting layers of when trying to achieve highly n-type crystals using Sn dop-
Ga2O3:Mn, Ga2O3:Cr, or Ga2O3:Eu have shown stable oper- ing in CZ, there are growth instabilities that rapidly produce
ation and high brightness.20,124,125 spiral formation and it was necessary to use either Bridgman
or gradient freeze approaches to obtain high quality, con-
BULK GROWTH TECHNOLOGY OF Ga2O3 ducting crystals.59
Hoshikawa et al.144 published one of the most detailed
As we stated previously, the progress in growing large studies of growth of b-Ga2O3 by vertical Bridgman (VB).
diameter bulk crystals of b-Ga2O3 has been one of the driv- Pt–Rh alloy crucibles were used to contain the melt, which
ing forces in pushing interest in the use of this materials for was carried out in ambient air and single crystals were grown
power electronics and solar-blind photodetectors. In particu- without seeding. The growth direction was perpendicular to
lar, the melt-growth methods developed for Ga2O3 are likely the (100) faceted plane produced in this approach136 and sin-
to be produce crystals at lower cost than the sublimation gle crystals 25 mm in diameter were obtained. Examples of
techniques commonly used for the growth of SiC substrates. the excellent crystals grown by VB in either a full-diameter
This could be a factor in helping to reduce manufacturing crucible or in a conical crucible with a thin seed well are
costs for Ga2O3 based power electronics. Note that since the shown in Fig. 6. The crystal shown in Fig. 6(a) was grown in
b-polymorph is the most thermally stable, all melt-grown a full diameter crucible and that shown in Fig. 6(b) was
crystals are in this form. This monoclinic structure gives rise grown in a conical crucible with a thin seed well.136 This
to anisotropic electrical, optical, and thermal properties. approach led to facile release of the crystals without destruc-
b-Ga2O3 crystals have been grown by all the common tive adhesion to the crucible.
techniques, including the Czochralski method (CZ),59,126–129
floating-zone (FZ),9,66,130–133 edge-defined film fed
(EFG),69,134 or Bridgman (horizontal or vertical, HB and
VB)135–138 growth methods. The bulk crystals can obviously
be used for rectifiers, but more importantly, provide a tem-
plate for growth of epitaxial films of controlled thickness
and doping for active channel and contact layers to allow
achievement of targeted device parameters such as break-
down voltage, on-state resistance, and reverse recovery
time.139–147 EFG is one of the leading candidates of the
established melt growth method in terms of bulk size and
quality for high-volume production of large-size single-crys-
tal b-Ga2O3 wafers. This method is currently used for mass
production of sapphire wafers69,134 and is the method used
FIG. 5. Bulk b-Ga2O3 crystals of 2 [(a) and (c)] and 5 cm [(b) and (d)] in
by the current commercial suppliers of bulk Ga2O3 crys-
diameter grown at low [(a) and (b)] and high [(c) and (d)] oxygen concentra-
tals.16 The charge material in EFG is typically high-purity tion. Reprinted with permission from Galazka et al., ECS J. Solid State Sci.
(5N is typical) Ga2O3 powder which is inductively melted in Technol. 6, Q3007 (2017). Copyright 2017 The Electrochemical Society.59
transistors (MESFETs) and metal-oxide-semiconductor field

effect transistors (MOSFETs) has previously been demon-
strated using MBE, MOCVD, and the variant called Mist
chemical vapor deposition (Mist-CVD). This is a nonvac-
uum, solution-processed approach that offers a highly scal-
able, low-cost route for oxide epitaxy at low growth
temperatures. One major challenge to date for all these
approaches is to overcome the generally slow growth rates.
Alema et al.149 reported high rate growth of epitaxial b-
Ga2O3 thin films on c-plane sapphire substrates using a close
coupled showerhead MOCVD reactor. This uses a small sepa-
ration of the substrate-to-showerhead gas entry of only approxi-
mately 1 cm. They examined use of a few different Ga
precursors, namely, Ga (DPM)3 (DPM ¼ dipivaloylmethanate),
FIG. 6. b-Ga2O3 crystals grown by the Vertical Bridgman method in either triethylgallium (TEGa), and trimethylgallium (TMGa), with
(a) a full-diameter crucible (a0 ), and (b) or in a conical crucible (b0 ). separate injection of molecular oxygen used for the oxygen
Reprinted with permission from Hoshikawa et al., J. Cryst. Growth 447, 36 source. The Ga(DPM)3 was able to be sublimated at 155 C
(2016). Copyright 2016 Elsevier.144
while the TEGa and TMGa sources were kept at or slightly
below room temperature and Ar was used as a carrier gas.149
Zhang et al.138 reported Sn-doped b-Ga2O3 single crys- The sapphire substrates were rotated at 170 rpm to enhance
tals grown by FZ. Compressed powders of b-Ga2O3 (5 N) deposition uniformity. Figure 7 shows transmittance (a) and
and SnO2 (5 N) were used as source materials and crystals growth rates (b), with up to 10 lm/h achieved using TMGa at a
grown at 1500 C for 10 h in air. Post-growth annealing was substrate temperature of 900 C. The bandgap was 4.9 eV.
used to control the conductivity of the crystals and this could The fast growth rates and excellent quality of the films indi-
be varied from semiconducting to insulating. cates this approach can overcome the previously reported low
The defect density in currently available EFG and CZ growth rates in MOCVD.149
crystals is of order 103 cm3, and in EFG, it is possible to Gogova et al.150 grew Si-doped b-Ga2O3 on both Al2O3
achieve an undetectably low density of twin boundaries.69 (0001) and b-Ga2O3 (100) substrates by MOCVD. The
Nakai et al.148 investigated defects in EFG grown (010) b- layers had similar dislocation density to those of the melt
Ga2O3 single crystals by X-ray topography, selective etch- grown substrates. The structural properties were not
ing, and transmission electron microscopy. Both screw dislo- degraded by Si-doping up to 1018 cm3, and in fact, the den-
cations with Burgers vector parallel to [010] and nanopipes sity of twins and stacking faults were up to an order of mag-
were found. The dislocations produced etch pits of approxi- nitude lower in doped layers than in undoped films.150
mately 2 lm dimension after selective etching in potassium Baldini et al.157,158 grew homoepitaxial layers on (100)
hydroxide (KOH) and these had a density in the range b-Ga2O3 substrates by MOCVD and investigated the role of
106–107 cm2. The nanopipes were hollow pipes of 0.1 lm In as a surfactant, which influences the growth dynamics of
diameter and approximately 15 lm length, elongated along the Ga2O3 layers. It was found that In was essential in
[010].148 These led to etch pits approximately 10 lm in achieving the best layer quality. By using the In as a surfac-
dimension and density 102 cm2. Studies of this type that tant, the density of stacking faults and twins was significantly
reveal the origin of the defects have aided in optimizing the reduced and step-flow growth mode was achieved.158 The
current state-of-the-art in which much lower defect densities presence of extended defects is also found to influence the
are obtained.69 effectiveness of n-type doping in homoepitaxial layers grown
by MOCVD.157
PROGRESS IN METAL ORGANIC CHEMICAL VAPOR Zhou et al.170 examined the growth regimes in which b-
DEPOSITION (MOCVD) GROWTH OF Ga2O3 and e-Ga2O3 were stable when grown on c-plane sapphire by
MOCVD. The phase composition was a function of both
As we discussed earlier, bulk growth provides the growth temperature and VI/III ratio, with some conditions
wafers needed for subsequent epitaxy and can also be used leading to a mixture of b- and e-Ga2O3. The structural qual-
to make bulk devices such as vertical rectifiers. However, it ity of e-Ga2O3 was superior to that of b-Ga2O3 under similar
is critically important to develop high quality epitaxial growth conditions.170 With decreasing growth temperature,
growth processes to enable more complex devices to be real- the MOCVD layers went from pure b-Ga2O3 to a mixture of
ized.149–177 The two dominant epitaxial growth methods the b- and e-Ga2O3 polymorphs and finally to microcrystal-
used in GaAs and GaN-based technologies have been Metal line structures at the lowest temperatures investigated.170
Organic Chemical Vapor Deposition (MOCVD) and Figure 8(a) shows a cross-sectional TEM image for a thicker
Molecular Beam Epitaxy (MBE). These provide high quality sample (391 nm) grown at 505 C. The film contains two
layers with practical growth rates and excellent control of pillar-like inclusions. An HRTEM image of inclusion A is
purity, uniformity, and composition. In these early stages of shown in Fig. 8(b). The lattice planes correspond to pure e-
Ga2O3 electronic device development, the fabrication of Ga2O3(0001). The selected area electron diffraction (SAED)
transistors such as metal-semiconductor field effect patterns of island A and the sapphire substrate are shown in
FIG. 8. Cross-sectional TEM analysis for 391-nm-thick Ga2O3 sample with

electron beam parallel to Al2O3 substrate: (a) TEM micrograph with two dis-
tinct islands marked as A and B; (b) HRTEM micrograph of island A; (c)
HRTEM micrograph of island B (“ⵧ” area). Two rotated domains of
b-Ga2O3 with-orientation present either (I) 010 or (II) nearly 132 parallel to
Al2O3; (d) HRTEM micrograph of disordered b-Ga2O3 in island B (“䉫”
area); (e) SAED pattern of island A; (f) SAED pattern of sapphire substrate.
The insets in (b)–(d) present magnified HRTEM images with the ball-and-
stick models of b- and e-Ga2O3 overlaid on the respective projections (the
green and red balls represent Ga and O atoms, respectively. Reprinted with
permission from Zhou et al. Appl Surf. Sci. 420, 802 (2017). Copyright
2017 Elsevier.170
bending.176 In concert with this, there is a recent work that

investigated the role of ambient on the growth modes of
Ga2O3 during sputtering, but this also has relevance to the
initial stages of MOCVD growth.174 Films were deposited
by RF magnetron sputtering on a- or c-plane sapphire or Si
(100) substrates either in an O2 or H2O vapor ambient. The
orientation of the deposited films was a strong function of
ambient and growth temperature.174 For example, when
FIG. 7. (Top) UV–Visible transmission spectra for b-Ga2O3 thin films deposited under O2 gas flow on c-plane sapphire at 300 C,
grown on (0 0 0 1) sapphire substrates by MOCVD. The black, red, and (201)-oriented b-Ga2O3 films were obtained.174 By sharp
green traces are representative of films grown using Ga(DPM)3, TEGa, and contrast, deposition on a-plane sapphire at 600 C led to
TMGa, respectively. The inset shows the plot of (ah)2 versus h used to
obtain the bandgap. (Bottom) Growth rate as a function of substrate temper-
(110)-oriented a-Ga2O3. The use of Si substrates led to depo-
ature for films grown with Ga(DPM)3 precursor using MOCVD. Reprinted sition of randomly oriented polycrystalline b-Ga2O3.174 It is
with permission from Alema et al., J. Cryst. Growth 475, 77 (2017). clear that the activation barrier of nucleation for the b-
Copyright 2017 Elsevier.149 polymorph is reduced in the presence of hexagonally packed
oxygen atoms on c-plane sapphire. Pseudomorphic growth
Figs. 8(e) and 8(f), respectively, and inclusion B was shown of a-Ga2O3 occurred at higher temperatures since it has the
to be composed of several small b-Ga2O3 grains. The meta- same corundum structure as the sapphire. The presence of
stable e-Ga2O3 tends to stabilize at lower growth tempera- water vapor in the ambient promoted nucleation of (311)-ori-
ture, but there was a limited temperature range for the ented c-Ga2O3 on c-plane sapphire.174 Pseudomorphic
formation of e-Ga2O3.170 growth of a-Ga2O3 occured on a-plane sapphire at 800 C
Kim and Kim175 investigated effects of annealing on and above, but with the incorporation of small c-Ga2O3 crys-
MOVCD films grown at 750–1050 C. Postdeposition tallites and both phases were stable at these temperatures.
annealing of amorphous Ga2O3 showed the onset of crystalli- The crystalline phases of Ga2O3 films deposited on sapphire
zation of the b-Ga2O3 phase and the photoluminescence and Si(100) substrates using O2 or H2O ambient are summa-
spectra of as-deposited Ga2O3 went from a dominant blue– rized in Fig. 9 (top).174 Solid-phase epitaxy proceeds from
green and ultraviolet emission, to a longer wavelength UV the interface with the substrate, which provides a templating
band and a new green band after annealing. In addition to the effect. The stable polymorph that grows is determined by the
temperature cycling of epitaxial layers, surface cleaning of choice of the substrate and oxygen source gases. The differ-
substrates prior to epi growth is a key step in promoting ences in the chemical potential between a-Ga2O3, b-Ga2O3,
uniformity and high yield.176 The surface carbon contamina- and c-Ga2O3 are responsible for the dependence on the sur-
tion on Ga2O3 substrates was measured as a function of face structure of the substrate.
annealing temperature up to 800 C under UHV conditions. The bottom of Fig. 9 summarizes the synthesis and
The concentration of adventitious carbon could be reduced interconversion of the different polymorphs of Ga2O3 from
by approximately 70% by annealing at 800 C. Annealing Playford et al.8 They also discovered a gallium oxyhydrox-
also produced defects that led to increased surface band- ide, which upon thermal treatment produces the transient,
lack of control of doping, especially for lightly doped layers

and difficulty in growing ternary alloys. These same disad-
vantages pertain to the use of HVPE for growth of Ga2O3.
Murakami et al.180 demonstrated that use of GaCl and
O2 precursors in HVPE of Ga2O3 produces successful
growth on (001) b-Ga2O3 substrates. They were able to
achieve very smooth surface morphology with the presence
of (100) macrosteps for a growth temperature of 1000 C and
a growth rate of approximately 5 lm/h. This slower growth
rate was a key to controlling morphology. Figure 10 shows
impurity depth profiles measured by Secondary Ion Mass
Spectrometry (SIMS) for homoepitaxial b-Ga2O3 grown at
1000 C on a Sn-doped b-Ga2O3 substrate. The C and H
impurity concentrations were below the background detec-
tion limits of this particular SIMS system, but that is too
high to conclude their accurate concentrations. Similarly, the
Sn and Si levels in the HVPE layer appear to be low and Cl
was detected at a level of approximately 1016 cm3. The net
carrier concentration (ND–NA) measured in the HVPE layer
was approximately 1013 cm3, demonstrating that this tech-
nique is certainly capable of growing thick, high purity
layers with acceptable morphology for diodes and rectifiers.
FIG. 9. (Top) Dependence of crystalline phases of Ga2O3 deposited on sap- Nomura et al.181 also showed from a thermodynamic
phire (0001), sapphire (11
20), or Si (100) substrates with O2 (Ga2O3) or analysis that use of GaCl and O2 as precursors is appropriate
H2O gas. Reprinted with permission from H. Akazawa, Vacuum 123,
8 (2016). Copyright 2016 Elsevier.174 (Bottom) Summary of the synthesis
for the growth of b-Ga2O3 by HVPE. The concentration of
and interconversion of the polymorphs of Ga2O3 (PPT ¼ precipitate, H2 as a carrier gas was a strong factor in the growth kinetics
DEA ¼ diethanolamine and MEA ¼ monoethanolamine). Reprinted with and with GaCl and O2 precursors, the growth is thermody-
permission from Playford et al., Chem.-A Eur. J. 19, 2803 (2013). Copyright namically controlled.
2013 Wiley.8
PULSED LASER DEPOSITION (PLD)

8
polymorph j-Ga2O3. The c-Ga2O3 defect spinel has local Pulsed laser deposition (PLD) is a versatile deposition
disorder with similar octahedral site structure to a- and technique that is valuable in doping and heterostructure stud-
b-Ga2O3.8 Thermal decomposition of gallium nitrate leads to ies because of its relatively low deposition tempera-
the formation of e-Ga2O3, which has a hexagonal close- ture.183–195 However, the main drawbacks for thick device
packed array of oxide ions with partial filling of octahedral structures are the low deposition rate and lower material
and tetrahedral sites.8 Thermal treatment always leads to the
formation of the thermodynamically stable b-Ga2O3 phase.
The d-Ga2O3 was found to be a nanocrystalline form of
e-Ga2O3, while the j-Ga2O3 is a transient phase.8
HALIDE VAPOR PHASE EPITAXY (HVPE)

Halide vapor phase epitaxy (HVPE) is a relatively inex-
pensive approach that achieves extremely high deposition
rates, but usually with a rough surface morphology that is
unsuitable for device processing unless some type of planari-
zation by polishing is employed.177–182 HVPE has been
used to produce b-Ga2O3 thin films at growth rates up to
250 lm/h on sapphire substrates,178–180 but with rough surfa-
ces containing a high density of defects and pits, even when
grown on bulk Ga2O3. It is possible to use chemical mechan-
ical polishing to achieve flat morphologies, but this usually
requires removing as much as 10 lm of material, which is
not economic. The experience from growing GaN by HVPE
shows that there are a number of challenges with this
method. These can include the uncontrolled etching of sub- FIG. 10. SIMS depth profiles of an HVPE layer grown at 1000 C. Arrows
represent the background concentrations of the elements. Reprinted with
strates by the HCl or Cl2 that are used for the formation of permission from Murakami et al. Appl. Phys. Express 8, 015503 (2015).
GaCl vapor, the high growth temperature, uniformity issues, Copyright 2015 The Japan Society of Physics.180
quality compared to the variants of CVD and MBE methods. chloride as the source of Sn-doping. Ultrasonic energy is
PLD has often been used to create doped layers of Ga2O3 as applied to the Ga-containing aqueous solution to create a
n-type transparent conducting layers. The normal dopants mist, which is carried to the heated substrate in the reactor
used have been Sn and Si, and the layers have primarily been using N2 carrier gas, with growth temperatures in the range
deposited on sapphire substrates and with conductivities in of 500–630 C,196–200 i.e., the solution is atomized via ultra-
the range 1 to 10 S cm1. The high end of this range is insuf- sonication at a frequency of 2.4 MHz. The diameter of the
ficient for application as contact layers. Films of Sn-doped mist droplets is approximately 3 lm. The mist of the CVD
Ga2O3 on silica and sapphire with conductivity of approxi- precursors is evaporated and chemically reacts in a heated
mately 1 S cm1 have been deposited at 500 C–550 C for quartz tube containing the substrate. To take advantage of
heteroepitaxial Sn-doped Ga2O3.183–189 For a lower growth the fact that a-Ga2O3 has the same corundum crystal struc-
temperatures of 380 C, Sn doping of Ga2O3 on sapphire, typ- ture as sapphire, the company Flosfia has developed stress-
ically leads to poorer conductivities. For Si-doped Ga2O3 free epitaxy of Ga2O3 on sapphire.209 The optimized growth
on sapphire, Zhang189 achieved a doping level of 9.1 process is now able to produce dislocation-free epilayers on
1019 cm3 with a maximum conductivity of 2 S cm1 and 4 in. diameter sapphire substrates, in an effort to minimize
M€ uller190 was able to control conductivity by adjusting the production costs. They have also developed a proprietary
oxygen partial pressure during deposition. Varying the Si con- lift-off process to transfer the a-Ga2O3 layers from the sap-
tent in Ga2O3 targets allowed control of carrier density from phire substrate to a high thermal conductivity metal support,
1015 to 1020 cm3.191 alleviating the thermal management issues that are still
The growth of Si-doped Ga2O3 by PLD on semi- issues with SiC and GaN power devices and require the use
insulating (010) b-Ga2O3 and (0001) sapphire was used to of diamond-based heat management approaches in some
investigate the effect of substrate on crystal quality.192 The cases.209
films deposited on b-Ga2O3 were single crystal, while those Oshima et al.206 deposited undoped defective-spinel-
deposited under the same conditions on Al2O3 were single type c-phase and corrundum-type a-phase on (100) MgAl2O4
phase, polycrystalline b-Ga2O3.192 The conductivity was and (0001) sapphire substrates, respectively, using Mist-
improved by two orders of magnitude for films on homoepi- CVD. These metastable phases were possible at low growth
taxial substrates relative to those of sapphire, with a value of temperatures, while at higher temperatures, they were trans-
732 S cm1 (mobility of 27 cm2 V1 s1 and carrier concen- formed to the stable b-Ga2O3 polymorph. For the pure c-
tration 1.74 1020 cm3) for the former. This approach phase films, the refractive index was determined to be 2.1 in
might be useful in making low resistance Ohmic contact the visible region, the direct gap was 5 eV, and the indirect
layer in b-Ga2O3 devices, but realistically, the devices need gap was 4.4 eV.206 It was also noted that the growth kinetics
to be grown by one technique and the PLD probably has lim- of Mist CVD were similar to those of that of conventional
ited utility. CVD206 and similar trends have been noted with sol–gel
It has also been found that the presence of a large oxy- derived phase pure a-Ga2O3 nanocrystalline thin films.210 In
gen deficiency in PLD-grown Ga2O3 can lead to a nanocom- that case, at low process temperatures, a mixed phase of a-
posite system that has an insulator to metal transition.193 GaO (OH) and a-Ga2O3 was found, and above this range, a
This mixed system was comprised of crystalline b-Ga2O3 mixed phase of a-Ga2O3 and b-Ga2O3 was detected. The
and an amorphous gallium sub-oxide, Ga2O2.4-a. Nagarajan pure b-phase was observed at higher annealing tempera-
et al.193 also found that, as expected, the oxygen partial pres- tures.210 They determined an optical band gap of a-Ga2O3 of
sure and deposition temperature were key parameters in 4.98 eV from transmittance measurements.210
determining the properties of the deposited films. For a depo- Mist CVD has been used in numerous device demonstra-
sition temperature of 400 C and low oxygen partial pressure, tions, including rectifiers and MESFETs. Dang et al.199 devel-
the film was found to be comprised of Ga metallic clusters oped a AgOx Schottky contact process on Mist CVD grown
embedded in a stoichiometric Ga2O3 matrix. This occurs due films, where Sn-doped Ga2O3 MESFETs and Schottky diodes
to a phase separation of oxygen deficient metastable Ga2Ox and MESFETs showed high rectification and ON–OFF ratios
(x ¼ 2.3) into the stable phases (Ga and Ga2O3).194 and promising breakdown voltages.199 The Sn-doping was able
Finally, PLD has been used to study the effects of Er to achieve a maximum carrier concentration of 7 1018 cm3
doping in Ga2O3 films on sapphire.195 Chen et al.195 was and a mobility of 0.23 cm2 V1 s1.
able to observe the green luminescence line at 550 nm, which Nishinaka et al.205,207 used Mist CVD to grow a-, m-,
was temperature-insensitive up to at least 180 C. While the and r-plane a-Ga2O3 layers grown on a-, m-, and r-plane sap-
intensity of the green emission is quenched with increasing phire using a-Fe2O3 buffer layers. The direct bandgap of the
temperature, it was less severe than comparable Er-doped a-Ga2O3 films were found to be in the range of 5.15–5.2 eV
GaN, consistent with the wider bandgap for Ga2O3. from transmittance and reflectance measurements.205,207
MIST-CVD MBE OF Ga2O3

The Mist CVD method has been successfully applied to From a power device viewpoint, where thick layers are
grow of a-, b-, and c-polymorphs of Ga2O3.196–208 It is a typically needed to achieve high breakdown voltages, the
low-cost approach that employs chemicals such as gallium main drawbacks of MBE are a relatively low growth rate and
(III) acetylacetonate in water as the Ga source, and tin (II) high production cost per wafer. However, MBE is capable of
growing very high quality Ga2O3.211–216 It is a scalable, uni-

form approach that is well-established in other materials sys-
tems. A typical approach is to grow b-Ga2O3 on c-plane
sapphire substrates using an O2 plasma source and a standard
effusion cell for gallium. The Ga2O3 is grown at a substrate
temperature typically around 700 C, with the growth rate
under these conditions of around of 50 nm h1,217 but this
can be increased to approximately 700 nm h1.214
Sasaki et al.213 pointed out that typical vertical geometry
power b-Ga2O3 devices require thick layers (usually well in
excess of 2 lm), with low background doping (<1016 cm3).
The initial studies typically grew homoepitaxial Ga2O3 on
(100) b-Ga2O3 substrates. Excellent morphology and control
of doping was maintained at much higher growth rates by
changing from the (100) plane to the (010) plane.213 The
cleavage planes of b-Ga2O3 have a high decomposition rate
under MBE conditions so epitaxial growth on the cleavage
planes yields a low growth rate. By investigating the growth
rate on different growth planes, it was noted that the rates
were significantly higher (more than an order of magnitude)
on the (010) non-cleavage plane. The same group211 investi-
gated the growth temperature dependence of Sn-doped
homoepitaxial films and found optimum growth tempera-
tures in the range 540–570 C. At higher temperatures, there
was an issue with Sn segregation.211
Okumura et al.214 showed that for b-Ga2O3 (010) homo-
epitaxial growth under Ga-rich conditions and for growth
temperatures >650 C, the growth rate was independent of
the Ga/O ratio (>1). They achieved a growth rate of approxi- FIG. 11. (a) (020) x–2h scan diffraction peak of 135-nm-thick b-
mately 130 nm h1with rms surface roughness of approxi- Ga2O3(010) film grown under slightly Ga-rich condition with an oxygen
flux of 1.2 105 Torr beam equivalent pressure (BEP). (b) Relation of the
mately 1 nm under slightly Ga-rich conditions.214 It is also growth rate of b-Ga2O3(010) at 700 C under slightly Ga-rich conditions to
noted that use of ozone as the oxygen source in MBE produ- the oxygen flux (BEP). Reprinted with permission from Okumura et al.,
ces higher growth rates than O2 plasma MBE (also called Appl. Phys. Express 7, 095501 (2014). Copyright 2014 The Japan Society of
Physics.214
plasma-assisted or PAMBE). However, the latter has better
control of doping since it operates at lower pressures.211–214 exacting layers. It is applicable to a wide variety of materials
Okumura et al.214 employed a PAMBE system with a low including oxides, nitrides, and various types of metals.
background pressure of approximately 105 Torr and were In terms of ALD of Ga2O3, one of the key issues is the
able to grow high quality b-Ga2O3 with good morphologies choice of the Ga precursor.217 Most commonly available gal-
and decent growth rates, as shown in the (020) x–2h scan lium precursors used in ALD require a temperature >300 C
diffraction spectrum of Fig. 11(a). They were able to observe for deposition. The list of these precursors (with their respec-
7th order satellite thickness fringes. The growth rate of the tive required temperature) includes gallium tris-hexafluoro
b-Ga2O3 (010) films was found to be a strong function of the acetyl acetonate (470 C), gallium alkoxide (300 C), gallium
O2 flux at a growth temperature of 700 C under slightly Ga- trichloride (450 C), trimethyl gallium (500 C), and triethyl
rich conditions, as shown in Fig. 11(b).214 gallium (470 C).217 These temperatures are not an issue for
Ga2O3 intended for power electronics structures or for insu-
ATOMIC LAYER EPITAXY (ALE) AND ATOMIC LAYER
lators or passivation layers in microelectronics. By sharp
DEPOSITION (ALD) Ga2O3 contrast, the high process temperature is limiting for flexible
and transparent electronics.
Atomic Layer Epitaxy (ALE) and Atomic Layer Choi et al.217 reported on the properties of ALD Ga2O3
Deposition (ALD) refer to a growth method based on using gallium tri-isopropoxide [Ga(OiPr)3, GTIP] as a Ga pre-
sequential, self-limiting surface reactions. This approach has cursor at 150–250 C and H2O as the oxygen source. These
been used for numerous growth studies of Ga2O3.217–226 films were obviously amorphous, but exhibited smooth surface,
When epitaxy occurs (typically at higher temperatures) on a low carbon background, and high transparency (>90%). The
homoepitaxial substrate, ALE is the correct description, films also had high breakdown fields (7.56 MV/cm at 250 C)
while ALD refers to the more general case of deposition. and low leakage current (1.92 1011 A at 1 MV/cm). The
ALD features excellent thickness control and conformal cov- growth rate was 0.25 nm/cycle at 250 C. The Ga2O3 films
erage on substrates and is widely used in the microelectron- show outstanding transmittance (over 90%) with an optical
ics industry for depositing dielectrics and other thin, bandgap energy level of 5.4 eV.
Altuntas et al.226 deposited Ga2O3 dielectric thin films

(7.5 nm thick) on p-type silicon wafer using plasma
enhanced atomic layer deposition (PEALD). After the depo-
sition, annealing was performed up 800 C to obtain the b-
phase. These b-Ga2O3 thin films deposited by PEALD tech-
nique at low temperatures showed promise as oxide layer for
MOS devices.
Ga2O3 NANOSTRUCTURES
The various forms of low dimensional, crystalline
Ga2O3 that includes nanowires (NWs), nanobelts (NBs), and
nanosheets (NSHs) are the subject of an extensive litera-
ture.227–303 This is due to both the interest in the fundamental
properties of the material at reduced dimensions and the pos-
sible application to devices, including nanowire-based field
effect transistors (NW-FETs), gas sensors, solar blind photo- FIG. 12. SEM images of Li doped Ga2O3 microwires. (a) General view of
detectors, and switches. It is well-established that devices the microwires grown on the substrate surface (b)–(d) details of several
structures with lateral pyramids. (e) Representative image from a sample
such as gas sensors fabricated on nanostructures can have grown in the same conditions, but with no presence of Li2CO3. Reprinted
advantages in sensitivity and response time due to the high with permission from L opez et al., Semicond. Sci. Technol. 31, 115003
surface to volume ratio. Of course, the reproducibility of the (2016). Copyright 2016 IOP.279
structures, making high quality contacts with high yield and
the reliability are all issues, as well as the need to incorporate revealed the presence of a peak at 270 cm1 only in the Li-
dopants to control their electrical and optical properties doped samples.279 This was accompanied by a sharp lumi-
(including both doping during growth and post growth dop- nescence peak at 717 nm, as shown in the cathodolumines-
ing by ion implantation or diffusion). cence (CL) emission from a Li-doped structures (solid line)
There have been a large number of approaches used to and an undoped sample (dashed line) of Fig. 13(a). Figure
grow Ga2O3 nanostructures, with the most common being 13(b) shows an expanded view of the Li-related transition/
the vapor–liquid–solid (VLS) or vapor–solid (VS) pro- via Ga2O3 interband excitation and by defect band
cess.228 In this process, metal catalysts or metal alloys are excitation.279
used which form supersaturated alloy droplets by dissolving Hosein et al.283 synthesized high aspect ratio b-Ga2O3
the vapor component.249,261,284 Supersaturation in alloy nanowires with facetted morphology by VLS. The long axis
droplets induces the precipitation of a solid phase for nano- of the nanowires was in the direction parallel to the crystal b
wire growth. Other methods employed for Ga2O3 have axis ([0 1 0]). The nanowires showed strong blue lumines-
included physical evaporation,252 thermal annealing of com- cence under UV laser excitation, ascribed to native
pacted gallium nitride powder,259 solution processing,265 arc defects.283 Figure 14(a) shows a TEM image of ab-Ga2O3
discharge,253 laser ablation,254 carbothermal reduction,255 nanowire, along with the selected area diffraction pattern
chemical vapor deposition (CVD),256,264,266,267,278 metal–or- (SAED) pattern. The streaking in the diffraction pattern was
ganic CVD (MOCVD),264,268 microwave plasma reaction,238 ascribed to the presence of line defects, such as stacking
and vapor phase epitaxy (VPE).267 Kumar et al.278 reported faults and micro-twins.283 This was not present in all areas
a comparative study of Ga2O3 nanowires grown on different and the SAED from the defect-free regions and lattice imag-
substrates by CVD. ing [Figs. 14(b) and 14(c)] showed the growth direction was
The most common approach to growing b-Ga2O3 NWs
has been to use Au as a catalyst.228 This always has the risk
of some level of incorporation of the catalyst into the b-
Ga2O3. The electrical activity of Au in Ga2O3 is not clear,
but it likely to be a recombination center. Johnson et al.275
reported the use of Fe as the catalyst, an approach that is
commonly used for the growth of carbon nanotubes.
Lithium is of potential interest as a dopant in Ga2O
nanostructures order to modify its conductivity for electrical
energy storage applications.279 Lopez et al.279 reported the
results of a study of in situ doping with Li on the morphol-
ogy and optical properties of large Ga2O3 nanostructures
grown by VLS.279 Figure 12(a) shows SEM images of Li-
doped Ga2O3 microwires on the substrate, where images FIG. 13. (a) Room temperature CL emission from Ga2O3:Li microstructures
(solid line) and a reference undoped sample (dashed line). (b) Details of the
(b)–(d) show the typical pyramidal surface morphology.
red luminescence band observed in Ga2O3:Li microstructures. Reprinted
These Li-doped structures have quite different structure from with permission from L opez et al.279 Semicond. Sci. Technol. 31, 115003
the undoped case, as shown in Fig. 12(e). Raman spectra (2016). Copyright 2016 IOP.279
modes compared to an undoped bulk Ga2O3 single crystal

grown by CZ, while the CL spectra both had a strong broad
UV–blue emission band and a weak red emission band. The
bandgap of this nanowire was measured to be 4.94 eV. There
was little difference in the structural, morphological, and
optical properties of Fe-catalyzed b-Ga2O3 nanowires com-
pared to those of Au-catalyzed wires grown under similar
conditions.249 Figure 15 shows TEM images of the b-Ga2O3
nanowire with the expanded area showing the atomic struc-
ture of b-Ga2O3 in the ½112 orientation.
Ga2O3 can also be used for photocatalytic decomposi-
tion of acids in water.301 Ga2O3 nanostructures have been
used to efficiently remove trace perfluorooctanoic acid
(PFOA) in the secondary effluent of sewage. The decomposi-
tion of targeted trace pollutants such as PFOA is generally
inhibited by coexisting high-level natural organic matters
and bicarbonate, and because of its ubiquity, there is a strong
effort focused on its remediation by decomposition.301
While many of the VLS and other solid- and gas phase
synthesis methods require sophisticated equipment, it is also
possible to use simpler solution-based methods to synthesize
Ga2O3 nanostructures.302 Metastable c-Ga2O3 nanorods (and
cubic In2O) nanoparticles produced by solution processing
using ethylene glycolate precursors are crystalline with nar-
row size distribution and exhibit intense blue light
emission.302
PROPERTIES AND PROCESSING OF b-Ga2O3

In this section, we focus more specifically on the proper-
ties of the b-polymorph,303–310 since that is the focus of most
of the recent interest in power electronics and solar-blind
UV detectors. To summarize, a b-Ga2O3 single crystal has a
thermodynamically stable monoclinic structure (b-axis
unique) belonging to the space group C2/m. The crystal
structure of b-Ga2O3 has two crystallographically non-
equivalent Ga positions (tetrahedrally and octahedrally coor-
dinated) and two cleavage planes [(100) and (001)]. b-Ga2O3
has two different Ga sites, denoted as Ga(1) and Ga(2), and
FIG. 14. TEM analysis of faceted nanostructures. (a) Low magnification
three different O sites, denoted as O(1), O(2), and O(3). As a
TEM image of a representative nanowire. Inset shows the corresponding
SAED pattern. The arrow indicates the growth direction, which corresponds
to [0 1 0]. Scale bar, 0.5 lm. (b) SAED along the [0 0 1] zone axis, on the
corner formed by the side wall and the tip. The reciprocal lattice has a unit
cell angle b ¼ 104.3 . Interplanar spacings determined from the diffraction
spots match the corresponding labelled directions. Inset shows the nanowire
region analyzed. Scale bar, 20 nm. (c) Lattice-resolved TEM image of the
corner of the nanowire. Lattice spacing measurements parallel and perpen-
dicular to the growth direction indicate [0 1 0] and [1 0 0] crystalline direc-
tions, respectively, in agreement with the SAED. The unit cell face parallel
to the lattice plane is indicated, with a ¼ 12.01 Å and b ¼ 2.98 Å, magnified
by a factor of 2 for clarity. Scale bar, 5 nm. Reprinted with permission from
Hosein et al., J. Cryst. Growth 396, 24 (2016). Copyright 2016 Elsevier.283
along the [0 1 0] direction and the lattice spacing perpendic-

ular to the growth direction corresponded to {1 0 0} planes.
Kumar et al.249 used Fe catalyst to synthesize single
crystal b-Ga2O3 nanowires with diameters of 30–80 nm and FIG. 15. Annular dark-field scanning transmission electron microscope
(ADF-STEM) image of the b-Ga2O3 nanowire with a zoomed in area framed
lengths up to tens of micrometers by VLS using CVD. The
on the nanowire, which illustrates the atomic structure of b-Ga2O3 in (1–12)
Raman peak positions from the b-Ga2O3 nanowire had a red orientation. Reprinted with permission from Kumar et al., J. Phys. D: Appl.
frequency shift of 0.3–1.4 cm1 for all three active frequency Phys. 47, 435101 (2014). Copyright 2014 IOP.249
result of this, b-Ga2O3 has an anisotropy for the optical and discussed earlier, theoretical predictions suggest that b-
electrical properties.34,304 There are also orientation- Ga2O3 has an indirect gap which is only 30–40 meV smaller
dependent etching and contact effects that are relevant for than the direct band gap.6,51 The Varshni parameters were
device fabrication. Ueda et al.13 concluded that the electrical a ¼ 4.45 103 eV/K and b ¼ 2000 K in the relation Eg
anisotropy is caused by the anisotropic effective mass, and (T) ¼ Eg(0) aT2/(bþT).311 The absorption data and
the optical anisotropy is caused by the selection rule of the bandgap as a function of n-type carrier concentration are
band-to-band transition.305 The Ueda et al.57 data showed shown in Fig. 17.
Hall mobility measured along the b and c axes of high qual- Through extended H€uckel method calculations together
ity FZ crystals to be 46 and 2.6 cm2 V1 s1, respectively, with measurements of the Electron Paramagnetic Resonance
corresponding to electron masses of 0.44 mo and 1.24 mo, (EPR) and electron nuclear double resonance, Binet et al.306
respectively. This shows a higher curvature of the bottom of concluded that the electrical anisotropy observed in b-Ga2O3
the conduction band along the b-axis in b-Ga2O3.57 A num- originates from the anisotropy of the conduction electrons.
ber of groups have suggested that the CBM belongs to C1þ, These electrons exhibit an anisotropic distribution along the
and the valence band maximum (VBM) and the second octahedral chains of the monoclinic structure. Based on
valence band belong to C1 and C2, respectively.305 semiempirical quantum-chemical calculations, Hajnal
It is worth pointing out that the bandgap reported for b- et al.307 also concluded that the optical and the electrical ani-
Ga2O3 is somewhat dependent on the technique used to mea- sotropies are attributed to the anisotropy of the electronic
sure it. For example, Fig. 16 shows the bandgap derived structure of the lattice. Hajnal et al.307 also carried out calcu-
from Reflection electron energy loss spectroscopy (REELS). lations on neutral and doubly charged vacancies and were
The bandgap was determined from the onset of the energy able to explain the temperature dependence of the electrical
loss spectrum. The bandgap of Ga2O3 is a function of the conductivity in terms of the donor levels.307 Yamaguchi34
polytype and it is common in the literature to quote a value and Litimein et al.304 produced some of the earliest compre-
of 4.8 eV for b-Ga2O3, but we and others have noted that hensive studies of the electronic structure of b-Ga2O3 using
optical transmittance and REELS data show a consistent the first principles full-potential linearized augmented plane
value of 4.6 eV.69,71 However, note that the REELS spectrum wave method. It was found that b-Ga2O3 has an indirect
exhibits a tail commonly ascribed to defects and this can band gap with a conduction band minimum (CBM) at the C
affect how the bandgap is derived from the intersection of
the slopes. Experimentally reported variations in determina-
tion of the gap may also result from the anisotropy of the
crystal structure, which causes a dependence of the absorp-
tion on the polarization of the incident light.
Rafique et al.311 recently used photoluminescence exci-
tation (PLE) and absorbance spectra to measure the tempera-
ture and doping dependence of bandgap of Si-doped layers
grown by low pressure CVD in sapphire substrates. The
bandgap decreased with temperature and increased with
electron concentration up to 2.5 1018 cm3. Beyond that
density, the Burstein-Moss shift was swamped by the onset
of the Mott transition to metallic-like conduction.311 As we
FIG. 17. Carrier concentration dependence of bandgap for b-Ga2O3 grown

FIG. 16. Typical REELS spectrum from bulk b-Ga2O3. Reprinted with per- on sapphire by CVD. Reprinted with permission from Rafique et al., Opt.
mission from Appl. Phys. Rev. 4, 021301 (2017). Copyright 2017 American Mater. Express 7, 3561 (2017). Copyright 2017 Optical Society of
Institute of Physics.560 America.311
point and a valence band maximum on the E line. The aniso-

tropic optical properties were explained by the selection rule
of the band-to-band transitions. On the other hand, the shape
of the CBM was found to be almost isotropic, and, therefore,
the observed electronic anisotropy in the n-type semicon-
ducting state should not be attributed to the properties of a
perfect lattice but to vacancies.34,304,308,309 The anisotropic
optical properties were explained by the selection rule of the
band-to-band transitions.304 For the monoclinic phase, it was
shown that the component with y-direction was more pro-
nounced than those along the x and z.304
The surfaces of b-Ga2O3 are also of strong interest from
a practical viewpoint since their stability will determine how
passivation schemes need to be incorporated. One of the big-
gest drawbacks of ZnO, another electronic oxide that
attracted interest for blue light emitters and transparent elec-
tronics,312–319 was the sensitivity of its surface to oxygen
and water vapor.313–318 Similarly, InGaZnO4, (IGZO) an
amorphous transparent conducting oxide of interest for trans-
parent thin film transistors (TFTs), has a strong surface sensi-
tivity to water vapor.312,316–319 The adsorption of oxygen
and water vapor from ambient atmosphere on the exposed
channel layer create changes in the conductivity of the near-
surface on both these materials and an instability of the chan-
nel conductivity during device operation of light-emitting
diodes (LEDs) and thin film transistors (TFTs). Under vac-
uum and hydrogen ambient, the threshold voltages of IGZO
TFTs showed a negative shift and the drain current increased
in the typical transfer curves.316–319 The trend was in the
opposite direction under oxygen-containing ambients.316–319
The change in the TFT positive threshold voltage shift under
the gate-bias stress was larger in hydrogen than in vacuum,
or even in oxygen, suggesting that hydrogen plays an impor-
tant role in the instability of a-IGZO TFTs under gate-bias
stresses.
Thus, it is of interest to understand the fundamental
properties of the surfaces of Ga2O3. Bermudez88 reported
on the physical and electronic structure of the (100), (010),
(001), and (10 1) faces of b-Ga2O3 using ab initio theory.
All surfaces exhibited a decrease in surface energy upon
relaxation, and the local bonding at the surface is analyzed
by comparing nearest-neighbor bond lengths and overlap
populations with those in the bulk. The (101) surface,
which exhibits a high energy when ideally terminated,
undergoes large displacements and changes in bonding FIG. 18. Polarized Raman spectra of (a) (010) Mg-doped, (b) (100) Si-
during relaxation leading to a substantial lowering of the doped, and (c) (001) undoped b-Ga2O3 substrates at room temperature (RT).
surface energy. Reprinted with permission from Onuma et al., J. Cryst. Growth 401, 330
(2014). Copyright 2014 Elsevier.309
Klinedinst and Stevenson308 measured the Gibbs energy
of formation of b-Ga2O3 to be 252 kCal mol1 over the
temperature range of 873–1273 K. Onuma et al.309 measured of b-Ga2O3 and was made possible by the high quality of the
lattice vibration modes from the complete set of polarized crystals produced in this study by EFG and FZ methods.309
Raman spectra from bulk (010) Mg-doped, (100) Si-doped, Defect-related luminescent centers in bulk, FZ, Ga2O3
and (001) unintentionally-doped b-Ga2O3. The Ag and Bg crystals doped with 1 at. % In, Tl, Sn, Pb, Sb, and Bi (ns2)
Raman optically active modes corresponding to the direction ions were examined by PL and scintillation decay analy-
of the electric field vector for the incident light, and that for sis.310 The Sn-doped samples showed the highest scintilla-
the back scattered light, respectively, were perfectly sepa- tion light yield. Peak emission was observed around 2.8 eV
rated in the spectra according to the polarization selection in photoluminescence (PL) under the excitation energy of
rules. As shown in Fig. 18, this was the first experimental 4.68 eV and around 3.0 eV in scintillation under X-ray irradi-
observation of the complete set of polarized Raman spectra ation. Both the PL and scintillation decay times could be
well-fit by assuming the presence of three exponential decay contacts or activating dopants affects the conductivity, since
functions with time constants ranging from tens of nanosec- this is a function of the ambient employed. In Ga2O3, it is
onds to several microseconds. The slow component was often observed that annealing in oxygen reduces the free
ascribed to nsnp-ns2 transitions while the fastest and inter- electron density, while annealing in nitrogen or hydrogen
mediate components were ascribed to defect-related leads to an increase in n-type conductivity.69,127,329,330 N-
transitions.310 type doping of b-Ga2O3 in both bulk crystals and epitaxial
The mechanical properties of b-Ga2O3 under deforma- films has shown to produce controllable carrier concentra-
tion during indentation have recently been studied.320 This is tions from 1016 to 1019 cm3, with an upper limit near
relevant due to the current need for polishing of thick HVPE 1020 cm3. In MBE, the usual dopants have been Sn and
layers on bulk substrates to provide acceptable surface mor- Ge,334 while Si and Sn are the most common dopants in
phology, but also for the grinding and polishing of the initial MOCVD.2,12,16 Note that there is a good correlation of elec-
bulk wafers cut from boules. Wu et al.320 examined the trical (from Hall data) and total dopant concentration (from
deformation patterns of b-Ga2O3 under different loading SIMS measurements up to approximately 1019 for Sn doping
conditions using TEM. They measured the load and displace- and even higher for Si doping in MOCVD,72 as shown in
ment curves from indentation under loads ranging from 0.2 Fig. 19).
to 10 mN and found that as the load was increased, there was This choice of dopants for each growth technique
firstly, formation of stacking faults along the (200) lattice appears to be simply a function of availability of those dop-
planes and twinning structures with the (201) plane as the ants on the particular growth systems used and not due to
twin boundary, then formation of dislocations on (101) lat- superiority of one n-type dopant over another. Ion implanta-
tice planes and, finally at the highest loads, there was bend- tion using Si has also been used to enhance the conductivity
ing of the (201) planes and cracking that propagated along of n-type layers for Ohmic contact formation.337,338 The
the (200) planes.319 This behavior is unique to Ga2O3 and maximum carrier concentration achieved in this case was
implies that effective grinding and polishing processes for approximately 5 1019 cm3, although the parameter space
this material must be developed and not simply transferred for optimized annealing conditions has really not been
from recipes for other materials.320 explored in detail and it is likely even higher carrier concen-
To summarize, a number of parameters, in b-Ga2O3 trations can be achieved. Kang et al.333 have examined the
including electron mobility, thermal conductivity, and even theoretical fundamental limits to mobility in Ga2O3. To give
bandgap are crystal orientation-dependent, anisotropic, and some idea of typical electron mobilities achieved for these
also a function of the synthesis method. This is expected at doping levels, in Sn doped b-Ga2O3 grown by MBE,
this relatively early stage of materials development and Ahmadi334 achieved a mobility of 39 cm2 V1 s1 at a donor
much of the spread in reported parameters will be reduced as concentration of 1 1020 cm3. In Si doped b-Ga2O3 grown
the material quality becomes more standardized. by MOVPE, Baldini72,157 achieved a mobility of 50 cm2
V1 s1 at a donor concentration of 8 1019 cm3.
DOPING AND DEFECTS IN Ga2O3 Understanding the origin of unintentional doping in
Ga2O3 is necessary in reducing the background conductiv-
As with any new wide bandgap materials technology, ity.338–354 This has implications for devices, where the
the control of conductivity through doping and mitigation of
trap states is key to realizing device applications.321 This has
been the subject of extensive recent summaries of the chal-
lenges and opportunities in the field by Tsao et al.321 and
specifically for Ga2O3 by Bayraktaroglu.322 The fact that
large area bulk substrates and high quality epitaxial growth
processes is available for Ga2O3 is a major plus for this
material, but the low thermal conductivity must be addressed
by active measures to transfer the Ga2O3 to more thermally
conducting heat-sinks, as has been done with various
approaches to incorporating diamond into GaN power device
design to facilitate heat removal during high power continu-
ous or pulsed operation.322–325 It is important to note that
this has been to be done in conjunction with a careful analy-
sis of where the thermal bottle-necks occur in the full device
structure and that this may require not just a simple heat-sink
under the active region, but conformal layers around the
gate.321–325
Doping in electronic oxides is always complicated by
self-compensation, solubility, and defects issues.326–341
FIG. 19. Hall free carrier concentration versus the dopant (Si and Sn) con-
There is also an asymmetry in that one conductivity (usually
centration obtained by SIMS. Reprinted with permission from Baldini et al.
n-type) is favored over the other.327–329 An additional con- ECS J. Solid State Sci. Technol. 6, Q3040 (2017). Copyright 2017 The
sideration is how post-growth annealing when forming Electrochemical Society.72
breakdown voltage of rectifiers is a strong function of doping Nakano356 used photocapacitance measurements to
level, and obviously, the mobility is degraded by the pres- show there are four dominate traps in state-of-the-art EFG
ence of existing impurities. If they could be minimized, Ma material, with levels at 0.8, 2.04, 2.71, and 3.87 eV
et al.353 showed that polar optical phonon scattering would below the conduction band, plus a deep-level defect with its
then be the factor limiting electron mobility for lightly doped optical onset at 3.71 eV above the valence band Ev.356
samples. Analysis of the main impurities in EFG bulk Ga2O3 These are shown schematically in Fig. 20.356 This material
shows that Si, Fe, and Ir are the main impurities.69 Varley also showed strong near band-edge luminescence, an indica-
et al.51 used theory to suggest that oxygen vacancies are tor of its high quality and an optical bandgap of 4.49 eV.356
deep donors with ionization energies larger than 1 eV and Wang et al.357 investigated formation energies and tran-
that these cannot explain the unintentional n-type back- sition energies of possible donor-like defects in GaInO3
ground conductivity. In their work, Si, Ge, Sn, F, and Cl using hybrid density functional theory. The goal was to
were found to be shallow donors51 and hydrogen in either understand possible sources of the experimentally observed
interstitial or substitutional sites can act as donors. n-type conductivity in this material, which has similar chem-
By contrast to the case of n-type doping, there is little istry to Ga2O3. They also found that O vacancies are deep
experimental work on p-type doping of Ga2O3 where it was donors, while interstitial Ga and In were shallow donors but
found that Mg doping increases resistivity of the material, with much higher formation energies (>2.5 eV), making
but does not lead to p-type conductivity. Theoretically, it has them less likely to be present.357 They found that intrinsic
been suggested from the band structure calculations that the defects were not responsible for high levels of n-type con-
holes have large effective mass and tend to form localized ductivity. This study found that impurities such as Sn and
polarons rather than mobile free holes in the valence band. Ge, as well as other commonly incorporated impurities like
This will obviously be an area of active research as novel H, can act as shallow donors.357
methods of acceptor incorporation are examined. The properties of rare earth ions incorporated in b-
Son et al.347 used electron paramagnetic resonance to Ga2O3 by ion implantation have also been investigated.
examine the properties of the background donor in bulk EFG Peres et al.358 used ion implantation at temperatures of Eu at
crystals. They found an effective donor ionization energy of temperatures up to 1000 C, using a fixed energy 300 keV
44–49 meV, which was reduced due to banding at higher Euþ beam. Using an elevated implantation temperature in
concentrations. The donor exhibited negative-U behavior, the range of 400–600 C increased the substitutional Eu frac-
with the negative charge state below the neutral charge state, tion present without annealing and the corresponding density
and the concentration was a strong function of annealing of Eu3þ ions.358 The ratio of Eu in the 2þ and 3þ charge
temperature.347 In their samples, Si was the dominant chemi- states was a strong function of the implantation and subse-
cal impurity present.347 Other studies have reported the ioni- quent annealing temperature.358 For Ga2O3 damaged by ion
zation level of unintentionally incorporated donors present at implantation, it was found, as expected for a binary lattice
a concentration of approximately 1016 cm3 in commercially with elements of vastly different vapor pressures, that the
available (201) Ga2O3 substrates to be 110 meV from tem- damage recovery during annealing is complex and a recov-
perature dependent Hall effect measurements and 131 meV ery of the crystalline damage was not complete by 1000 C.
from admittance spectroscopy.352 The presence of such rela- Figure 21 shows cathodoluminescence spectra from samples
tively deep donors and their ionization percentage at device implanted at either (a) 300 C or (b) 600 C, which exhibit
operating temperatures could have an effect on the on-state the sharp Eu3þ intra-ionic 4f transition lines in the
resistance and breakdown voltage of rectifiers.352 The analy- 600–700 nm region. The broad emission peaked near 400 nm
sis352 showed that to achieve 10 kV breakdown voltage in is due to defects and impurities.359–364
Ga2O3 Schottky diodes, made on materials containing these
deeper unintentional donors, the concentration of these
donors must be below 5 1014 cm3.
Korhonen et al.354 investigated the electrical compensa-
tion in n-type Ga2O3 by Ga vacancies in Ga2O3 thin films
using positron annihilation spectroscopy, which is a tech-
nique for studying vacancy defects in semiconductors. They
estimated a VGa concentration of at least 5 1018 cm3 in
their undoped and Si-doped samples.354 Since theoretical
calculations51 predict that these VGa should be in a negative
charge state for n-type samples, they will compensate the n-
type doping.354 Kananen et al.355 used electron paramagnetic
resonance (EPR) to demonstrate the presence of both doubly
ionized (VGa2) and singly ionized (VGa) acceptors at room
temperature in CZ Ga2O3. The singly charged state had the
FIG. 20. Schematic of positions in the bandgap of deep level traps in state-of-
two holes localized on oxygen ions on either side of the VGa
the-art EFG Ga2O3. Reprinted with permission from Y. Nakano, ECS J. Solid
and indicated that both types of vacancies are unlikely to be State Sci. Technol. 6, P615 (2017). Copyright 2017 The Electrochemical
shallow acceptors.355 Society.356
FIG. 21. RT CL spectra of b-Ga2O3 samples implanted at 300 C (a) and at 600 C (b) annealed at different temperatures. Reprinted with permission from
Peres et al., J. Phys. D: Appl. Phys. 50, 325101 (2017). Copyright 2017 IOP.358
Johnson et al.365,366 examined three-dimensional imag- devices.368–381 When GaN is heated at 900 C, thick films
ing of individual point defects using selective detection (0.5 lm thick) of b-Ga2O3 can be grown in a day on this sur-
angles in annular dark field scanning transmission electron face.374 When non-native oxides are grown on the GaN, they
microscopy (STEM). This was related to trying to under- can be contaminated by surface contaminants from this
stand the origin of residual n-type conductivity in Ga2O3 and material.368–372 By directly growing the Ga2O3 on GaN, this
the fact that unlike in some other transparent conducting issue is mitigated and the grown oxide has been shown to be
oxide (TCO), oxygen vacancies may not be the reason both chemically stable and have a low density of interface
because they are deep donors in b-Ga2O3. The other sus- states.374–380 There have been a wide variety of techniques
pected causes, such as Si and H, may also complex with employed to grow or deposit Ga2O3 on GaN, including low
native defects such as Ga vacancies. Johnson et al.365,366 temperatures approaches such as photoelectrochemical
reported the imaging of point defects in Ga2O3 using elec- (PEC) oxidation, PLD, excimer laser induced oxidation,
tron channeling contrast.366,367 The channeling contrast microwave plasma oxidation, oxygen plasma oxidation, and
method detects the de-channeling of the electron caused by saturated water vapor oxidation, in addition to the high tem-
individual point defects, including vacancies and impurity perature thermal oxidation.368 As expected, the latter produ-
atoms, using multiple narrowly selected detection angles in ces a significantly lower interface trap density (usually at
STEM.367 The technique can be used to determine the struc- least an order of magnitude) relative to Ga2O3 that is depos-
ture and positions of individual point defects, as shown sche- ited and also compared to other deposited insulator/GaN
matically in Fig. 22(a) and in an actual image in Fig. interfaces.368 An example of an XRD spectra from GaN
22(b).365,366 The left inset shows a simulated image that
matches the structure in (a), and the inset at right shows the
experimental position averaged convergent beam electron
diffraction pattern taken from a unit cell of b-Ga2O3.365,366
The realization of such images requires TEM samples of
<10 nm thickness. This is usually achieved by cleaving and
mechanical wedge polishing, which limits the orientations
that are available.
Thin layers of Ga2O3 can also be used as a gate dielec-
tric on GaN for high power metal-oxide-semiconductor
FIG. 22. (a) Potential lattice defects in b-Ga2O3. (b) An experimental STEM
high angle annular dark field (HAADF) image of b-Ga2O3. Inset (left) is the
simulated image that matches the structure in (a), and (right) is the experi-
mental position averaged convergent beam electron diffraction pattern taken FIG. 23. XRD spectra of GaN epilayers oxidized at 800 C, 900 C,
from a unit cell of b-Ga2O3. Reprinted with permission from Johnson et al., 1000 C, 1050 C, and 1100 C for 1 h in dry oxygen ambient. Reprinted
Microsc. Microanal. 23(S1), 1454 (2017). Copyright 2017 Microscopy with permission from Chen et al., Appl. Phys. A 71, 191 (2000). Copyright
Society of America.365 2000 Springer.381
oxidized at oxidized at 800 C, 900 C, 1000 C, 1050 C, other of these types, but not both with the carrier concentra-
and 1100 C for 1 h at dry oxygen ambient is shown in Fig. tions needed for light-emitting devices. Intentional doping
23.381 This shows the clear signature of b-Ga2O3.381 with the n-type dopants Si and Sn has allowed demonstration
Gallium oxide films have also shown potential as passiv- of numerous types of devices and the doping levels are con-
ation layers on silicon solar cells.382–387 This is of particular trollable in the range of 1015–1019 cm3 and higher in some
interest for high efficiency solar cells, due to the reduction of cases when needed for Ohmic contact regions. The develop-
the surface recombination of excited carriers. There have ment of p-type doping in b-Ga2O3 is much less advanced,
been many transparent dielectric films investigated for pas- and there is even a question as to whether p-type conductiv-
sivation on solar cells, including SiO2, silicon nitride (SiNx), ity can ever be achieved in Ga2O3. The heavy valence band
aluminum oxide (Al2O3), amorphous silicon, and multilayers effective mass and a strong Fr€ohlich interaction favors the
of these materials.382 Ga2O3 is known to effectively passiv- formation of polarons with very low conductivity.
ate p and pþ crystalline Si surfaces, and has been used as an Nitrogen has been explored as a dopant for other oxides
alternative to Al2O3 films.383 Based on its transparency and such as ZnO, but was found theoretically to not lead to
conductivity, Allen and Cuevas383 have suggested that acceptor doping.393,394 The difficulties in achieving p-type
Ga2O3 might be alternatives to the currently used conductive doping for ZnO is related to the formation of compensating
layers in the photovoltaic devices, such as doped amorphous defects, the presence of donor impurities, the low solubility
silicon and tin-doped indium oxide. Xiang et al.382 reported (typically <1018 cm3) of the acceptor dopants, and the large
on the characterization of spin-coated Ga2O3 films for use as ionization energy (170–380 meV) of all of the acceptor can-
passivation layers on Si. They used gallium nitrate hydrate didates. Nitrogen was predicted to have the lowest ionization
(Ga(NO3)39H2O) dissolved in deionized water with addition energy of the possible acceptors. N is a natural choice for an
of a surfactant, namely, 1 vol. % of polyethylene glycol acceptor dopant, since it has about the same ionic radius as
monomethyl ether [HO(CH2CH2O)nC12H25, n ffi 25].382 The that of O, and thus should readily substitute for the latter.
films showed an amorphous-to-crystalline phase transforma- While there were reports of successful p-type doping of ZnO
tion above 550 C.382 As the temperature was increased from with nitrogen,312–314 the conductivity was often unstable and
450 to 850 C, the interface state density decreased from degraded with time after growth. In addition, the doped ZnO
4.21 1011 to 1.53 1011 eV1 cm2. The crystalline was often found to remain either semi-insulating or n-type
b-Ga2O3, being monoclinic and thus biaxial, has a triaxial due to the sensitivity of the state of the nitrogen on several
ellipsoidal refractive-index surface.386 factors. For example, the N incorporation may be affected by
polarity of ZnO and even when N is incorporated in ZnO its
OXIDE p-n JUNCTION HETEROSTRUCTURES efficiency as an acceptor is expected to largely depend on
the local surrounding and the presence of other impuri-
In the absence of clear demonstrations of p-type doping
ties.393,394 For example, nitrogen substituting for oxygen,
and the theoretical work that suggests it may not be possible
NO, is calculated to be a deep acceptor with an energy level
to get free holes in the material, an alternative strategy is to
of Ev þ (1.6–1.7) eV.393,394 On the other hand, the acceptor
use a natively p-type oxide to produce p-n heterojunctions
ionization energy is predicted to be significantly reduced if
involving n-type Ga2O3. Kokubun et al.387 demonstrated
(i) NO is surrounded by isovalent group II atoms substituting
NiO/b-Ga2O3 all-oxide p–n heterojunction diodes using
Zn; (ii) NO is a part of complexes with residual group III
p-type NiO epitaxial layers grown on n-type b-Ga2O3 sub-
contaminants (Al, Ga, In); or (iii) NO forms the NO-H-NO
strates.387 NiO is one of the few wide-bandgap p-type semi-
complex.392,393 In addition, the electrical activity of N can
conductors, has a bandgap of approximately 3.7 eV, and thus
be largely affected by the presence of various intrinsic
has potential applications in heterojunction devices compris-
defects which can act as efficient compensating centers and
ing n-type oxide semiconductors. The diodes exhibited recti-
also facilitate the formation of various N-related complexes,
fying ratios >108 at 63 V, with a built-in voltage was 1.4 V.
including donor complexes of ON or ZnI-NO.
The energy band diagram showed this to be a type-II hetero-
In Ga2O3, N-doped nanowires with a p-type electric
junction, with conduction band offset of 2.2 eV and valence
conductivity were reported.395 However, it can be difficult to
band offset (VBO) of 3.4 eV, respectively.387
definitely assign conductivity type unless several different
techniques are used to establish it and the lessons from ZnO
THEORY OF DEFECTS IN Ga2O3
that N acceptors are strongly compensated by the intrinsic
There have been a significant number of theoretical defects, and N atoms cannot be effective acceptors due to
studies of intrinsic defects and p-type dopants in b- their deep levels certainly introduce a note of caution. Dong
Ga2O3.38,45,51,388–392 The electrical properties of all wide et al.391 reported first-principles calculations based on den-
band gap semiconductors, and especially oxides, depend crit- sity functional theory (DFT) to study the compensation
ically on the concentration and charge state of defects. As mechanism and interaction of dopants and native defects in
we discussed earlier, b-Ga2O3 has intrinsic n-type conductiv- N-doped b-Ga2O3 with native defects. The introduction of N
ity because of the presence of unintentional dopants like Si was found to expand the lattice due to the larger radius
and other donor impurities such as H. The dopants determine of N3. The band structure calculations found that N dopants
whether the current (and, ultimately, the information proc- act as deep acceptors and cannot be effective p-type
essed by the device) is carried by electrons or holes. In semi- dopants.391 They also found that N formed a variety of com-
conducting oxides, it is generally possible to achieve one or plexes with native defects, including oxygen and Ga
vacancies and interstitials.391 Interstitial Ga and O vacancies

were found to be energetically favorable under Ga-rich con-
dition and that the presence of these in N-doped material
could compensate any p-type conduction.388
Xiao et al.390 carried out first-principles spin-polarized
density functional theory calculations to examine the elec-
tronic and magnetic properties of N-doped b-Ga2O3. They
found the spin-polarized state was stable with a magnetic
moment of 1.0 lB per nitrogen-dopant. The magnetic
moment mainly originated from the p-orbital of the nitrogen
and hole-mediated short-range p–p exchange was predicted
to cause the predicted ferromagnetism in this material.
Ma et al.389 performed first principles calculations to
look at the effect of Al addition to Ga2O3 to increase the
bandgap. They calculated the effect on absorption (blue-shift)
and the energetics of formation of defects in the presence of
Al addition. They found that O interstitials were likely to
form increasing n-type conductivity.389 Dong et al.388 exam-
ined the stability of O vacancies in b-Ga2O3 using the density
functional theory. These vacancies were deep donors and lead
to deep emission bands in PL spectra.
The work of Varley et al.38 compared the behavior of
holes in many wide bandgap oxides and found in Ga2O3 that
self-trapped holes form with very low mobility of these polar-
FIG. 24. (Top) Velocity-field characteristic of b-Ga2O3 at room temperature
ons (approximately 106 cm2 V1 s1), precluding the achieve- in three different directions. (Bottom) Calculated mobility (blue open trian-
ment of p-type conductivity. This certainly highlights the gles) for a wide temperature range (30 K–650 K). Experimental data shown
challenges of trying to synthesize p-type b-Ga2O3 with practi- as green closed circles. Reprinted with permission from J. Appl. Phys. 122,
cal conductivities and suggests that the experimental efforts 035702 (2017). Copyright 2017 American Institute of Physics.397
should focus on non-equilibrium approaches that minimize
other sources of background n-type conductivity and also maxi- Parisini and Fornari398 published a very detailed analy-
mize the solubility of the acceptors, perhaps to form bands. sis of the scattering mechanisms in n-type Ga2O3 and optical
phonons provide the main scattering mechanism, via lattice
LOW AND HIGH FIELD TRANSPORT IN Ga2O3 deformation. They compared their calculated transport prop-
Ghosh and Singisetti396,397 reported calculations of the erties to their own experimental values as well as those in
high-field transport in b-Ga2O3 using a combination of ab- the literature and found that the high-temperature experimen-
initio calculations and Monte Carlo simulations. They tal mobility behavior nearly follows the T3/2 power law,
examined the contribution of electron-phonon interactions which is the typical temperature dependence for a classical
on the velocity-field characteristics in this material for elec- electron gas limited by non-polar acoustic (deformation
tric fields ranging up to 450 kV/cm in different crystal direc- potential) scattering.398 Very good agreement with experi-
tions. The latter is important because of the known mental transport data was obtained assuming polar phonons
anisotropy of properties in b-Ga2O3.394–398 Figure 24 (top) of energy approximately 100 meV, and phonons with energy
shows the calculated velocity-field curves in three different approximately 30 meV that undergo non-polar interaction
Cartesian directions in the crystal.397 The velocity saturation with carriers.398 Ghosh and Singisetti396,397 also calculated
and negative differential conductivity in b-Ga2O3 result the impact ionization coefficient for avalanche breakdown, a,
from intra-valley short-range scattering, which contrasts to as being given by a ¼ a exp(b/E), with a ¼ 2.5 106/cm
the cases of both GaAs and GaN where these result from and b¼ 3.96 107 V/cm.396,397
inter-valley scattering.396,397 The calculated average peak The reported anisotropy of electron mobility by some
velocity at a typically encountered electric field of 200 kV/ groups as well as some of the optical properties398–402 has
cm is approximately 2 107 cm/s, which is just slightly been ascribed to an anisotropy in the conduction band effec-
smaller than in GaN. These same authors reported low field tive mass, but this is not consistent with some calculations
electron mobility calculations in this material from first that show a nearly isotropic conduction band. Kang et al.403
principles396 and investigated the interaction between elec- used first principles calculations to examine whether
trons and polar optical phonon modes. At low electric fields electron–phonon scattering could show significant anisot-
strengths, these interactions controlled the electron mobility, ropy in the monoclinic lattice. Another important aspect of
which was found to be 115 cm2 V1 s1, in good agreement transport is the dependence of the mobility on the carrier
with experiment.398 Figure 24 (bottom) shows the fit of the density. Their results indicated that the strong dependence
calculated mobilities as a function of temperature396 to of electron mobility on the electron density, which is more
experimental data.398 pronounced than in other oxide semiconductors such as
In2O3 and ZnO and the anisotropy in mobility (with mobili-

ties differing by approximately 20 depending on the direc-
tion) is not intrinsic to b-Ga2O3, but is a result of the
presence of a high density of extended defects such as twin
boundaries. This is easily experimentally testable in state-of-
the-art EFG crystals, which are available without significant
densities of twins. Kang et al.403 also found that polar longi-
tudinal optical (LO)-mode electron–phonon interactions con-
trol mobility at low carrier densities. The phonon spectrum
was anisotropic, but did not lead to appreciable anisotropy in
mobility.403
As reported in Tsao et al.,321 experimental mobilities in
n-doped (Si and Sn-doped) b-Ga2O3 are typically
100–200 cm2 V1 s1 at room temperature, 500 cm2 V1 s1
at 100–200 K in lightly doped bulk material and 5000 cm2
V1 s1 at 80 K in epitaxial layers. It is generally accepted as
discussed above that mobility is limited by optical phonon
scattering at high temperature and ionized impurity scatter-
ing at low temperature.399–401 The thermal conductivity for
the cases of Sn and Fe-doping has also been determined and
compared to the case of undoped b-Ga2O3, as shown in Fig.
25.404 The thermal conductivity in the [001], [100], [010],
and [201] directions were measured from 80 to 495 K. The
lowest thermal conductivity at 300 K was along the [100]
(10.9 W/m K), while the highest was along the [010] (27 W/
m K).404 The thermal conductivity of the Sn-doped and Fe-
doped b-Ga2O3 was lower than undoped material due to the
enhanced phonon-impurity scattering contribution,404 while
the anisotropy in conductivity results from difference in
sound velocities, the anharmonicity in phonon spectrum, and
the directional dependence of Debye temperature.404
There has been little work to date on transport in b-
Ga2O3-based alloys and heterostructures, and the properties
of 2D electron gases in heterostructures will be of interest
for transistor structures.321,322 Ternary alloys of b-(AlxGa1-
405–408
x)2O3/Ga2O3 have been reported, but the growth is still
being optimized. It was established that the phase stability
limit of Al2O3 in Ga2O3 was <18%.405 Similarly, first-
principles calculations408–410 have indicated that b-(Ga1-
xInx)2O3 can be stable for low concentrations of In. Clearly,
as the materials technologies progress, there will be more FIG. 25. Temperature dependence of the thermal conductivity for different
studies of transport in both InGaAlOx epitaxial layers over a orientations of undoped (a), Sn-doped (b), and Fe-doped (c) Ga2O3.
Reprinted with permission from J. Appl. Phys. 121, 235104 (2017).
range of compositions. Copyright 2017 American Institute of Physics.404
BAND STRUCTURE orientations are needed to deduce accurate band-gap values

As we discussed earlier, the valence band maximum in from absorption measurements.411 It was found that the fun-
b-Ga2O3 is mainly composed of weakly interacting oxygen damental band gap is indirect,411–413 but the minimum direct
2p orbital states with contribution of gallium 3d and 4s orbi- gap is only 29 meV higher in energy, causing the strong
tals,3 while the conduction band minimum is mainly com- near-edge absorption.411 They predicted that deep-UV lumi-
prised from the gallium 4s states.3 Holes in the valence band nescence is possible at sufficiently high excitation powers.411
are localized by lattice distortions to form polarons, while Li et al.414 combined hard X-ray photoelectron spectros-
the conduction band is fairly isotropic with an effective mass copy (HAXPES) and first-principles band structure calcula-
for electrons of around 0.3 mo.3 Mengle et al.411 carried out tions to investigate the valence properties and confirmed the
first principles calculations of the near-edge optical proper- main contribution to its formation as being from Ga 4sp states.
ties of b-Ga2O3 explained the broad range of experimentally Onuma et al.412 used polarized reflectance spectra of b-
reported band-gap values (4.4–5.0 eV at room temperature) Ga2O3 crystals measured as a function of temperature to
as resulting from the optical anisotropy of the crystal. They determine the energies of the absorption edge as 4.48 eV,
pointed out that linearly polarized light and defined crystal 4.57 eV, and 4.70 eV, respectively, for an electric field vector
of incidence light parallel to the c-axis (E//c), a*-axis (E//

a*), and b-axis (E//b). The absorption coefficient a for the
direct transitions from the upper most valence band to the
conduction band minimum at C was 102 cm1
a
103 cm1, but increased rapidly with increasing transition

energy (a > 105 cm1 for E > 5 eV). They again emphasized
that the optical anisotropies were the cause of differences in
the reported Eg for b-Ga2O3. Exciton energies were
179–268 meV at 5 to 300 K,412 while the longitudinal optical
(LO) phonon energies were in three ranges of 35–48, 70–73,
and 88–99 meV. The large exciton changes with temperature
result from the exciton-LO-phonon interaction.412
Ghose et al.415 used the Cauchy dispersion relation to
calculate the bandgap of epitaxial films grown by PA-MBE
on sapphire as 5.02 eV. The refractive index of the films is
1.89 at 6328 Å.415 The observed intrinsic broad optical
emission of gallate host lattices has been attributed to a self-
trapped exciton416 or a charge transfer-type mechanism.417
Binet and Gourier107 pointed out that donors in Ga2O3
exhibit impurity band conduction and motionally narrowed
electron spin resonance (ESR) signals, but also blue lumines-
cence characterized by strong electron–phonon coupling.
They suggested that the blue luminescence would result
from the fast recombination of an exciton trapped at an
acceptor site after a rate determining tunnel capture of an
electron from a donor cluster.32,107 b-Ga2O3 typically exhib-
its up to three different emissions, UV, blue and green,
which are dependent on growth type and purity, except for
FIG. 26. (Top) Luminescence spectrum of an as-grown undoped single crys-
the UV emission which is oftern attributed to recombination tal under excitation at 4.67 eV and excitation spectrum of the blue emission,
of a self-trapped exciton.32,107,360,412,417,418 The green emis- observed at 2.95 eV. Also shown is the absorption spectrum. All the spectra
sion is related to impurities, while the blue luminescence is were recorded at room temperature. (Bottom) Time evolution of lumines-
generally seen only in conducting samples and may be cence spectrum of an as-grown single crystal at room temperature after a
laser pulse at 4.67 eV: (a) t ¼ 0.5 ls; (b) t ¼ 0.75 ls; (c) t ¼ 1 ls; (d) t ¼ 2
related to oxygen vacancies.107 This emission can be selec- ls; (e) t ¼ 5 ls; (f) t ¼ 7 ls; (g) t ¼ 10 ls; (h) t ¼ 20 ls; and (i) t ¼ 50 ls.
tively excited with less bandgap light at low temperatures. Reprinted with permission from L. Binet and D. Gourier, J. Phys. Chem.
Both UV and blue emissions are broad and have strong Solids 59, 1241 (1998). Copyright 1998 Elsevier.107
Stokes shifts, characteristic of a strong electron–phonon
Si, Zr, or Hf doped b-Ga2O3 and ascribed to conduction elec-
coupling.
trons trapped around oxygen vacancies. Figure 27 shows a
Figure 26 (top) shows the emission at room temperature
single narrow ESR line, measured a t room temperature,
from undoped b-Ga2O3 for sub bandgap excitation at
with an anisotropic g value of approximately 1.96 observed
4.67 eV (266 nm). The peaks at 2.85 eV are in the blue range
and can be excited in three overlapping bands, indicated by
arrows in Fig. 26.107 The Stokes shift is 1.7 eV. The bottom
of Fig. 26 shows the time dependence of the PL after differ-
ent delays after excitation. For short delays (t
1 ls), a fast
decaying UV component increases the blue emission. This
UV peak disappears for long delays. Harwig et al.418
reported decays with a characteristic time of 120 ls. The
kinetics is not exponential, i.e., it is not driven by the transi-
tion probability of the recombination center.107
EPR OF Ga2O3
Electron spin resonance (ESR) and Electron
Paramagnetic Resonance (EPR) techniques are powerful
tools for exploring the properties of impurities and defects in
Ga2O3.354,419,420 Yamaga et al.40 examined valence, occupa-
FIG. 27. The ESR spectrum observed for Si4þ doped b-Ga2O3 at room tem-
tion site, and symmetry of Si, Zr, Hf, Cr, and Mn dopants in
perature, with a magnetic field parallel to the b axis. Reprinted with permis-
b-Ga2O3 using ESR. A single narrow ESR line with an sion from Yamaga et al., J. Non-Cryst. Solids 358, 2548 (2012). Copyright
anisotropic g value (approximately 1.96) was observed for 2012 Elsevier.420
in Si doped b-Ga2O3 with a magnetic field B parallel to the b single crystal ZnO to a 2H plasma. The incorporation depths
axis.420 The fine structure observed in the ESR spectrum for are much larger than in other wide bandgap semiconductors
Mn or Cr doped b-Ga2O3 suggested that Mn2þ and Cr3þ ions such as GaN, where a similar process leads to incorporation
substitute for octahedral Ga3þ ions in the lattice without depths of less than 1 lm.419,420,422,425 Hydrogen was found
charge compensators.419 The ESR lines were not observed to be a shallow donor in ZnO and to play a significant role in
for Si, Zr, or Hf doped b-Ga2O3, indicating that Si, Zr, and the background n-type conductivity.438,441 Similarly, the role
Hf are non-paramagnetic with a valence of þ 4.419 played by hydrogen impurities in the conductivity of indium
Kananen et al.355,419 used EPR to identify Ga vacancies oxide (In2O3) was controversial.440,448 Some studies, based
in Ga2O3, as discussed earlier. The V2Ga exhibited a hyper- on the effect of oxygen partial pressure in growth or anneal-
fine structure due to equal and nearly isotropic interactions ing environments,438,446,447 argued that oxygen vacancies
with the 69,71Ga nuclei at two Ga sites (the hyperfine parame- were the cause of the conductivity of In2O3. However, both
ters are 1.28 and 1.63 mT for the 69Ga and 71Ga nuclei, theoretical and experimental work found that hydrogen cen-
respectively).355 This same group used EPR to characterize ters can be important shallow donors in In2O3.422,436,447
neutral Mg acceptors (Mg0Ga) in Mg-doped b-Ga2O3.419 Muon-spin-resonance experiments found that implanted
These acceptors were observed after x-ray irradiation at muons, whose properties mimic those of hydrogen, form
77 K. This forms neutral acceptors when holes are trapped at shallow donors in In2O3.443–445 In2O3 thin films containing
singly ionized Mg acceptors (MgGa), i.e., the hole is local- hydrogen show n-type conductivity with high mobility.438 It
ized in a nonbonding p orbital on a threefold-coordinated was theoretically predicted that interstitial hydrogen (Hiþ)
oxygen ion adjacent to an Mg ion at a sixfold-coordinated and hydrogen trapped at an oxygen vacancy (HOþ) were
Ga site.419 The Mg0Ga acceptors converted back to the non- shallow donors, giving rise to n-type conductivity or com-
paramagnetic MgGa charge state above 250 K.419 They also pensate acceptors in In2O3.436
observed electron traps due to the unintentional impurities The role of hydrogen impurities on the conductivity of
Fe3þ (3d5) and Cr3þ (3d3).419 In2O3 single crystals was studied using IR spectroscopy and
theory.437,438 The annealing of In2O3 crystals in an H2 or D2
HYDROGEN IN Ga2O3 ambients at 500 C produced a thin conducting layer near the
surface with thickness 0.06 mm and with a carrier concen-
There is a particular interest on the properties of hydro- tration determined by Hall measurements of 1.6 1019 cm3.
gen in wide bandgap semiconductors, in general,421–457 and An OH vibrational line at 3306 cm1 was assigned to the
in Ga2O3, in particular,433,443–446,456,457 because of the pre- interstitial H shallow-donor center that is responsible for the
dictions from density functional theory and total energy cal- hydrogen-related conductivity.438,447 The corresponding Di
culations that it should be a shallow donor.51,437,438 The center had an OD line at 2464 cm1. The Hi center was stable
generally observed n-type conductivity, therefore, may at up to 500 C. Investigations of hydrogen shallow-donor cen-
least in part be explained by the presence of residual hydro- ters in ZnO and SnO2 found that Hi is only marginally stable
gen from the growth ambient, rather than to native defects at room temperature and HO is a more thermally stable donor
such as Ga interstitials or O vacancies, which are suggested that dominates the n-type conductivity of hydrogenated sam-
to be deep donors.51 There is some experimental support to ples of these materials stored for substantial times at room
the fact that hydrogen may be a shallow donor in Ga2O3 temperature. The experimental results and complementary
from experiments on its muonium counterpart and from elec- theory showed that the conductivity produced by the thermal
tron paramagnetic resonance of single-crystal sam- treatment of In2O3 in hydrogen can be explained primarily by
ples.443–445 This would be similar to the case of ZnO, where a thermally stable Hi center, consistent with theoretical pre-
hydrogen is a shallow donor and contributes significantly to dictions that HO has a higher formation energy.438,447
the observed native n-type conductivity in that material in Interstitial hydrogen (Hi) and hydrogen at an oxygen
many cases.422 There is evidence that a number of transpar- vacancy (HO) behave as shallow donors in SnO2.437,438
ent conducting oxides have commonality of impurity Annealing SnO2 in a hydrogen-containing ambient gives rise
properties.435,438 to strong n-type conductivity and produces several
It is well known that atomic hydrogen can passivate vir- hydrogen-containing centers. Bates and Perkins449 in exam-
tually all impurities in semiconductors like Si, GaAs, and ining the vibrational properties of OH, OD, and OT centers
GaN through the formation of neutral dopant-hydrogen com- in TiO2 found single sharp vibrational lines, while Herklotz
plexes.421,422,437,438,447 In most semiconductors, hydrogen is et al.450 discovered a more complicated OH (or OD) spec-
a negative-U center, so in the p-type material, hydrogen is a trum with a multiline structure. Herklotz et al.450 proposed
compensating donor, whereas for n-type conductivity, hydro- that OH and OD centers in TiO2 are shallow donors with
gen is a compensating acceptor, and therefore, in these binding energies of 10 meV and that the multiline structure
materials, hydrogen always counteracts the prevailing con- is due to the thermal ionization of the OH or OD center.
ductivity.421,433,435,450 However, different behavior has been SnO2 and TiO2 both have the rutile structure and have inter-
found for conducting oxides.437,438 stitial H centers with similar structures.437,438 However,
In terms of what is known about hydrogen in other elec- interstitial H gives rise to an effective mass theory (EMT)
tronic oxides, it has been found previously that hydrogen shallow donor in SnO2 (and In2O3) with strong free carrier-
diffuses extremely rapidly in ZnO,425,431 producing incorpo- absorption and to a small polaron in TiO2 that leads to a dis-
ration depths of almost 30 lm for a 0.5 h exposure of bulk, tinctive OD vibrational absorption spectrum.438 The origin
of the conduction band states is critical in determining the switching speeds as well as reliability issues due to local
behavior of donor states in these materials, being s- and p- contact heating during current flow during device operation.
like in SnO2 (and In2O3) and having substantial d-character Wide bandgap semiconductors typically have large barrier
in TiO2. These differences cause delocalization of the elec- heights for metals deposited on them.458–463 To decrease the
trons in effective-mass-like states in SnO2 and In2O3 and to barrier height and/or width, the insertion of an intermediate
be self-trapped at Ti in TiO2.437,438 semiconductor layer with a smaller band gap and/or a higher
We have done some preliminary work on determining doping concentration is one approach to improve carrier
the properties of hydrogen in Ga2O3. The first set of experi- transport across the interface.464–490 In other wide bandgap
ments involved implantation of 2H ions into bulk Ga2O3 and semiconductors, Ohmic contacts are typically realized by
subsequent annealing to measure the thermal stability. As metal-semiconductor structures in which the potential barrier
shown in Fig. 28 (top), the deuterium evolves from the sam- is reduced by heavily doping the semiconductor under the
ple for annealing up to 650 C, but the remainder at each metal, which enhances electron tunneling.458–463 The doping
temperature decorates the residual ion-induced damage. The levels achievable in most wide bandgap semiconductors are
fraction of the remaining implanted deuterium is shown at usually well below of what can be obtained in Si, and com-
the bottom of the figure and corresponds to a lower activa- bined with the lower electron affinity of these materials, the
tion energy than in materials like GaN. The temperature at Schottky barrier heights of most metals are very high
which deuterium is retained is also higher when it is (> 2 eV).321 This leads to the unacceptably high contact
implanted compared to introducing it by diffusion from a resistances. The usual approaches to mitigate these issues in
plasma. This is a result of the deuterium trapping onto resid- oxides involve surface etching or cleaning to reduce barrier
ual implant damage in the former case. height or increase of the effective carrier concentration of
the surface through preferential loss of oxygen. Another
OHMIC CONTACTS TO Ga2O3 approach used in other semiconductors has been to use com-
positionally graded layers specifically designed to lower the
The achievement of acceptable device characteristics for barrier to electron transport.321,322
any Ga2O3-based structure relies heavily on developing low It is desirable to have reliable Ohmic contacts capable
specific contact resistance Ohmic metallization schemes. of withstanding high temperatures to maximize process flexi-
Additional contact resistance leads to slower device bility and also the range of device applications. In addition,
in devices like rectifiers, their resistance must be negligible
compared to the semiconductor drift layer, in order to mini-
mize the device specific on-resistance (Ron) and, hence, the
power losses of the system. Typically, electronic devices
based on wide bandgap semiconductors require specific con-
tact resistance (qc) values in the range of 105–106 X
cm2.321,322
For an n-type semiconductor, to achieve an Ohmic con-
tact means that the work function of the metal must be close
to or smaller than the electron affinity of the semiconduc-
tor.460 The electron affinity of b-Ga2O3 is reported to be
4.00 6 0.05 eV,481 and thus potential choices if the surface is
unpinned include Hf (work function 3.9 eV), Sc and La (both
3.5 eV), and Gd (2.9 eV). It will be interesting to see if these
are effective on n-type b-Ga2O3, most likely with bilayers of
these with Au to reduce the contact sheet resistance. The
transport mechanisms in the contacts should then be deter-
mined by the temperature-dependent transmission line
method (TLM) or circular-TLM (CTLM) measurements.
The contact resistance is an important parameter charac-
terizing the metal/semiconductor interfaces.460 It is the total
resistance of the metal/semiconductor junction (expressed in
X) and, hence, it depends on the area of the contact. A more
useful physical parameter describing the performance of
Ohmic contacts is the specific contact resistance qc, which is
independent of the contact geometry, and is typically
expressed in X cm2. The relation between the contact resis-
tance and the specific contact resistance in a metal/semicon-
ductor contact can be viewed analogously to that between
FIG. 28. Profiles of 2 H implanted Ga2O3, subsequently annealed at different
the resistance and the resistivity in a resistor.
temperatures to measure the thermal stability (top) and percentage of 2 H In Ohmic contacts, the specific contact resistance qc
remaining as a function of anneal temperature (bottom). depends on the metal/semiconductor Schottky barrier height
UB and on the doping concentration of the semiconductor. resistance. Table III shows a summary of results from the lit-
For an n-type semiconductor, according to the Schottky– erature. Specific contact resistances of approximately
Mott theory, the metal/semiconductor Schottky barrier 8 106 X cm2 were reported for Ti/Au source-drain con-
height UB obeys the relation.458–460 tacts on n-Ga2O3 epitaxial layers in which Si was implanted
and annealed at 925 C, followed by dry etching, metal depo-
eUB ¼ eUm vs ; sition, and annealing at 470 C.467 Other reports with a simi-
lar process have achieved 4.6 106 X cm2 for Ti/Au
where Um is the work function of the metal, and vs is the
contacts on Si-implanted epitaxial layers.464
electron affinity of the semiconductor.
Oshima et al.470 used the inter-layer approach, but
For an n-type semiconductor, depending on the semi-
pointed out that the often-used In- and Ti-based electrodes
conductor doping level ND, different mechanisms of carrier are often utilized as Ohmic contacts, are probably not opti-
transport at the interface can occur, owing to the different mal due to the low melting point of In of 157 C, while Ti
thickness of the space charge region W formed at the metal/ contacts degrade above 600 C due to oxidation reactions
semiconductor interface. Consequently, the specific contact with b-Ga2O3. We have found previously in GaN, for exam-
resistance qc will have a different dependence on the barrier ple, that there is usually a strong relationship between the
height UB, on the carrier concentration ND, and also on the electrical performance and interfacial microstructure of
temperature T. Ohmic contacts over different annealing temperatures. The
A traditional Ohmic contact on semiconductors follows formation of interfacial nitrides, for example, is found to be
a thermionic field effect (TFE) or field-effect (FE) mecha- critical in achieving Schottky-to-Ohmic transition as well as
nism depending on the semiconductor effective doping and/ excellent Ohmic contacts. In addition, the interdiffusion,
or temperature.458–460 If, for example, there was no signifi- intermixing, and metallurgical reactions within and between
cant dependence of the contact resistance on measurement the metal layers lead to the formation of second phase con-
temperature for the samples after annealing, this would indi- sisting of intermetallics, solid solutions, or precipitates.
cate that the dominant current transport mechanism is field Auger Electron Spectroscopy elemental profiling of the con-
emission. Thermionic emission has a significant temperature tacts for different anneal temperatures and times will help us
dependence, and thermionic field emission is operative at the to identify any metal reactions with the Ga2O3. These phases
doping range of 1016–1018 cm3. In thermionic emission, the can be examined in detail with TEM, microprobe, and XPS.
specific contact resistance is given by458–460 According to all the above considerations, metals with a
low Schottky barrier height are recommended to obtain an
k q/Bn Ohmic contact with a low specific contact resistance qc. It is
qC ¼ exp ;
qA T kT worth noting that most of the data reported in the literature
(Table III) refer to multilayers in which a capping layer is
where k is the Boltzmann’s constant, A* the Richardson con-
introduced over a Ti/Al stack (Ti/Al/Ni/Au, Ti/Al/Ti/Au, Ti/
stant, and T the measurement temperature, may be explained
Al/Pd/Au, and Ti/Al/Mo/Au). In GaN technology, a typical
by the tunneling process, which is given by460
Ohmic stack consists of a contact layer like Ti or Ta, an
" pffiffiffiffiffiffiffiffiffiffi # overlayer of Al, a barrier layer of a metal such as Ni, Ti, Pt,
2 eS m /Bn
qC exp pffiffiffiffiffiffi ; Pd, Mo, or Ir, and then a cap layer of Au to reduce the sheet
h ND resistance of the contact.
The presence of surface states on Ga2O3 is crucial to the
where q is the electronic charge, /Bn the barrier height, eS contact, since high density allows carriers to tunnel through
the semiconductor permittivity, m* the effective mass, h the the barrier. This was reported for Au/Ti on Si-implanted b-
Planck’s constant, and ND the donor density. In the FE Ga2O3.491 If there are a few surface states, it should be a
mechanism, the specific contact resistance of Ohmic contacts Schottky contact.490 Yao et al.475 also reported recently that
is strongly dependent on both the Schottky barrier UB and the surface states appear to have a more dominant effect on
the semiconductor doping ND level. For fixed values of the Schottky diode characteristics than the actual choice of the
barrier height UB and carrier concentration ND, the tempera- metal. Thus, the orientation of the wafer surface, which
ture determines the carrier transport mechanism through the defines both the Ga-to-O ratio and the density of dangling
metal/semiconductor interface. In particular, when the ther- bond states, makes a difference. Yao et al.475 investigated
mal energy kT is much lower, tunneling of the carriers by the the use of Ti, In, Ag, Sn, W, Mo, Sc, Zn, and Zr as electrical
FE mechanism is predominant. At higher temperatures, the contacts to n-type b-Ga2O3 substrates as a function of
carriers can pass through the barrier by the TFE mechanism. annealing temperature up to 800 C. Some metals displayed
Finally, at much higher temperatures, thermionic emission Ohmic (Ti and In) or near-Ohmic (Ag, Sn, and Zr) behavior
over the barrier becomes relevant for the current transport. over a particular range of conditions, but the contact mor-
To date, contact schemes involving indium zinc oxide phology was problematic.475 They concluded that the metal
(IZO) interlayers between the metal and the Ga2O3, dry etch- work function is not a dominant factor and the interfacial
ing in BCl3/Ar to enhance the surface n-type conductivity, reactions were key. Baik et al.492 found that Ti (200 Å)/Au
followed by Ti/Au,466 Ti/Al annealed at 500 C (Refs. 464 (1500 Å) metallization deposited on the two different orien-
and 465) or implantation of Si to increase surface conductiv- tations and annealed at 450 C showed Ohmic current-
ity,464,480 have been employed to achieve acceptable contact voltage (I-V) behavior for ð201Þ but rectifying characteristics
TABLE III. Summary of Ohmic contact properties on b-Ga2O3.
n-type doping
Metal stack (cm3) Anneal conditions qc (X cm2)/method Comments Reference
Ti(50 nm)/Au(300 nm) 3 1019 450 C for contacts 4.6 106, CTLM Si implanted to increase dop- Sasaki et al.469
(950 C for implant ing in contact region
activation prior to
metallization)
Pt(100 nm)/ITO(140 nm) 2 1017 800–1200 C for Pt/ Not measured, but Compared use of ITO inter- Oshima et al.471
ITO, 450–700 C for Ohmic for anneals layers to enhance Ohmic con-
Pt/Ti above 900 C tact formation
Ti/Al/Au (15/60/50 nm) 2.7 1018 None Contact resistance of Ar plasma pre-treatment to Zhou et al.485
2.7 X mm, TLM enhance conductivity under
contact
Ti(20 nm)/Au(230 nm) 3 1019 470 C, 60 s Not measured-field- Si implant annealed at 950 C Wong et al.486
plated MOSFET and BCl3 dry etch to enhance
surface conductivity
Ti(20 nm)/Au(230 nm) 7 1017 None Not measured-MBE BCl3/Ar RIE to enhance sur- Higashiwaki et al.487
grown MESFET face conductivity
Ti(20 nm)/Au(230 nm) 3 1019 470 C, 60 s 7.5 106 Si implant annealed at 950 C Wong et al.473
and BCl3 dry etch to enhance
surface conductivity
Ti/Al 1018 450 C, 60 s (1100 C, 2.1 105, TLM Diffusion of Sn from spin-on Zeng et al.474
5 min for Sn drive-in glass source, as well as BCl3/
prior to contact Ar RIE
deposition)
Ti/Al/Ni/Au 1019 470 C, 60 s Contact resistance of e-mode MOSFET Chabak et al.470
4.7 X mm, TLM
Ti/Au/ITO (10/20/80 nm) 1019 600 C 6.3 105, TLM Si implant activated at 950 C Carey et al.475
to enhance conductivity
Ti/Au/AZO (10/20/ 1019 400 C 2.8 105, TLM Si implant activated at 950 C Carey et al.477
80 nm) to enhance conductivity
Ti, In, Ag, Sn, W,Mo, Sc, 5 1018 400–800 C Not measured In and Ti produced linear Yao et al.476
Zn and Zr (20 nm), with I-Vs after anneals
Au (100 nm) overlayers
for (010). Since the net doping density is the same in both breakdown strength over 10 times larger and on-state resis-
samples, the surface chemistry determines the effective bar- tance (RON) approximately four hundred times lower at a
rier height and hence the dominant carrier transport given voltage.500–508 These characteristics make Ga2O3 devi-
mechanism.492 ces attractive for hybrid electric vehicles and power condi-
Carey et al.476,477 also reported the use of IZO or ITO tioning in large industrial motors.
interlayers to improve the contact resistance. Pt/ITO contacts As we discussed in “Ohmic Contacts to Ga2O3” section,
on n-Ga2O3 showed superior Ohmic contacts to Pt/Ti and this the interface-induced gap states, made up of both valence-
was attributed to the formation of an interfacial layer with band and conduction-band states in electronic oxides, often
lower bandgap and higher doping concentration than the play a determining role in the barrier heights of metal con-
Ga2O3 alone. The band alignment at the heterointerface is tacts.505,508–515 Some studies have reported a correlation
also critically important in determining the favorability of the between Schottky barrier heights of metals on Ga2O3 and the
carrier transport.480,481 Several authors have found that the electronegativity difference between the oxide and the
presence of upward band bending in low conductivity Ga2O3 metal.501,506,507 The presence of the upward band bending in
complicates the achievement of Ohmic contacts.464–469,481,484 low conductivity Ga2O3 complicates the achievement of
Ohmic contacts,495,505,508–510 and there was a strong depen-
dence of the properties of the barrier height on the energy of
SCHOTTKY CONTACTS TO Ga2O3
incident atoms during deposition.516,517 The latter implicates
Schottky rectifiers are attractive applications for b- surface defects and stoichiometry as being one factor in
Ga2O3 because of their fast switching speed, which is impor- determining the properties of the contact.
tant for improving the efficiency of the inductive motor con- Table IV summarizes the reported values for the
trollers and power supplies and also their low turn-on Schottky barrier height, ideality factor, and thermal stability
voltages compared to p-n junction rectifiers. Their break- of different metals schemes used for rectifying contacts on
down voltage is a strong function of bandgap.493–499 Ga2O3 b-Ga2O3. Most groups extract the effective barrier height
Schottky diodes have numerous advantages over more con- from the linear portions of the I-V characteristics of
ventional Si rectifiers, achieving a maximum electric field Schottky diodes, assuming the ideal thermionic-emission
behavior. However, capacitance-voltage (C-V) and internal electron affinity of Ga2O3 (4.00 eV).508–510 A similar esti-
photoemission (IPE) measurements are also employed for a mate for Pt would yield a value of 1.12–1.53 eV based on
complete understanding of the current transport mecha- the reported work functions of 5.12–5.93 eV. Many of
nisms.494–497,518–523 The typical barrier heights are in the the experimental values are consistent with the Schottky-
range of 1–1.4 eV. Note that different metals yield different Mott predicted values for the barrier height. However, the
barrier heights within a limited range, and so it is of interest presence of surface states will modify the resultant effec-
to examine possible Fermi-level pinning effects on Ga2O3 tive barrier heights and this may explain the fact that the
surfaces.499 The predicted barrier height for the Ni in the experimental value for some Ni Schottky diode barrier
absence of pinning would be the difference in the Ni work heights is higher than Pt, despite the fact that the Ni work
function and electron affinity of Ga2O3 from the Schottky- function is lower than Pt.501 The ns values reported the
Mott relation. This would translate to 1.04–1.35 eV,501 based range from very near the ideal value of 1 (unity) to approx-
on the reported work function of Ni (5.04–5.35 eV) and the imately 1.5. The reasons for ns values to be larger than 1
TABLE IV. Table of reported Schottky barrier contacts to b-Ga2O3.
Metal Barrier Height (eV) Ideality Factor Process/measurement condition Comments References
Ni 1.25 1.01 Evaporation, I-V Thermionic emission Oishi et al.499

Ni 1.05 Not measured Dc magnetron sputtering, I-V No photocurrent in forward bias Armstrong et al.503
Ni 1.08 1.19 Sputtering, I-V Piranha clean,BOE,water clean Jayawardena et al.507
Ni 1.55 1.04 E-beam deposition, I-V Similar values obtained with C-V, inter- Zhang et al.505
nal photoemission
Ni 1.07 1.3 E-beam Pt/Au, I-V Consistent with Schottky-Mott model Ahn et al.502
Ni 0.95 3.38 E-beam Ni/Au, I-V Epi layer, barrier height increased with Oh et al.504
temperature
Ni 1.04 1.3 E-beam, I-V and C-V Much larger barrier height (1.61eV) with Yao et al.500
C-V
Ni 0.8–1.0 1.8–3.2 Evaporation, I-V (AlxGa1-x)2O3 with x up to 0.164. Ahmadi et al.407
Barrier height increases slightly with Al
content
Ni 0.97–1.22 Not measured Evaporation, I-V and C-V Cleaved surface, higher barrier measured Irmscher et al.496
with C-V
Ni 1.54 1.04 E-beam deposition, I-V Similar values obtained with C-V, inter- Farzana et al.515
nal photoemission
Pd 1.27 1.05 E-beam deposition, I-V Similar values obtained with C-V, inter- Farzana et al.515
nal photoemission
Pt 1.58 1.03 E-beam deposition, I-V Similar values obtained with C-V, inter- Farzana et al.515
nal photoemission
Au 1.71 1.09 E-beam deposition, I-V Interface consistent with inhomogeneous Farzana et al.515
barrier
Au 1.04 1.01 E-beam deposition, I-V electron affinity Ga2O3 4.00 6 0.05 eV, Mohamed et al.516
work function of Au 5.23 6 0.05 eV
Au 0.98 1.09 E-beam, I-V Significant degradation above 200 C Suzuki et al.517
Pt 0.71–1.1 1.41–1.61 Thermal evaporation or sputtering, I-V Long throw sputtering generally better Muller et al.518
PtOx 1.09–1.34 1.52–2.18 Sputtering,I-V Long throw sputtering generally better Muller et al.518
Pt 1.39 1.1 Pt/Ti/Au sputtering, I-V Barrier height stable to at least 150 C He et al.513
Pt 1.46 Pt/Ti/Au evaporation, I-V Barrier height may have been enhanced Konishi et al.520
by presence of F at interface
Pt 1.04 1.28 E-beam Pt/Au, I-V Consistent with Schottky-Mott model Ahn et al.502
Pt 1.35–1.47 Evaporation of Pt/Ti/Au, I-V, C-V Sasaki et al.498
Pt 1.01 1.07 E-beam, I-V Thermionic emission Tadjer et al.514
TiN 0.98 1.09 ALD at 350 C, I-V Thermionic emission, degrades above Tadjer et al.519
170 C
Cu 1.32 1.03 Sputtering, I-V Stable up to at least 250 C Splith et al.509
Cu 1.13 1.53 E-beam, I-V and C-V Optimal cleaning was organic solvent, Yao et al.500
HCl and H2O2 and DI water rinse
W 0.91 1.4 E-beam, I-V and C-V Higher barrier heights for (010) cf. Yao et al.500
(201)
Ir 1.29 1.45 E-beam, I-V and C-V Not a strong correlation of barrier height Yao et al.500
with work function
Pt 1.05 1.4 E-beam, I-V and C-V Larger barrier heights in C-V Yao et al.500
Pt 1.09–1.15 1 Evaporation of Pt/Ti/Au, I-V and C-V Richardsons’s constant A* of Higashawaki et al.501
28–41 A/cm2 K2
are generation–recombination in the depletion region, WET ETCHING OF Ga2O3

thermionic field emission tunneling, and edge leakage,
There is generally a need to pattern Ga2O3 when fabri-
which tend to degrade the electron device performance
cating devices such as UV solar blind photodetectors, vari-
characteristics.493,505–510
ous types of transistors, as well as sensors. The patterning is
Farzana et al.514 carried out a systematic study of
carried out by etching the semiconductor, using dielectric or
Schottky barriers fabricated on (010) b-Ga2O3 for Pd, Ni, Pt,
photoresist masks to protect the active areas. There are two
and Au Schottky diodes. The diodes exhibited nearly ideal I-
basic classes of etch processes, those carried out in the liquid
V characteristics with the ideality factors from 1.03 to 1.09.
phase (known as wet-etching) and those performed in the gas
Thermionic emission was the dominant mechanism for Ni,
phase (called dry etching, especially when a plasma is used
Pt, and Pd. The barrier height depended on the metal, rang-
to provide the reactive species for etching). Etch processes
ing from 1.27 V for Pd, 1.54 V for Ni, 1.58 V for Pt, and
may be classified by their rate, selectivity, uniformity, direc-
1.71 V for Au.514 While most of the metals showed a good
tionality (isotropy or anisotropy), surface quality, and repro-
agreement between I-V, C-V, and IPE measurements, for Au
ducibility. All etching processes involve three basic events:
a large variation was noted between the methods. This may
(i) movement of the etching species to the surface to be
indicate the presence of a more complex interface for Au/
etched, (ii) chemical reactions to form a compound that is
Ga2O3. Figure 29 shows the dependence of the barrier
soluble in the surrounding medium, and (iii) movement of
height on the metal work function.499 The lack of a strong
the by-products away from the etched region, allowing fresh
correlation indicates that these metal-(010) b-Ga2O3 inter-
etchant to reach the surface. Both (i) and (iii) usually are
faces are not fully pinned. It is also important to note that
referred to as diffusion, although convection may be present.
the literature will always contain a small number of outlier
The slowest of these processes primarily determines the etch
results due to experimental control issues, but the general
rate, which may be diffusion or chemical-reaction limited.
trend to date is that the barrier heights are not controlled by
While not as chemically resistant as sapphire (Al2O3),
surface Fermi level pinning, that the deposition method
Ga2O3 is still very difficult to wet etch, and there are only a
makes a difference through its effect on the energy of the
few reports on wet etching of thin films and single crystal
incident species arriving at the surface, the effectiveness of
substrates.
the surface cleaning process, and the orientation of the
Wet etching is performed by immersing the wafers in an
Ga2O3 surface all play a role in determining the effective
appropriate solution or by spraying the wafer with the etch-
barrier.499,501,508–510,518–521
ant solution. Wet-chemical etching is superior to dry etching
Muller et al.522 examined both Pt and PtOx as Schottky
in terms of effectiveness, simplicity, low cost, low damage
contacts, deposited by thermal evaporation, and various sput-
to the wafer, high selectivity and high throughput. However,
tering methods. In general, they and others520–523 have found
the main limitations of wet etching include its generally iso-
that oxygen-containing ambients produce better Schottky
tropic nature, which results in roughly equal removal of
characteristics on oxides like Ga2O3 and that the kinetic
material in all directions, making it incapable of patterning
energy of the impinging metal species on the oxide surface
sub-micron features, and the need for disposal of large
plays a role in defining the quality of the rectifying contacts.
amounts of corrosive and toxic materials. It is possible to
This is an important practical consideration when trying to
achieve anisotropy with wet etching in specific cases where
fabricate rectifiers with reproducible and optimized break-
there is a strong dependence of the etch rate on crystal orien-
down voltages.
tation, such as KOH etching of crystalline silicon. However,
the inability to pattern sub-micron features and the need for
a high degree of uniformity and reproducibility limit the
application of wet etching for most microelectronic devices.
One of the major needs at the moment for Ga2O3 is the
lack of high quality patterning processes that exist for the
more mature semiconductors. As in materials like GaAs and
GaN, etching is needed for intra-device isolation or for
exposing sub-surface layers for the fabrication of Ohmic
contacts. A number of wet etch solutions have been reported
to work for Ga2O3, including HNO3/HCl, H2SO4, H3PO4,
and HF.492,524–537 Wet etching of b-Ga2O3 in an HF solution
can be attributed to the following reaction:526
Ga2 O3 þ 6HF ! 2GaF3 þ 3H2 O:
The etch rate in acids and bases strongly depends on the

FIG. 29. Schottky barrier heights as a function of metal work function for crystal quality. For example, in sputtered Ga2O3 films depos-
Schottky diodes on (201) bulk and (010) epitaxial b-Ga2O3. Schottky barrier ited at substrate temperatures from 400 to 1000 C, the etch-
height values on (201) Ga2O3. The Schottky–Mott predicted line is calcu-
lated based on the Schottky–Mott model. Reprinted with permission from
ing rate of gallium oxide grown at 400 C was about 490 nm/
Yao et al., J. Vac. Sci. Technol. B 35, 03D113 (2017). Copyright, 2017 s, while for a 1000 C film, the etching rate was about
American Vacuum Society.500 0.196 nm s1.536 Table V shows a summary of the wet etch
TABLE V. Summary of wet etch results for b-Ga2O3.
Etch solution Sample description Etch rate (Å min1) Comments Reference
H3PO4 Float zone (100) orientation 700 at 150 C 85 wt. % phosphoric, no etching at 25 C, Oshima et al.528
activation energy 20.2 kCal mol1
H2SO4 Float zone (100) orientation 200 at 150 C
97 wt. % sulphuric, no etching at 25 C, acti- Oshima et al.528
vation energy 26.3 kCal mol1
HF Undoped, Float zone (100) orientation 10 at 25 C
47% HF Ohira and Arai527
HF Undoped, Float zone (001) orientation 5 at 25 C Rate approximately half that of (100) Ohira and Arai527
orientation
HF Sputtered, with substrate temp 2.94 105 at 25 C (400 C films) 49% HF Ou et al.537
400–1000 C to 120 (1000 C films)
NaOH Undoped, Float zone (100) orientation 1 at 25 C 20% NaOH Ohira and Arai 527
HNO3 Undoped, Float zone (100) orientation 14 at 120 C 60.5% HNO3 Ohira and Arai527
HCl/H2O MBE thin films deposited at 100 or 1000 for high temp films, 8000 1:3 ratio of HCl:H2O, activation energy Ren et al.526
535 C for low temp films 14.1 kCal mol1
HF E-Beam deposited thin films grown at 700–2400 depending on growth 10% HF, etch rate lower for films grown at Passlack et al.525
40–350 C temperature higher temperatures or annealed postgrowth
HCl MIST grown films 300 at 60 C 32% HCl Dang et al.538
rates reported for Ga2O3.524–537 It is worth noting that amor- and 4-fold-coordinated O(III) with 1 and 3 dangling bonds.
phous Ga2O3 is a component of the native oxide on GaAs The atomic densities can be calculated for which the parame-
wafers, and it must be always chemically removed prior to ters were determined from plots of the crystal structure. The
deposition of contacts on that material. number of dangling bonds for Ga or O can then be calculated
For photoelectrochemical (PEC) etching,492,528 both in a similar fashion, using the number of bonds formed to
ð
201Þ and (010) Ga2O3 single crystal wafers were immersed each atom and the number of atoms per unit cell. For (010)
in a 5 M potassium hydroxide (KOH) solution for 30 min at Ga2O3, there exist 2 types of surfaces with Ga atomic density
a stirring rate of 300 rpm at 80–95 C with ultraviolet illumi- of 0.58 1015 cm2 and O of 0.87 1015 cm2. The dan-
nation using a 120 W Hg lamp.528 The photochemical etch- gling bond densities of Ga and O atoms on for both (010)
ing rate in KOH solutions of ð201Þ oriented, n-type bulk surfaces are estimated to be the same: 0.58 1015 cm2 and
single crystals grown by the edge-defined film-fed growth 0.87 1015 cm2 for Ga and O atoms, respectively. For the
method is approximately 3–4 times higher than for the (010) two types of ð201Þ surfaces, the Ga and O atomic densities
planes.492 The activation energy for etching was 0.498 eV are 0.89 1015 cm2 and 1.34 1015 cm2, respectively. If
and 0.424 eV for ð201Þ and (010) orientations, respec- the surface is terminated with O, the dangling bond densities
tively,492 suggesting that etching is reaction-limited with the of O are 1.78 and 2.68 1015 cm2 for type I and II, respec-
same rate-limiting step. This energy is characteristic of that tively. Figure 30 (top) shows an SEM image of the etched
expected for the reaction-limited etching, whose other char- surface with Miller plane indices, while the bottom of the
acteristics include a linear increase in the etch depth with figure shows the crystal structure of (115), ð115Þ, and (010)
time and an independence of the etch rate on solution agita- planes of Ga2O3. For the etch pits observed after Pec etching,
tion. SEM images of the surface morphology of the two each side of the isosceles triangle with an angle of 70 was
types of samples after being etched for 120 min at 95 C found to be (115) and ð115Þ planes, and the base is (010)
showed triangular shapes formed on the surface of the ð201Þ plane. The atomic bond configurations of the cleaved (115)
Ga2O3 after KOH PEC etching. By contrast, the (010) surfa- and ð115Þ planes demonstrate that the Ga and O atoms coex-
ces maintained flatness and smoothness after etching at a ist on those planes. In the case of Ga-O coexisting surface,
depth of approximately 10–19 nm. Thus, it is necessary to OH ions may form Ga-O compounds with Ga dangling
examine the difference in atomic configurations in detail for bonds, but these are not easily soluble due to strong Ga-O
both ð201Þ and (010) Ga2O3 crystal surfaces. bonding in the plane and the repulsive force between Ga-O
It is clear that the atomic arrangement for the (010) and bonds and OH ions. Consequently, wet etching may be
ð
201Þ planes are different, leading to different atomic config- impeded on (115), ð115Þ, and (010) planes in Ga2O3, thus
urations and dangling bond densities on a particular crystal resulting in chemically stable surfaces in KOH wet chemi-
orientation. The (010) plane consist of Ga (I) atoms in the cals under UV illumination.
tetrahedral site and Ga(II) in the octahedral site, and O atoms The number of dangling bonds on the surface is believed
in a distorted octahedral arrangement. It is notable that the to control the etching behavior of the different orientations
respective Ga and O atoms have different numbers of dan- of Ga2O3. For example, on an oxygen-terminated surface,
gling bonds depending on the sites of the Ga2O3 surface, as the oxygen atoms which exist on the surface after removing
follows: 4-fold-coordinated Ga(I) with the number of dan- the first Ga-layer by OH ions have dangling bonds. In
gling bonds of 1 and 3, 6-fold-coordinated Ga(II) with 2 and KOH-based wet etching, it was suggested that Ga atoms
3 dangling bonds, 3-fold-coordinated O(I) with 1 dangling react with OH to also form Ga2O3, which subsequently dis-
bond, 3-fold-coordinated O(II) with 1 and 2 dangling bonds, solved in base solutions. In H3PO4-based etching, the Ga2O3
aluminum in Cl2 gas (when there is no native oxide to inhibit

the etching), or etching of Si using a fluorinated gas such as
XeF2. By far the most important advantage of dry etching
over wet etching is that it provides a higher resolution trans-
fer of a pattern by having anisotropic removal rates of the
exposed material, i.e., it etches much faster in the vertical
direction than in the lateral direction.538 Dry etching also
generally has less chemical waste (although the exhaust
gases must be scrubbed). As discussed earlier, for
manufacturing of microelectronics, the processing must be
automated and have very controlled and reproducible condi-
tions.538 This is much easier to achieve with dry etching as
the patterning technique.
For strongly bonded materials like Ga2O3, the so-called
high-density plasma etch techniques are preferred. The tech-
niques vary in the plasma source they employ.538 These dif-
ferent variants include electron cyclotron resonance (ECR)
which operates at microwave frequencies (2.45 GHz), and
inductively coupled plasma (ICP) and magnetron RIE
(MRIE), which operate at rf frequencies (2–13.56 MHz).
These techniques are generally flexible enough in terms of
the higher ion and reactive neutral densities they produce rel-
ative to RIE, that they are favored in most critical etch appli-
cations in the industry. The plasma densities are 2–4 orders
of magnitude higher than RIE, thus improving the bond
breaking efficiency in the semiconductor being etched, while
the increased ion density also enhances the desorption of
FIG. 30. (a) SEM image of the etched surface with Miller plane indices. (b) etch products formed on the surface. Additionally, since the
11
The crystal structure of (115), ð 5 Þ; and (010) planes of Ga2O3.
ion energy and the ion density can be more effectively
decoupled as compared to RIE, the plasma-induced damage
was suggested to dissolve directly in the acid solution. is more readily controlled and minimized.
Clearly, in the case of Ga2O3, we can expect that the differ- ICP has become the standard high density plasma etch-
ence in Ga-to-O ratio and the dangling bond density on dif- ing variant platform for patterning semiconductors. ICP plas-
ferent planes will play a strong role in the PEC etching mas are formed in a vacuum chamber encircled by an
behavior. The ð 201Þ-oriented Ga2O3 crystal etches roughly inductive coil into which rf power is applied at frequencies
three to four times faster under PEC conditions in KOH solu- of 2–13.56 MHz to couple through a dielectric window. The
tions.492 We speculate that the higher etch rates for the ð201Þ alternating electric field between the coils induces a strong
Ga2O3 may be due to the higher density of O dangling bonds, alternating magnetic field, which acts to trap electrons in the
which are exposed on the surface.488 The O dangling bond center of the chamber and generates a high-density plasma.
density for the ð
201Þ surface is 2–3 times higher than for the Since ion energy and plasma density can be effectively
(010) surface. Thus, the (010) surface is more chemically decoupled, uniform ion density and energy distributions are
stable than the ð
201Þ surface due to low surface energy.492 transferred to the sample while keeping ion and electron
energies low to minimize the surface damage. Thus, ICP
DRY ETCHING OF Ga2O3 etching can produce low damage while maintaining fast etch
rates. Anisotropy is achieved by superimposing of rf bias on
Dry etching methods are the standard for patterning the sample.
small feature sizes, such as the one that occur in integrated Studies of dry etching of Ga2O3 are still in their infancy
circuit manufacturing.538 As discussed in “Wet Etching of and there is much to establish,539–543 including the appropri-
Ga2O3” section, wet etching is problematic for Ga2O3, and ate gas chemistries to employ, the etch mechanisms, and the
thus, dry etch processes are needed. Plasma-driven chemical effects of the plasma exposure on the near-surface electrical
reactions and/or energetic ion beams are used to remove and optical properties of the Ga2O3.547–549 The high bond
materials in all the different dry etching techniques.539–549 strength of b-Ga2O3 suggests that ion-assisted etching will
Rather than being carried out in beakers or liquid etch baths be the likely mechanism needed to achieve practical etch
as is the case for wet etching, a dry etch system is comprised rates. We can enhance this using the Ar added to the chlori-
of a stainless steel or aluminum vacuum chamber that can be nated gases. There are a number of things that can occur in
fitted with endpoint detection tools and vacuum load locks to this near-surface region, including a change in stoichiometry
ensure reproducible conditions. Dry etching includes gas relative to the bulk, the displacement of lattice atoms by ion
processes where neutral atoms are the etching species, such bombardment from the plasma, and the presence of etch resi-
as the removal of photoresist using oxygen atoms, etching of dues consisting of the reactive species from the plasma
TABLE VI. Summary of dry etch results for b-Ga2O3.
Plasma Maximum etch

Plasma chemistry Ga2O3 samples source/frequency rate (Å min1) Comments Reference
SF6/Ar Thin film (MOCVD) ICP/2 MHz 350 Chemical enhancement over pure Ar Liang et al.541
Cl2/BCl3 Bulk EFG RIE/13.56 MHz 120 Faster rates for (010) and (201) relative Hogan et al.542
to (100)
BCl3, BCl3/SF6, CF4/O2 Bulk EFG ICP/13.56 MHz 450 BCl3 produced fastest rates Hogan et al.542
BCl3/Ar Bulk (201) EFG ICP/2 MHz 1600 Almost vertical sidewalls Zhang et al.543
O2/Ar, SF6/Ar, CHF3/Ar, Bulk (201) EFG ICP/13.56 MHx 1440 Only BCl3/Ar showed some selectivity Shah and
BCl3/Cl2, Cl2/Ar, BCl3/ to SiNX masks. No temperature depen- Bhattacharya540
Ar dence of etch rate up to 210 C
SF6 Bulk EFG RIE/13.56 MHz 160 Used for thinning of exfoliated quasi-2D Kwon et al.549
flakes
BCl3/Ar Bulk (201) EFG ICP/2MHz 800 Damage study-barrier height decreased Yang et al.548
by 28% and ideality factor increased
after etch
Cl2/Ar, BCl3Ar Bulk (201) EFG ICP/2 MHz with different 1300 Surfaces become oxygen-deficient under Yang et al.544
frequency chuck biasing high power conditions
(13.56 or 40 MHz)
BCl3/Ar Bulk (201) EFG ICP/2 MHz, with 13.56 700 Annealing study-450 C anneals found to Yang et al.550
MHz chuck biasing restore barrier height after etching
attached to the dangling surface bonds or lattice atoms. A

summary of the reported etch rates and damage characteris-
tics is shown in Table VI. Notable amongst these studies are
those of Hogan et al.,541 who measured GaO3 etch rates in
plasmas of Cl2/BCl3 under RIE conditions and plasmas of
BCl3, BCl3/SF6, and CF4/O2 under inductively coupled
plasma (ICP) conditions. Compared to the removal rates
achieved with other oxides such as ZnO, the etch rates
achieved for Ga2O3 were lower and the etched surfaces has
poorer morphology. Shah and Bhattacharya539 found that
there was basically no temperature dependence of the etch
rate in plasmas of BCl3/Ar or Cl2/Ar under ICP conditions
up to 200 C. In general, the plasma-induced damage in
Ga2O3 is of n-type character and thus will increase the con-
ductivity of the etched surface. This can be employed to
improve the properties of n-type Ohmic contacts. This same
process was used to enhance the Ohmic contact behavior in
lightly n-type GaN.538,544
Figure 31 (top) shows SEM micrographs of features
etched into the bulk Ga2O3 using a Cl2/Ar discharge and a
SiO2 mask, which is still in place on the sample. The etching
in this case used a high ICP power (800 W) and a moderate
rf power (150 W). There is some sidewall roughness, which
is replicated from the initial photoresist used for patterning
the SiO2 mask and the morphology on the etched field is
somewhat rough. Figure 31 (bottom) shows similar micro-
graphs, this time using BCl3/Ar, which leads to a smoother
morphology, perhaps because of the ability of the BCl3 to
remove the native oxide effectively. When using Cl2, the
native oxide will be removed in a slower, less uniform man-
ner, leading to variations when the actual etching commen-
ces and hence a rougher morphology.
Shah and Bhattarcharya539 also examined the etching of
b-Ga2O3 (201) substrates using ICP-RIE. Fluorine-based FIG. 31. SEM micrographs of bulk Ga2O3 dry etched in an ICP discharge of
Cl2/Ar using a SiO2 mask, which is still in place. The image at the top shows
plasmas produced low etch rates for Ga2O3, as expected
the anisotropic nature of the etching, while the image at the bottom shows
since GaF3 has a low vapor pressure and is not volatile the surface morphology after an etch depth of over 2 lm. The etching in this
enough to produce fast removal rates. Much higher etch rates case used high ICP power (800 W) and moderate rf power (150 W).
were obtained using chlorine-based plasmas.539 The study

looked at the effect of varying the Cl2/BCl3 gas content and
also the temperature-dependence in the range of 22–205 C,
which is above the sublimation temperature of GaCl3.539
Despite the chemical similarity between Ga2O3 and GaN, the
trends for Ga2O3 were significantly different from GaN. This
was ascribed to the fact that GaN etching in Cl2/Ar is due to
the formation of GaCl3, which has a significant vapor pressure
and leads to high removal rates.539 By contrast, it was sug-
gested that the etching of Ga2O3 is limited by the reaction of
BCl2þ ions from the BCl3/Ar plasma and then removing oxy-
gen from the Ga2O3 substrate to form volatile B3Cl3O3 and
nonvolatile B2O3, with the latter removed by physical sputter-
ing by positive ions incident on the surface.539 It must be
emphasized that these possible reaction products were not
directly observed and additional studies are needed to conclu-
sively establish the etch mechanism and products. As seen in
Table VI, the BCl3-based discharges produce higher etch rates
than Cl2-based. A possible explanation is that BCl3 reacts
with the oxygen in the Ga2O3, while at the same time, the
BCl2þions provide ion-enhanced sputtering. By contrast, in
Cl2-based discharges, the atomic and molecular chlorine spe-
cies do not have the same ability to remove oxygen.
In terms of electrical effects to the surface as a result of
dry etching, Fig. 32 shows the reverse I-V characteristics of
diodes fabricated on the etched surfaces using either 40 MHz
(top) or 13.56 MHz (bottom) rf chuck biases with different
ICP powers and etch times, all with BCl3/Ar discharges.549
The reference diodes that were not exposed to the plasma
exhibited reverse breakdown voltages of approximately 50 V.
In the diodes exposed to the plasma, the reverse breakdown
voltage was significantly reduced as a result of both ion- FIG. 32. Reverse I-V characteristics of diodes fabricated on the etched sur-
induced damage and non-stoichiometry induced in the surfaces using either the 40 MHz (top) or 13.56 (bottom) rf chuck biasing con-
face. The extent of this degradation depended on the self-bias. ditions with BCl3/Ar discharges. Reprinted with permission from Yang
This controls the incident ion energy and hence the density of et al., J. Vac. Sci. Technol. B 35, 031205 (2017). Copyright 2017 American
Vacuum Society.544
point defects created by the impinging ions. The damage
induced by plasma exposure could be essentially completely
BAND ALIGNMENTS OF DIELECTRICS ON
removed by annealing at 450 C. Photoluminescence (PL) is a
(2201) Ga2O3
non-destructive method for monitoring changes to the near-
surface region after plasma processing. If a large concentra- There are three main criteria that a gate dielectric must
tion of non-radiative centers is introduced by the plasma possess to be considered acceptable.550–552 First, it must be
exposure, this would be expected to decrease the overall PL
intensity. We did not observe a significant change in the total
intensity from the surface after introduction of the etch dam-
age, as shown in Fig. 33. Based on a comparison of these
spectra with the literature, we assign the transitions as due to
an O defect donor band to the valence band (EDB
¼ 3.411 eV) and to the Ga vacancy band (ED2 ¼ 3.002 eV to
ED3 ¼ 2.39 eV).
In summary, Ga2O3 can be etched at practical rates
using chlorine-based plasma chemistries, especially under
high ion density conditions such as ICP discharges. Due to
the high bond strengths, the rates are lower than for other
electronic oxides like ZnO where the same plasma condi-
tions produce etch rates of a factor of two higher than for
Ga2O3. Schottky barrier height measurements reveal the
presence of ion-induced damage that can be effectively
removed by annealing at 450 C. There is a little change in
the optical properties of stoichiometry of the etched surface. FIG. 33. 10 K PL spectra from reference and plasma damaged Ga2O3.
thermodynamically stable with the semiconductor and not core levels in the heterojunction. The determination of DEv
react during processing. Second, it must provide a high qual- comes by combining those three quantities. Heterojunction
ity interface with low defect and trap density to ensure high samples, consisting of a thin (1–2 nm) layer of dielectric
carrier mobility. Finally, it must act as a barrier to both elec- deposited on the semiconductor, are prepared in which the
trons and holes which is among the most important physical separation between reference core levels in each material is
parameters for a given heterojunction system. Because these measured. The separation between the reference core levels
discontinuities can form a barrier for carrier transport across can be translated directly into a value for the valence band
the interface, the knowledge of these quantities is essential offset (VBO) using the previously measured single layer
for calculating the transport properties of the interface, or the sample core-level to VBM energies by the following
electrostatic potential in a heterojunction device. There are relationship:553
three types of band alignments: type I, type II staggered, and
type III broken gap.550–555 DEV ¼ E1core E1VBM IGZO E2core E2VBM Dielectric
Numerous experimental methods are used to determine
E 1core E2core Heterostructure
the oxide/semiconductor band alignment. These include
external photoemission spectroscopy,555–558 internal photo- To determine the conduction band offset, it is necessary to
emission spectroscopy,555–557 and x-ray photoelectron spec- measure the band gap of each material. The samples are
troscopy (XPS) core-level based method.553 The method of measured via UV/Vis or REELS, and the conduction band
Kraut et al.553 using x-ray photoemission spectroscopy has offset is calculated as follows:
been established as a reliable way to determine band offsets
at the hetero-junction interface. This method has also been DEC ¼ Edielectric
g EIGZO
g DEV :
successfully used to provide insights into interfacial proper-
ties between different materials.553,559 It is based on using an From these values, it is then possible to construct the flat
appropriate shallow core-level position as a reference. band diagram and determine if this is a type I, II, or III
Generally, this approach is based on the assumption that the heterostructure.555,557
energy difference between the core-level positions and Table VII contains a list of reported band offsets
valence-band maximum (VBM) are both fixed in the bulk. obtained for different dielectrics deposited by different
The basic method, shown in Fig. 34, is to first measure methods on Ga2O3.560–573 Figure 35 shows a summary of
the energy difference between a core level and the VBM for the measured band offsets for different dielectrics on
both single layer dielectric and semiconductor of inter- Ga2O3.560–573 This shows that there are a limited number of
est.553,559 One measures the reference core level binding choices that are appropriate for gate dielectrics on Ga2O3,
energies in thick films of each material and then measures where it is desirable that there be adequate conduction and
the binding energy difference between the two reference valence band offsets. These include SiO2 and Al2O3.
However, the other films could still be an option as a surface
passivation layers on Ga2O3 to prevent surface conductivity
changes upon exposure to ambient. The sensitivity of Ga2O3
to water vapor and hydrogen is not yet firmly established,
but many oxides like ZnO and InGaZnO4 do show such sen-
sitivity and require surface passivation to provide device sta-
bility in humid ambients. The other thing to note from Fig.
35 is the fact that there is a difference in band offsets
reported by different groups for nominally the same types of
films. Some of the reasons include metal contamination,
interface disorder, dielectric composition, carbon/hydrogen
contamination, annealing, stress/strain, and surface termina-
tion.557 The presence of these effects can result in differ-
ences in the bandgap of the dielectric and this affects the
conduction band offset since the valence band offset is
directly measured. However, the latter can also be affected
by most of these same issues. For example, sputtered films
are more likely to contain metallic contaminants and have
interfacial disorder due to the sputter-induced damage and
therefore have possible Fermi level pinning effects than a
more controlled process such as ALD. Interfacial defects,
such as oxygen or metal atom vacancies, may have a pro-
nounced effect on the band offsets. The literature shows that
energy band alignment variations of sometimes more than
FIG. 34. A schematic energy band diagram illustrating the basic principle of
XPS band offset measurements. Reprinted with permission from Appl. Phys. 1 eV depending on interface preparation can be obtained,557
Rev. 4, 021301 (2017). Copyright 2017 American Institute of Physics.560 due to the presence of high defect concentrations in the
TABLE VII. Reported values for band offsets for different materials on Ga2O3.
Dielectric material
(crystalline nature of Ga2O3) Synthesis method DEC (eV) DEV (eV) Alignment type Reference
SiO2 (single crystal) PECVD 3.1(60.2) 1.0(60.2) I Konishi et al.567

SiO2 (single crystal) ALD 3.63–3.76 0.3-0.43 I Jia et al.568
SiO2 (single crystal) ALD 2.9 (60.7) 1.2 (60.2) I Carey et al.565
Al2O3 ALD 1.5–1.6(60.2) 0.7(60.2) I Kamimura et al.569
cAl2O3 (single crystal) PLD 1.9 0.5 I Hattori et al.570
Al2O3(single crystal) ALD 2.23(60.2) 0.07(60.2) I Carey et al.564
Al2O3(single crystal) PVD 3.16(60.2) 0.86(60.2) II Carey et al.559
LaAl2O3(single crystal) PVD 2.01(60.60) 0.21 (60. 02) II Carey et al.563
Si(amorphous) PLD 0.2(60.1) 3.5(60.1) I Chen et al.571
GaN(polycrystalline) oxidation 0.1(60.08) 1.4(60.08) I Wei et al.572
6H-SiC(amorphous) PVD 0.89(60.1) 2.8(60.1) II Chang et al.573
ZrO2 (single crystal) ALD 1.2 0.3(60.04) II Wheeler et al.574
HfO2(single crystal) ALD 1.3 0.5(60.04) II Wheeler et al.574
HfSiO4 (single crystal) ALD 2.38 (60.5) 0.02 (60.003) I Carey et al.565
AZO (single crystal) PVD 0.79(60.34) 0.61(60.023) I Carey et al.561
ITO (single crystal) PVD 0.32(60.34) 0.78(60.34) I Carey et al.558
AlN(single crystal) PLD 1.75(60.05) 0.55(60.05) II Sun et al.756
materials and on a cation effect that will increase the VBM of determined to be inversely proportional to the volume of the
that material. unit cell. This generally scales with bandgap, so that these
materials have intrinsically higher radiation resistance than
RADIATION DAMAGE IN Ga2O3 Si. Similarly, wide bandgap devices generally employ higher
The strong bonding in a wide bandgap semiconductor critical fields and smaller active volumes that reduce
gives them an intrinsically high radiation resis- radiation-induced charge collection. Tsao et al.321 summa-
tance.321,574–593 The fluence of ionizing radiation at which rized the fact that there are four basic kinds of semiconductor
materials and devices such as transistors and light-emitting device radiation effects, including total ionizing dose that
diodes made from SiC, GaN, and related materials start to results in charge accumulation in field oxides in MOS-based
show degradation is about two orders of magnitude higher devices. This results in the well-known shifts in the threshold
than in their GaAs equivalents.574–579 This difference is voltage, but since most wide-bandgap transistors use
attributed to the stronger bonding of these materials.574–582 Schottky metal gates, this is less of an issue. There are also
A measure of this bond strength is the energy required to dis- single event upset effects that result from the transit of ener-
place an atom from its lattice position or simply the atomic getic ions passing through the semiconductor, creating
displacement energy, denoted by Ed. This parameter has electron-hole pairs. Again, the device structures employed in
been measured in several semiconductors and empirically wide bandgap semiconductors, involving heterostructures,
tend to mitigate this effect. The same applies to the dose-rate
radiation effects, which are sensitive to the total volume of a
device. The last issue is lattice displacements that typically
create traps and recombination sites in the device that
degrade the carrier density through trapping and the carrier
mobility, with both of these mechanisms scaling with the
radiation dose. Si MOSFETs also suffer from single-event
burnout when the charge from an energetic ion creates sus-
tained conduction of the parasitic bipolar transistor inherent
in Si MOSFETs that leads to thermal degradation and single-
event rupture when charge build-up near the gate causes a
breakdown in the gate oxide.321
Ga2O3 has been investigated as a radiation detection
material for fast (14 MeV) neutrons,588 utilizing the 16O
(n,a)13C reaction. Both diamond and 4H-SiC have previ-
ously been used as nuclear detectors under extreme condi-
tions, including temperature up to 700 C for 4H-SiC and
200 C for diamond.588 Pt Schottky diodes of conducting
Ga2O3 with rear Ti/Au ohmic contacts or insulating samples
with Ti/Au ohmic contacts on both sides were examined. The
FIG. 35. Summary of reported band offsets for dielectrics on Ga2O3. latter could be operated up to 1000 V.588 Neutrons could be
detected under these conditions.584 Wong et al.589 examined fluence, with the values in the range of 21–25 ns for all recti-
the gamma-ray irradiation tolerance of Ga2O3 MOSFETs to fiers. The changes in device characteristics were accompa-
doses of 230 kGy(SiO2). Hysteresis in the transfer characteris- nied by a decrease in electron diffusion length from 325 to
tics was negligible after exposure to the highest dose, while 240 lm at 300 K, as shown at the bottom of Fig. 36.
degradations in the gate oxide were found to limit the overall Proton damage in back-gated field-effect transistors
radiation resistance.593 Initial DLTS measurements on (FETs) fabricated on exfoliated quasi-two-dimensional b-
neutron-irradiated samples show the introduction of a level Ga2O3 nanobelts were studied by exposure to 10-MeV pro-
near EC–1.88 eV, whose microstructure is unknown.515 tons.594 The proton dose and time-dependent characteristics
Yang et al.589 subjected vertical rectifiers fabricated on of radiation damaged FETs showed a decrease of 73% in the
epi Ga2O3 on bulk b-Ga2O3 to 1.5 MeV electron irradiation field-effect mobility and a positive shift of the threshold volt-
at fluences from 1.79 1015 to 1.43 1016 cm2. The elec- age after proton irradiation at a fluence of 2 1015 cm2,
tron irradiation caused a reduction in carrier concentration in which corresponds to approximately 105 times the intensity
the Ga2O3, with a carrier removal rate of 4.9 cm1. Figure 36 of a solar proton event. The on/off ratio of the exfoliated b-
(top) shows that the 2kT region of the forward current- Ga2O3 FETs was maintained even after proton doses of up to
voltage characteristics increased due to electron-induced 2 1015 cm2. The data are summarized in the drain-source
damage, with more than 2 orders of magnitude increase in characteristics of Fig. 37, which show the effect of proton
the on-state resistance at the highest fluence. There was a dose.594 The radiation-induced damage in b-Ga2O3-based
reduction in the reverse current, which scaled with electron FETs was significantly recovered after rapid thermal anneal-
fluence. The on/off ratio at 10 V reverse bias voltage was ing at 500 C.594 It will be interesting to compare the results
severely degraded by electron irradiation, decreasing from from nanobelt transistors with more conventional devices
approximately 107 in the reference diodes to approximately fabricated on bulk or epi Ga2O3.
2 104 for the 1.43 1016 cm2 fluence. The reverse recov- To summarize, the initial data on proton, electron, neu-
ery characteristics showed little change even at the highest tron, and gamma irradiation of photodetectors and transistors
show fairly similar radiation resistance to GaN devices under
the same conditions.
THIN FILM SOLAR-BLIND UV DETECTORS

One of the most promising applications for b-Ga2O3 is
for deep UV solar blind detectors (cut-off wavelength shorter
than 280 nm). These detectors only sense radiation with
wavelength shorter than 280 nm while being insensitive to
visible and infra-red radiation. Few photons in this wave-
length region reach the Earth’s surface from the sun due to
strong absorption by the stratospheric ozone layer, and thus,
FIG. 37. Output characteristics (IDS vs. VDS) of b-Ga2O3 nanobelt FET
before and after 10-MeV proton irradiation at different doses: (a) as-
fabricated, (b) 1 1015 cm2, and (c) 2 1015 cm2, and (d) transfer charac-
teristics (IDS vs. VGS) of b-Ga2O3 nanobelt FET at VDS ¼ 30 V before and
FIG. 36. (a) I-Vs from Ga2O3 diodes before and after 1.5 MeV electron irra- after 10-MeV proton irradiation at different doses. Reprinted with permis-
diation to different doses; (b) diffusion length of electrons as a function of sion from Yang et al., ACS Appl. Mater. Interfaces 9, 40471 (2017).
temperature after different electron irradiation doses. Copyright 2017, American Chemical Society.595
these devices’ photodetectors can respond to a very weak have also been reported.602,603,607,623,625 In particular, lower
signal even in sunlight or room illumination. These photode- dark current, higher photoresponse, and faster switching
tectors have a large number of applications including flame time under 254 nm light illumination were obtained by intro-
detection, missile guidance systems, underwater communica- ducing Au nanoparticles.626 The improved performance has
tions, photolithography, automatization, intersatellite com- been ascribed to localized surface plasmonic resonance from
munication, and biochemical detection.321,595–629 b-Ga2O3 the nanoparticles.622 Lu et al.615 showed that higher partial
has a cut-off wavelength in the range of 250–280 nm, which pressures of oxygen (PO2) in PA-MBE improved the crystal
leads to detection over the full range of deep ultra-violet quality, surface morphology, and the chemical performance
(DUV). In addition, the high chemical and thermal stabilities of the photodetectors, which exhibited the significantly
of this material are attractive for applications involving high improved photocurrent and responsivity characteristics in
temperatures or extreme environments. High performance comparison with devices grown with lower PO2 of 0.01
photodetectors based on nanostructure, thin-film, and bulk mbar. This was attributed to a reduction in the number of
Ga2O3 have been reported. Different device structures, oxygen vacancies.611 A similar result was found by An
including photoconductors, metal-semiconductor-metal et al.,607 and Feng et al.608 reported better performance of
(MSM) or metal-intrinsic-semiconductor-metal photodetec- thin film photodetectors on bulk Ga2O3 relative to sapphire
tors, avalanche photodiodes, and Schottky barrier photodio- substrates because of a lower defect density.
des, have been reported.623 Among these, MSM structures, The photodetection mode has also been investigated.
which consist of back-to-back Schottky contacts, are simple Qian et al.624 employed a four-terminal photodetector fabri-
to fabricate and have low dark currents. cated on b-Ga2O3 deposited by PA-MBE. A dark/photo volt-
Pratiyush et al.595 reported MBE grown epitaxial b- age ratio of 15 was achieved, comparable to interdigitated
Ga2O3-based solar blind metal-semiconductor-metal (MSM) MSM structures. The aperture ratio was >80%, roughly dou-
photodetectors fabricated on (201) b-Ga2O3 films grown ble that of MSM structures. The dark/photo voltage ratio was
using plasma-assisted MBE on c-plane sapphire. MSM nearly trebled with use of two Zener diodes.624
geometry devices were fabricated using Ni/Au contacts in an There has been particular interest in fabricating photode-
interdigitated geometry. The photodetectors exhibited peak tectors on b-Ga2O3 nanobelts exfoliated from bulk substrates
SR > 1.5 A/W at 236–240 nm at a bias of 4 V with a UV to because of the high surface-to-volume ratio, low power
visible rejection ratio > 105. The devices had low dark requirement, and flexibility in transferring to other sub-
current (< 10 nA at 20 V) with no persistent photoconductiv- strates. Oh et al.620,622 employed b-Ga2O3 micro-flakes exfo-
ity and a large photo-to-dark current ratio >103. Figure 38 liated from single crystalline b-Ga2O3 using a mechanical
(Ref. 595) illustrates both the steady-state photo, dark I-V exfoliation method by an adhesive tape, similar to mechani-
characteristics, and time-dependent photo-response at a bias cal exfoliation of highly oriented pyrolytic graphite to obtain
of 20 V after passivation with 20 nm of ALD Al2O3. This graphene. The exfoliated b-Ga2O3 micro-flakes were then
was an important step in optimizing performance.595 The transferred onto a Si substrate with a thermally grown
photo currents in the steady-state and transient measure- 300 nm-thick SiO2 layer, which was followed by a conven-
ments were found to be similar, which are 4.6 lA (at tional photolithography process to form a MSM configura-
20 V), while the dark current was observed to be in the nA tion. Figure 39 shows a plot of (ah)2 vs. photon energy,
range. Both the photo and the dark current decreased slightly with the inset showing the UV-VIS absorbance of the bulk
after the passivation of the devices. Yu et al. reported similar b-Ga2O3 from which the exfoliated flake was taken; (b)
results using PLD films,601 showing that the thin film deposi- Raman spectrum of the exfoliated b-Ga2O3 micro-flake; and
tion method appears to be less critical for photodetectors (c) cross-sectional TEM image of the fabricated b-Ga2O3.622
than other devices such as rectifiers and transistors, although The dark current was in the pA range, the ratio of photo-to-
higher UV to visible rejection ratio is generally obtained dark current was >103, responsivity of 1.68 A/W, the
with better crystal quality. Hybrid structures, including gra- 254 nm/365 nm rejection ratio was 1.92 103, and no persis-
phene, Si, BN, ZnO, SiC, Au nanoparticles, SnO2, and GaN, tent photoconductivity was observed.622 Oh et al.620 further
reported nanobelt photodetectors with responsivities under
illumination to 254 nm light of 9.17 104 A W1,
1.67 105 A W1, and 1.8 105 A W1 at VGS of
0 V, 15 V, and 30 V, respectively. They suggested that
the performance can be further improved by minimizing b-
Ga2O3 defects through optimizing the fabrication process
and increasing the size of the active region to absorb pho-
tons, thus reducing the high dark current that is a drawback
for such photoconductors.620 Notably, Oh et al.627 also
employed implantation of Si to improve the contact resis-
FIG. 38. (a) Steady-state photo and dark I-V characteristics at room temper- tance of Ohmic contacts on their thin film photodetector
ature after passivation (log scale). (b) Time-dependent photo-response under structures grown by MOCVD. Si ions were implanted
236 nm illumination at 20 V (log scale). The inset of (a) shows Schottky bar-
rier lowering variation with the bias voltage. The inset of (b) shows rescaled
at 30 keV energy at a dose of 1 1015 cm2 and post-
transient at 20 V (linear scale). Reprinted with permission from Appl. Phys. implantation annealing to activate the implanted Si atoms
Lett. 110, 221107 (2017). Copyright 2017 American Institute of Physics.596 was performed under Ar ambient at 900 C by rapid thermal
drift layers grown by the established epi methods.638–640

GaN is preferred for 1 kV range devices because of the lower
on-state resistance. Many of the figures-of-merit for device
performance in Table II scale with increasing bandgap in a
highly non-linear manner, so Ga2O3 has the potential for per-
formance far superior to that of SiC and GaN. For a low-
frequency unipolar vertical power switch, the Baliga figure
of merit (BFOM)321 is defined as V2BR/RON, where VBR is
the maximum voltage that the switch can block when it is
off, and RON is the specific on-resistance (the inverse of the
conductance per unit area when the switch is on).321,322 The
Johnson FOM relates the high frequency and high voltage
capability of a device. The rectifiers should have low on-
resistance while having very high blocking voltages in the
off state. The critical electric field in semiconductors scales
approximately as the square of the bandgap. More precisely,
a generally accepted empirical relationship between the
FIG. 39. (a) Plot of (ah)2 vs. photon energy, with the inset showing the breakdown field ec and bandgap (Eg) is ec ¼ (Eg)n, where n is
UV-VIS absorbance of the bulk b-Ga2O3 sample; (b) Raman spectrum of between 2 and 2.5. This means that the BFOM scales
the exfoliated b-Ga2O3 micro-flake; (c) cross-sectional TEM image of the approximately as the sixth power of bandgap.321,639,640
fabricated b-Ga2O3. Reprinted with permission from Oh et al. ECS J. Solid
Figure 40 shows how the wider bandgap really improves the
State Sci. Technol. 6, Q79 (2017). Copyright 2017 The Electrochemical
Society.623 rectifier performance, with lower on-state resistance at a
given reverse voltage.319 However, this applied to the case
annealing.627 The RS was 164.1 kX/ⵧ, while the electrical of low defect material, where the performance is not limited
activation efficiency of Si atoms implanted into Ga2O3 films by defect-assisted breakdown. In the presence of defect such
was reported to be approximately 75%.627 as screw dislocations, nanopipes, or voids, it has been dem-
onstrated that premature breakdown occurs in both GaN and
POWER RECTIFIERS SiC.627,628 Especially in GaN rectifiers, the reverse break-
down voltage increases dramatically as the contact size is
The development of high power electronics capable of decreased and is also much larger for vertically depleting
operating at elevated temperatures without the need for devices.627,629 Note that both SiC and GaN rectifiers are at a
extensive system cooling requirements is attractive in many commercialized stage. Figure 41 shows the current state of
industrial and, in particular, military applications. The cost the art for commercial SiC and GaN devices,639 which have
of the electronics components in radar, communication, and reached an impressive level of maturity.
guided missile systems is a significant fraction (typically There have been a large number of recent reports on
20%–30%) of the total cost of these systems, leading to the Ga2O3-based rectifiers.641–661 The device structures for the
strong support from military agencies in the past to develop most part have been relatively simple. For example, these
GaAs and GaN electronics.322 rectifiers show performance limited by the presence of
As shown in Table II, the figures of merit for power defects and by breakdown initiated in the depletion region
applications scale with bandgap, and therefore, Ga2O3 will near the electrode corners.641,646,647 In SiC and GaN
have advantages over both SiC and GaN with their smaller
bandgaps.321 Wide bandgap semiconductor power Schottky
diodes have numerous advantages over more conventional Si
rectifiers, achieving a maximum electric field breakdown
strength over 10 times larger and on-state resistance (RON)
more than four hundred times lower at a given volt-
age.630–640 These characteristics have made SiC and GaN
devices attractive for hybrid electric vehicles and power con-
ditioning in large industrial motors. In particular, Schottky
rectifiers are attractive because of their fast switching speed,
which is important for improving the efficiency of inductive
motor controllers and power supplies. Both GaN and SiC
power Schottky diodes have demonstrated shorter turn-on
delays than Si devices.630–640 The advantage of simple
Schottky rectifiers over p-n diodes is the shorter switching
times due to the absence of minority carriers. This, however,
leads to higher Ron values than in p-i-n rectifiers.
FIG. 40. Comparison of breakdown fields and on-state resistance in candi-
Currently, power rectifiers are made from SiC for very date materials for high power electronics (Adapted from Higashiwaki
large breakdown voltages because of the availability of thick et al.,15 Tadjer et al.,54 and Tsao et al.321)
with novel ALD TiN (65 nm thick) contacts on (201)

b-Ga2O3. Both contacts showed thermionic emission, similar
barrier heights of about 1 eV and near-unity ideality factor
values, independent of temperature. The TiN contact actually
had a much lower reverse current at room temperature.
Figure 42(a) shows Richardson plots for the two types of
contacts. The extracted barrier heights were 1.01 eV for Pt
and 0.98 eV for TiN, and this difference was much lower
than the 2.6 eV predicted by the difference in work func-
tions.518 This may result from image force lowering. Ideality
factors as a function of temperature were extracted at low
bias in order to eliminate series resistance effects. Both sam-
ples had ideality factors close to unity over the whole tem-
perature range investigated, as shown in Fig. 42(b). The only
drawback of the TiN was a tendency to oxidize at higher
temperatures.
Reverse breakdown voltages of over 1 kV for b-Ga2O3
FIG. 41. Performance of commercially available SiC and GaN power have been reported by several groups,646–649 either with or
diodes. Reprinted with permission from A. Q. Huang, Proc. IEEE 99, 1 without edge termination. A typical device structure is
(2017). Copyright 2017 IEEE.640
shown in Fig. 43 (top). The highest reverse breakdown vol-
tages have been achieved with similar layer structures, con-
sisting of a thick epitaxial layer grown on a high quality
rectifiers, a wide variety of edge termination methods have
EFG substrate,646–649 approximately 10 lm thick of lightly
been employed to smooth out the electric field distribution
Si-doped n-type Ga2O3 grown by HVPE on nþ bulk, (201)
around the rectifying contact periphery, including mesas,
Sn-doped (3.6 1018 cm3) Ga2O3 single crystal wafers.
high resistivity layers created by ion implantation, field
The dislocation density from etch pit observation was
plates, and guard rings.631,638–640 In addition, the design of
approximately 103 cm2. The reverse leakage current has
SiC and GaN devices is relatively sophisticated, taking into
been closely correlated to the dislocation density in (0–10)
account effects like incomplete ionization of dopants, tem-
oriented bulk b-Ga2O3 which revealed upon a hot H3PO4
perature dependence of mobility and bandgap, ionization
acid delineation etch for 1 h.658 Diodes were fabricated by
rates due to avalanche multiplication, carrier lifetimes, dop-
depositing full area back Ohmic contacts of Ti/Au (20 nm/
ing dependence of mobility, and a complete understanding
80 nm) by E-beam evaporation, while the Schottky contacts
of how edge termination affects reverse breakdown voltage
were patterned by lift-off of E-beam deposited Schottky con-
for a given lateral or vertical device geometry.630,631,636,641
All of these need to be developed for Ga2O3. tacts Ni/Au (20 nm/80 nm) on the epitaxial layers.643,644
One thing that is relevant here is the fact that while a- Note that this device does not employ edge termination and
Ga2O3 actually has a larger bandgap (5.16 eV) than b- shows the capability of the Ga2O3 to withstand high field
Ga2O3, the few rectifiers demonstrated on this material to strengths. The diameter of these contacts ranged from 20 lm
date have not shown particularly large breakdown vol- to 0.53 mm. The bottom of Fig. 43 shows the optical images
tages,322 perhaps at this stage due to the need for continued of some of the completed diodes with different
optimization of material quality. diameters.643,644
Compared with lateral diodes grown on insulating sub- Figure 44 (top) shows the forward and reverse current
strates, vertical geometry Schottky diodes on conducting sub- density-voltage (J-V) characteristic from a 20 lm diameter
strates can deliver higher power with full back side Ohmic diode. The VBR was approximately 1600 V for this diode
electrodes and have higher current capability since they take
advantage of the entire conducting area. Edge termination can
also enhance the performance by preventing premature break-
down due to field crowding around the contact periphery. One
of the key components of a rectifier is the Schottky contact,
which we covered in an earlier section (Schottky Contacts to
Ga2O3). Schottky contacts to b-Ga2O3 have been characterized
using a number of metals, including Ni,478,500,503,504,511,651
Pt,500,518 Cu,469 and Au.136 The transparent Schottky contacts
such as poly(3,4-ethylenedioxythiophene) polystyrene
sulfonate (PEDOT:PSS) have also been demonstrated for
flame detection.482 To give an example of the control that is FIG. 42. (a) Richardson plots comparing Pt and TiN contacts to (201)
Ga2O3. (b) Ideality factor–thermal voltage (nVthermal ¼ nkT/q) for the
possible with different metals on b-Ga2O3, Tadjer et al.518
two contacts. Reprinted with permission from Tadjer et al., ECS J. Solid
carried out temperature I-V and Schottky barrier height stud- State Sci. Technol. 6, P165 (2017). Copyright 2017 The Electrochemical
ies on e-beam evaporated Pt/Au bilayers and compared these Society.519
FIG. 43. Schematic of vertical Ni/Au Schottky diode on Ga2O3 epi layer on
a conducting b-Ga2O3 substrate (top) and top-view microscope image of the
fabricated b-Ga2O3 diodes (bottom). Reprinted with permission from Appl.
Phys. Lett. 110, 192101 (2017). Copyright 2017 American Institute of
Physics.648
with a smaller diameter and 250 V for the largest diame-

ter.648,649 This trend is typical for newer materials technolo-
gies still being optimized in terms of defect density.656–661 FIG. 44. (Top) Forward and reverse current density-voltage characteristics
Kasu et al.658 examined the effect of crystal defects revealed from a 20 lm diameter diode. (Bottom) Diode on/off ratios for temperatures
in the range of 25–100 C as a function of reverse bias. The forward bias
by etch pit delineation and found that dislocations are closely was held constant at 1.3 V. Reprinted with permission from Yang et al.,
related to the reverse leakage current in the rectifier and that IEEE Electron Device Lett. 38, 906 (2017). Copyright 2017 IEEE.649
not all voids produce leakage current.658,659 Dislocation
defects along the [010] direction were found to act as paths
for leakage current, while the Si doping did not affect this under the anode edge was 5.1 MV/cm,647 much larger than
dislocation-related leakage current.658–660 By contrast, in the the theoretical limits for SiC and GaN and similar to the
[102] orientation, three types of etch pits were present, breakdown field for lateral Ga2O3 MOSFETs.662,663,667 Even
namely, a line-shaped etch pattern originating from a void higher breakdown voltage should be possible by demonstrat-
and extending toward the [010] direction, arrow-shaped pits ing a junction barrier Schottky (JBS) diode architecture,
in the [102] direction, and gourd-shaped pits in the [102] whereas in the reverse bias the drift layer depletes away
direction. Their average densities were estimated to be from the surface by employing a pn junction.664,665 In the
5 102, 7 104, and 9 104 cm2, respectively, but in this case of Ga2O3, pn type heterojunctions have been demon-
orientation there was no correlation between the leakage cur- strated using Li-doped NiO deposited on Ga2O3 (Ref. 666)
rent in rectifiers and these crystalline defects.658–660 Thus, and Ga2O3 deposited on 6H-SiC,386 where the b-Ga2O3/6H-
the orientation of the substrate used determines the sensitiv- SiC anisotype heterojunction appeared to have achieved
ity to the defect density.660,661 Figure 43 (bottom) shows the minority carrier injection.
on-off current ratio measured at a fixed forward voltage of In summary, there have been a number of 1 kV break-
1.3 V and reverse biases from 5 to 40 V.647,648 The on-of down voltage rectifiers reported. The maximum current den-
ratios ranged from 3 107 to 2.5 106 for this range of sity reported is 3 kA cm2, with a lowest reported Ron of 0.1
biases and showed only a small dependence on temperature mX cm2 and a turn-on voltage of 1.7 V.322 The growth meth-
in the range of 25–100 C. This is promising for device oper- ods have included HVPE, MBE, and Mist-CVD, and both
ating temperatures in this range, since there would be little vertical and horizontal geometries have been demonstrated.
change in performance characteristics. It is worth noting again that annealing of EFG-grown materi-
Konishi et al.647 obtained high reverse breakdown vol- als under oxygen ambients typically leads to a reduction in
tages in excess of 1 kV for diodes employing an SiO2 field net carrier concentration of up to an order of magnitude,
plate with 300 nm thickness and length 20 lm. This was which can lead to higher reverse breakdown voltages.69 The
designed using simulation software to optimize the break- usual Schottky contacts are Pt/Au, Au or Pt/Ti/Au, and the
down voltage.647 The simulated maximum electric field Ohmic metallization is usually Ti/Au.322
POWER MOSFETs AND MESFETs of enhancement (e-mode) operation.676–691 The channels

have been undoped, Si,682,688 Sn or Ge-doped,680 and
Modulation-doping for heterostructures
HfO2,678 and Al2O3 (Ref. 679) and SiO2 (Refs. 682 and 684)
Tsao et al.321 pointed out that heat removal is going to have been the most widely used dielectrics. The first Ga2O3
be a major issue for Ga2O3, more so if heterostructure-type transistors were reported in the 2012–2013 time frame by
devices become necessary since they will contain binary, Higashiwaki et al.,676,677 fabricated on homoepitaxial epitax-
ternary, and quaternary compounds and interfaces between ial Ga2O3 grown by MBE on native substrates. In 2014, a
dissimilar materials. Combined with the poor thermal con- back-gated MOSFET device was fabricated by exfoliating
ductivity of Ga2O3, this will lead to a complex set of internal a nm-thick slice of Ga2O3 from the (100) face of a
impedances to heat removal. Thus, there is a need to under- Czochralski-grown bulk Ga2O3 crystal.692 In early 2016,
stand the thermal properties of these alloys, in addition to Tadjer et al. reported a MOSFET fabricated on epitaxial
their electrical and structural properties. To this point, there Ga2O3 grown by metal organic CVD using a trimethylgal-
is little known about the solubility limits and carrier confine- lium (TMG) source in O2 atmosphere.54 Even though the
ment at heterostructures with Ga2O3. films were insulating, implantation of Si in the source/drain
The first reported heterostructure between a binary and a regions and the good surface roughness obtained by growth
ternary Ga2O3 system was by Kaun et al.406 While the x-ray on c-plane Al2O3 allowed a functional transistor to be fabri-
data demonstrated that an (AlxGa1-x)2O3/GaO structure was cated. In addition, Tadjer et al.678 fabricated a (001) b-Ga2O3
grown, the undoped AlGaO layer and the step-bunching MOSFET with þ2.9 V threshold voltage and HfO2 as the
induced surface roughness of the Ga2O3 surface prevented gate dielectric. Figure 46 shows the IDS-VDS characteristics,
the formation of a two-dimensional electron gas (2DEG).406 (b) IDS-VGS as a function of drain bias (0.1, 1, and 10 V), and
These issues were successfully overcome by (c) off-state IDS-VDS characteristic, showing the soft break-
Krishnamoorthy et al.668 who reported a modulation-doped down regime at around 80 V.678
two-dimensional electron gas (2DEG) at the b-(Al0.2Ga0.8)2O3/ The outstanding devices in the literature to date have gen-
Ga2O3 heterojunction by silicon delta doping. This was grown erally employed a planar, lateral geometry.342,343,346,347,677–692
on Fe-doped (010) b-Ga2O3 semi-insulating substrates. The While the report by Tadjer et al.678 was not an enhancement
device performance was limited by the quality of the Ohmic mode device, normally off operation was reported by deplet-
contacts, with a maximum current normalized to the channel ing the channel using a wrap-gate finfet architecture, where
width of 5.5 mA/mm. Figure 45(a) shows the output character- e-mode operation and a 600 V breakdown voltage were dem-
istics and (b) transfer characteristics of AGO/GO MODFET, onstrated.679 Furthermore, a prototype lateral MOSFET with a
exhibiting good charge modulation, pinch-off, and a high ON/ very low-doped channel (1014 cm3) and gate-overlapped
OFF ratio.668 The pinch-off voltage was 3 V, on-off ratio of S/D implants was demonstrated in early 2017, which to-date
2.5 105, and peak transconductance was 1.75 mS/mm. represents the closest demonstration to a textbook
Oshima et al.669 explored the solubility limits of (AlxGa1–x)2O3 enhancement-mode MOSFET in the absence of a pn source/
grown on Ga2O3, while Takayoshi et al.670 were able to drain junction technology in Ga2O3.683 Si ion implantation
observe carrier confinement at the interface of similar hetero- has been employed to improve source/drain resistance in
structures. Several groups have demonstrated b-(AlxGa1-x)2O3 some cases.683 As a group, it is clear that the devices do
(Refs. 671–674) and a-(AlGa)2O3 (Ref. 675) grown by MBE indeed exhibit high critical field strength and breakdown
or Mist-CVD. voltages, but the current densities have been low, typically
approximately 10–100 mA/mm or less, and the on-state resis-
MOSFETs tances have been large, especially for e-mode devices. The
The metal-oxide-semiconductor field-effect transistors highest output current density was reported by Zhou et al.,684
(MOSFETs) fabricated on Ga2O3 to date have predominantly by employing an nþ doped channel exfoliated from a Sn-
been depletion (d-mode) devices, with a few demonstrations doped substrate onto an SiO2/Si substrate. For high voltage
and high power ratings, vertical topologies are preferred since
chip area utilization is more efficient and device operation is
insensitive to surface effects. Two examples of vertical
Ga2O3-based devices have been reported to date.689,690 Wong
et al.689 adopted a current aperture vertical transistor design
(CAVET) with a Mg-implanted current blocking layer. Hu
et al.,690 on the other hand, employed a deep-etch process to
fabricate 1 lm thick vertical structures into a low-doped
Ga2O3 substrate. At this stage, thermal effects have not been a
major issue, but if Ga2O3 is to realize its promise, this will
need significant attention. Given the low thermal conductivity,
wafer thinning or backside heat sinks will not be sufficient
FIG. 45. (a) Output characteristics and (b) transfer characteristics of AlGaO/
and the use of thermal shunting on the active device surface
Ga2O3 MODFET showing FET operation with charge modulation, pinch-
off, and high on/off ratio. Reprinted with permission from Appl. Phys. Lett. will be required for MOSFETs to operate with high-efficiency
111, 023502 (2017). Copyright 2017 American Institute of Physics.668 at higher power.
extrinsic cutoff frequency (fT) and maximum oscillation fre-

quency (fmax) of 3.3 and 12.9 GHz, respectively. Figure 47
(top) shows the extrinsic small signal RF gain performance
at VGS ¼ 3.5 V (peak gm) and VDS ¼ 40 V. A gain decay
of 20 dB/dec is plotted with the dashed line. The bottom of
the figure shows the 800 MHz Class-A power sweep of a 2
lm 50 lm MOSFET.687
Zhou et al.684 realized b-Ga2O3 on insulator (GOOI) d-
mode and e-mode FETs by transferring a b-Ga2O3 nano-
membrane 50–150 nm thick to a SiO2/Si substrate, followed
by device fabrication. They demonstrated maximum ID of
600/450 mA/mm for d/e-mode FETs and on/off ratio of 1010.
Ar bombardment was used to achieve Ohmic source and
drain contacts. E-mode FETs with source-drain separation
0.9 lm had breakdown voltage of 185 V, corresponding to a
field strength of 2 MV/cm. Figure 48(a) shows a schematic
of a GOOI FET while (b) shows an atomic force microscopy
image of the b-Ga2O3 surface after cleavage. Figure 48(c)
shows the thickness dependence of the ID-VGS characteris-
tics. The threshold voltage shifts from negative to positive as
the membrane thickness is reduced. Figure 48(d) shows the
extracted thickness dependent threshold voltage.684
FIG. 46. (a) DC output (IDS-VDS) characteristics, (b) DC input (IDS-VGS) as

a function of drain bias (0.1, 1, and 10 V), and (c) off-state IDS-VDS charac-
teristic showing the soft breakdown regime at around 80 V. Reprinted with
permission from Tadjer et al. ECS J. Solid State Sci. Technol. 5, P468
(2016). Copyright 2016 The Electrochemical Society.678
Moser et al.680 reported MOSFETs on Ge-doped

(4 1017 cm3) b-Ga2O3 layers grown by MBE on (010)
Fe-doped semi-insulating substrates, after the successful
demonstration of Ge-doped MOCVD Ga2O3 epilayers. The
drain current on/off ratios were >108, and the saturated drain
current was >75 mA/mm at VG ¼ 0 V. A MOSFET with a
gate-drain spacing of 5.5 lm had a three-terminal breakdown
voltage of 479 V.680 The same group680 fabricated an
enhancement-mode device achieved by gate recess with
drain-current >20 mA/mm, on/off ratio >107, and 200-V
breakdown for 3-lm source-drain distance. The Si-doped
MBE grown channel was 200 nm thick with approximately
10 nm highly doped epitaxial cap layer to reduce Ohmic con-
tact resistance.680
Green et al.687 reported the first rf performance of a
Ga2O3 MOSFET, based on Si-doped b-Ga2O3 (channel dop- FIG. 47. (Top) Extrinsic small signal RF gain performance recorded at
VGS ¼ 3.5 V (peak gm) and VDS ¼ 40 V. A gain decay of 20 dB/dec is
ing 1018 cm3 and cap doping 1019 cm3) grown by
plotted with the dashed line. (Bottom) 800 MHz Class-A power sweep of a
MOCVD on a semi-insulating CZ-grown (100) substrate. 2 50 lm gate recessed MOSFET. Reprinted with permission from Green
The devices achieved a transconductance of 21 mS/mm and et al., IEEE Electron Device Lett. 38, 790 (2017). Copyright 2017 IEEE.687
nano-belts transferred to a SiO2/Si substrate. The d-mode

transistors exhibited improved channel modulation with both
front and back gates operational compared to either front or
back-gating alone. The maximum transconductance was
approximately 4.4 mS mm1 with front and back-gating,
with a maximum drain source current density of 60 mA
mm1 at 10 V and on/off ratios of approximately 105 at
25 C. The device characteristics were stable over more than
a month for storage in air ambient.
Theoretical studies697 have suggested that these nano-
belts films will not have quantum confinement effects and
exhibit the same electronic structure as bulk material because
of states that are strongly confined near the surface.
However, if the Ga2O3 layers are clad in a wider band-gap
material such as Al2O3, quantum confinement with small
effective electron mass of electrons is expected.697
In summary, for the epi MOSFETs, the substrate has
usually been semi-insulating Fe-or Mg-doped, with the epi
grown by MBE or HVPE and doped with Sn or Si.322 The
Ohmic contacts have usually been enhanced using ion
implantation or plasma exposure to improve conductivity,
with Ti/Au-based Ohmic metallization and gates of Pt/Ti or
FIG. 48. (a) Schematic view of a GOOI FET with a 300 nm SiO2 layer on Si
Ti/Au. The gate dielectric and passivation has usually been
substrate and (b) AFM image of b-Ga2O3 surface after cleavage. (a)
Thickness dependent ID - VGS plots of various GOOI FETs from d-mode to ALD Al2O3, and both e- and d-mode operations have been
e-mode. (b) Thickness dependent VT extracted at VDS ¼ 1 V. Reprinted with reported.322 The on-off ratios have been in the range of
permission from Zhou et al. IEEE Electron Device Lett. 38, 103 (2017). 107–1010, with field effect mobilities from 2 to 95.322 The
Copyright 2017 IEEE.684
interface trap densities reported to date range from the
1011 cm2 eV1 to values well over an order of magnitude
Wong et al.689 demonstrated a vertical Ga2O3 higher. This is an area where further optimization is needed.
MOSFET, in which the source was electrically isolated from
the drain by a current blocking layer formed by Mgþ implan- Ga2O3-BASED GAS SENSORS
tation except at an aperture opening through which the drain
current flows. The IDS was modulated through a gated chan- Resistive gas sensors based on metal oxide semiconduc-
nel above this current blocking layer CBL. tors have a long history and due to its high melting point
Krishnamoorthy et al.691 reported silicon delta doping (approximately 1800 C), Ga2O3 is one of the most important
in PA-MBE b-Ga2O3 to achieve a low sheet resistance of materials for high temperature gas sensing.698–737 Gas sen-
320 X/square (mobility 83 cm2 V1 s1, sheet charge sors based on semiconducting, typically polycrystalline,
2.4 1014 cm2). This delta-doped channel was used in a Ga2O3 thin films can be used either for sensing oxygen (T P
MESFET with a maximum drain current of 236 mA/mm and 900 C) or reducing gases (T < 900 C), depending on the
transconductance of 26 mS/mm.690 Dang et al.691 showed that operating temperature.700–709 Another advantage of Ga2O3 is
a-phase corundum-structured MESFETs could be realized by that it has only one stable structural modification, mono-
Mist-CVD ON low cost sapphire substrates. clinic, that can be easily obtained by thermal annealing at
Finally, a number of groups have demonstrated air- temperatures between 800 and 900 C. This structure is then
stable high power and high temperature capable MOSFETs stable in the temperature range up to the melting point.703–706
and MESFETs on two dimensional nanobelts or nanomem- A typical application for sensors based on high-temperature-
branes that have been mechanically exfoliated from bulk stable semiconducting gallium oxide thin films is monitoring
wafers and transferred to other substrates such as, but not of the composition of exhaust gases from internal combus-
limited to, SiO2/Si.692–695 The zero or small band-gap of two tion engines or furnace installations.702–705 The resistivity of
dimensional materials like graphene and transition metal Ga2O3 thin films changes with the concentration of oxygen,
dichalcogenides materials has precluded the use of devices leading to the ability to sense oxygen. Table VIII shows
fabricated on these materials for high power and high tem- a summary of the gas sensor applications, operating
perature applications. The use of quasi-two dimensional temperature range, and detection ranges for Ga2O3 gas
Ga2O3 has potential for these applications and a potential sensors.698,699,712,715,724,730,735,737–744
large-area, controllable mechanical transfer of Ga2O3 thin Fleischer and Meixner702,703,707–709 were the first to
films onto high thermal conductivity substrates can provide a show using polycrystalline thin films sputtered from a
low-cost thermal solution for this material.696 Ahn et al.694 ceramic target that Ga2O3 is a suitable material for high tem-
reported nanobelt FETs using SiO2 and Al2O3 as the gate perature operating gas sensors. Ogita et al.698,699 used poly-
oxides for the back and front sides, respectively, which were crystalline thin films deposited by sputtering from a powder
fabricated on exfoliated two-dimensional (2D) b-Ga2O3 target and fabricated oxygen sensors with interdigitated and
TABLE VIII. Summary of Ga2O3 gas sensors and performance.
Detection Operating Detection

Sample gas temperatures ( C) range Comments Reference
Sputtered, poly H2 400–650 0.5%–3% in Ar Bulk effects due to O2 exposure present Fleischer et al.738
Pt-gated thin films H2 400–550 100 ppm Ga evaporation in O2 plasma Nakagomi et al.715
Poly H2 600 1% Can be used for O2 sensing at lower temperature Fleischer et al.745
Poly CH4 Up to 700 15% Exposure to methane leads to oxygen vacancies Becker et al.739
Poly CO 550–700 4–100 ppm Excellent stability Schwebel et al.740
Pt functionalized nanostructures CO 100 10–100 ppm Pt nanoparticles of dimension tens of nm Kim et al.735
Sputtered thin films O2 600–900 1% Used CeO2,Mn2O3, La2O3 as modifiers to obtain Schwebel et al.743
selectivity
Sputtered, poly O2 300–1000 2–10 Pa Used Si substrates as template Ogita et al.698
Sputtered, poly O2 >900 n/a Examined dependence on sputtering conditions Ogita et al.699
Thin films O2 >800 n/a Role of oxygen vacancies Baban et al.742
Sol gel thin films O2 420–460 100–10 000 ppm Effect of Ce, Sb, W or Zn doping Li et al.702
Multiple nanowires O2, CO 100–500 50–500 ppm Fast response times Liu et al.737
Pt/Ga2O3/SiC H2 310–700 0.1%–1% Schottky diode Trinchi et al.724
Nanocrystallite NH3 30 0.5 ppm Selective in both dry and humid ambients Pandeeswari et al.730
SnO2 doped sputtered films NH3 900 30 ppm Sn doping enhances sensitivity Frank et al.744
Nanobelts NO2 25 0.5–1000 ppm 30–50 nm width, 10s of microns long Lin et al.741
Nanorods Humidity 25–40 1%–95% RH Doped with Na or K Wang et al.746
mesh electrodes which showed good sensitivity and response voltage at fixed forward bias, with a faster response above
times of 14 – 27 s at 1000 C. Below 700 C, the sensors 500 C. The decrease in bias voltage for exposure to 1% H2
were found to exhibit sensitivity to reducing gases such as was 210 mV.
CO, H2, and CHx due to surface reactions while above Koroncentov and Cho736 recently reviewed the use of
900 C there is a switch to sensitivity to oxygen because of metal oxide nanocomposites and complex metal oxides,
the presence of the oxygen vacancies inside the material. including SnO2, ZnO, Ga2O3, In2O3, WO3, and the metal
Lampe et al.710 used Ga2O3 sensors in an investigation oxide modifiers including Fe2O3, La2O3, Cr2O3, Co3O4,
of both real exhaust gas and mixtures of N2, O2, CH4, CO, V2O5, NiO, CuO, SiO2, MoO3, and CeO2. The application of
NO, and water vapor to produce a synthetic exhaust gas with nanocomposites and mixed metal oxides in gas sensors
very precisely defined composition. In the range of greatly improves the sensor performance, and Ga2O3 is often
1000–900 C, the gallium oxide sensors responded to the used to increase the response of SnO2 or In2O3-based sen-
oxygen partial pressure of the mixtures.710 With the knowl- sors.736 Figure 50(a) shows the influence of additives to
edge of the fuel composition (carbon-hydrogen ratio), they SnO2 on its sensitivity to N2O at 500 C,726 with Ga2O3
were able to measure the Lambda coefficient (k) from the being one of the most effective additives. Figure 50(b) shows
relationship between air and gasoline involved in combus- the N2O conversion percentage at similar temperatures.736
tion of the mixture.710 This is a standard parameter that Rahman et al.716 explored the options for use of Ga2O3
defines the efficiency of the gasoline engine by measuring biosensors. The existing techniques for DNA biosensors
the percentage of oxygen in the exhaust. Their value was include attaching a fluorescent label to a target molecule, but
k ¼ 1.2–0.85, with a resistance jump of about three decades label-free methods are attractive for improving the speed and
at the stoichiometric point.710 size, and lowering the sensor cost.716,747 Field-effect-transis-
Bartic et al.714 used thin films and FZ single crystals of tor-based detection is attractive because of its scalability,
Ga2O3 to investigate the high temperature oxygen sensitivity integration with on-chip wireless, and ease of making arrays.
of these materials. Figure 49 shows the normalized sensitiv- A single-crystal b-Ga2O3 electrolyte/oxide diode was inves-
ity in the form of (a) the dynamic response and (b) the recov- tigated for biosensing applications.716 The surface b-Ga2O3
ery of Ga2O3 sensors subjected to a change in oxygen functionalization involved silanization with 3-aminopropyl-
content between 0 and 20% in a gas stream at 1000 C.714 triethoxysilane (APTES) and modification with N-succini-
The grain boundaries in the sputtered films played some role midyl-6-maleimidylhexanoate (EMCS) to immobilize the
in the results, but the difference in response times between capture probe DNA (thiol-terminated single-stranded DNA).
single crystal and poly material was small.714 Upon exposure to bioactive electrolytes, the functionalized
Pt/Ga2O3/SiC Schottky diodes were characterized for b-Ga2O3 could be used for detecting complementary DNA
their hydrogen gas sensitivity as a function of operating tem- sequences from noncomplementary DNA with a nearly
perature and found advantages compared to the pure thin twentyfold relative resistivity difference.716 This showed the
film Ga2O3 conductometric sensor.724 The Ga2O3 thin films feasibility of Ga2O3-based integrated bioelectronics.716
were prepared by the sol–gel process and deposited onto the The development of nanostructured materials for sensor
transducers by spin-coating. Cycling the ambient from air to materials is a result of their large active interfacial and sur-
1% H2 in produced repeatable changes of the forward face areas, which can improve the sensor response and
FIG. 50. Influence of additives to SnO2 on (a) sensitivity to 300 ppm N2O
(Toper ¼ 500 C) and (b) N2O conversion. Reprinted with permission from G.
Korotcenkov and B. K. Cho, Sens. Actuators 244, 182 (2017). Copyright
2017 Elsevier.736
FIG. 49. Dynamic response and (b) recovery of Ga2O3 sensors as a function
of change in O2 content from 0% to 20% in a gas stream at 1000 C.
Reprinted with permission from Bartic et al., Jpn. J. Appl. Phys. 45, 5186 reversible dynamic gas responses to cyclic CO exposures at
(2006). Copyright 2006 The Japan Society of Physics.714
concentrations of 50–500 ppm at 100 C. The resistance
decreases reversibly upon each CO pulse. Wang et al.746
speed. This must be balanced with sensor stability, since also showed that Ga2O3 nanorods doped with Na or K could
high surface area structures are inherently unstable due to be used for relative humidity sensors.
their high surface energy. Lin and Wang734 synthesized b-
Ga2O3 nano/microbelts under different oxygen pressures by SUMMARY AND CONCLUSIONS
thermal evaporation and measured their oxygen sensing
response at 254 nm illumination with different oxygen pres- Ga2O3 is the least mature of the wide bandgap semicon-
sures and they were able to measure rapid changes in conductors currently being examined for use in high tempera-
ductance due to chemisorption and desorption surface ture, high power electronics and solar-blind UV
processes.734 detection.87,321,322 However, its combination of materials
Pandeeswari et al.730 examined ammonia detection properties (with the notable exception of thermal conductiv-
using spray pyrolysis films of Ga2O3 at levels near the ity) and the availability of large, high quality bulk substrates
Occupational Safety and Health Administration (OSHA) makes it an attractive option and worthy of a significant
maximum recommended exposure level of 25 ppm. Existing research focus. The electronic device demonstrations have
techniques are often time consuming and require sophisti- included kV-class Schottky rectifiers with breakdown
cated instruments.730 The resistive sensors detected ammonia voltage >1 kV and e- and d-mode MOSFETs with critical
concentrations ranging from 0.5 ppm to 50 ppm and suggest field strengths larger than GaN or SiC values.87 The theory
that b-Ga2O3 thin films can be utilized to sense ammonia at of band structure, defects, and low and high field transport
room temperature for environmental monitoring and disease properties is relatively mature, and strong experimental
diagnosis through exhaled human breath.730 efforts to confirm these predictions are needed. The absence
Liu et al.711 reported the gas sensing capabilities of of solid demonstrations of p-type conductivity and the pre-
Ga2O3 nanowires to O2 and CO gases. Figure 51 shows the diction that holes are self-trapped to form polarons limits the
dynamic responses of the nanowire gas sensor to 0.5%, 1%, current range of possible devices to unipolar conductivity.
and 5% O2 at 300 C. The resistance increases upon exposure Obviously one of the biggest issues for Ga2O3 power
to oxygen ranged from factors of 2–10 times for 0.5%–5% electronics is thermal management. The approaches devel-
oxygen, respectively. The bottom of Fig. 50 shows the oped for GaN in recent years are relevant here, involving
FIG. 52. Schematics of concepts for embedded cooling of vertical current

wide bandgap semiconductor devices. Reprinted with permission from Dede
et al., in 2017 16th IEEE Intersociety Conference on Thermal and
FIG. 51. (Top) Dynamic response of the Ga2O3 nanowire sensor to O2 gas Thermomechanical Phenomena in Electronic Systems (ITherm) (2017), pp.
pulse at 300 C. (Bottom) Dynamic response of the Ga2O3 nanowire sensor 508–515. Copyright 2017 IEEE.748
to O2 gas pulse at 300 C. Reprinted with permission from Zhifu Liu et al.,
Sens. Actuators B 129, 666 (2008). Copyright 2008 Elsevier.711 factor of 6 in the total RF output power of GaN power ampli-
fiers. Similar embedded cooling approaches are critical to
embedded cooling.748–753 It was clear that the near-junction the future success of Ga2O3.
thermal barriers severely limited the capability of GaN There are obviously a large number of areas that need
power transistors and a variety of approaches including the more development and improvements in the understanding
use of diamond substrates and efficient removal of the dissi- of the basic materials science.14,87,321,322 These include the
pated power with convective and evaporative microfluidics following:
have been developed.748–753 Three of the embedded cooling
approaches employed for the reduction of the near-junction (i) Identification of the dominant defects in bulk crystals
thermal resistance are shown schematically in Fig. 52.748 and epitaxial films and the effect on device perfor-
These are used for large (approximately 1 cm2) die size verti- mance. Experimental clarification of the predictions
cal current wide bandgap devices. Concept A uses micro- on the energy levels of native defects such as oxygen
channels etched directly into the substrate. Concept B vacancies and their role in residual conductivity rela-
employs a thick metal layer as the back electrode of the ver- tive to extrinsic impurities.
tical device, with this contact containing microchannels for (ii) Experimental efforts to obtain p-type conductivity,
embedded cooling. Concept C is a stacked wafer structure utilizing non-equilibrium doping methods such as ion
that comprises the device bonded to a Si cooling chip via an implantation/cycled rapid thermal annealing. If
intermediate metallization plus thermal interface material polaron formation does preclude p-type conductivity
layer.748 For Concept C, through-Si vias are interleaved in single films, can charge separation in heterostruc-
within the Si cooling chip. This approach allows electrical tures or other novel designs be used to obtain practical
connection from the bottom electrode of the wide bandgap hole transport? Evidence of p-type conductivity
device and the lowermost electrode in the stack.748 ascribed to deep Ga vacancies has recently been
The most developed method has been deposition or reported.757 It is also known that hydrogen has been
bonding of diamond to GaN, which enabled significant found to be complexes with the VGa defect and acts to
increases in power handling capability per transistor unit passivate this important deep acceptor.758
area, while the use of microfluidic cooling enabled heat (iii) Continued development efforts in epi growth of b-
fluxes of 30 kW/cm2 at the transistor level and 1 kW/cm2 at (AlxGa1-x)2O3/Ga2O3 and b-(InxGa1x)2O3 hetero-
the die-level.749–753 This translates to an improvement of a structures growth on single crystal Ga2O3 substrates
2
to establish the stability regimes and compositional S. I. Stepanov, V. I. Nikolaev, V. E. Bougrov, and A. E. Romanov, Rev.
Adv. Mater. Sci. 44, 63 (2016); available at http://www.ipme.ru/e-
limits.
journals/RAMS/no_14416/06_14416_stepanov.pdf.
(iv) Continued progress in growth of large diameter, high 3
H. Von Wenckstern, Adv. Electron. Mater. 3, 1600350 (2017).
4
quality EFG, CZ, and FZ crystals. S. Yoshioka, H. Hayashi, A. Kuwabara, F. Oba, K. Matsunaga, and I.
(v) Novel thermal management approaches for power Tanaka, J. Phys.: Condens. Matter 19, 346211 (2007).
5
H. He, R. Orlando, M. A. Blanco, R. Pandey, E. Amzallag, I. Baraille,
devices discussed above, including layer lift-off and and M. Rerat, Phys. Rev. B 74, 195123 (2006).
wafer transfer schemes to exploit the much higher 6
H. He, M. A. Blanco, and R. Pandey, Appl. Phys. Lett. 88, 261904
thermal conductivities available with common heat- (2006).
7
sink materials like Cu, SiC, or diamond. The design P. Kroll, R. Dronskowski, and M. Martin, J. Mater. Chem. 15, 3296
(2005).
of power devices needs to incorporate accurate ther- 8
H. Y. Playford, A. C. Hannon, E. R. Barney, and R. I. Walton, Chem.-A
mal modelling. Eur. J. 19, 2803 (2013).
9
(vi) Improved Ohmic contacts through optimized surface M. Saurat and A. Revcolevschi, Rev. Int. Hautes. Refract. 8, 291 (1971).
10
J. Kohn, G. Katz, and J. D. Broder, Am. Mineral. 42, 398 (1956).
cleaning, interface tailoring, and increased doping 11
J. Ahman, G. Svensson, and J. Albertson, Acta. Crystallogr. Sect. C
capabilities. Cryst. Struct. Commun. 52, 1336 (1996).
12
(vii) Surface passivation and encapsulation techniques for M. Higashiwaki, K. Sasaki, H. Murakami, Y. Kumagai, A. Koukitu, A.
ensuring stable device operation. Kuramata, T. Masui, and S. Yamakosh, Semicond. Sci. Technol. 31,
034001 (2016).
(viii) Better understanding of carrier removal rates, domi- 13
M. Kim, J.-H. Seo, U. Singisetti, and Z. Ma, J. Mater. Chem. C 5,
nant defects created, and transient dose effects result- 8338–8354 (2017).
14
ing from radiation damage in Ga2O3 and related M. A. Mastro, A. Kuramata, J. Calkins, J. Kim, F. Ren, and S. J. Pearton,
ECS J. Solid State Sci. Technol. 6, P356 (2017).
heterostructures. 15
M. Higashiwaki, K. Sasaki, A. Kuramata, T. Masui, and S. Yamakoshi,
(ix) Interface state density mitigation processes for dielec- Phys. Status Solidi A 211, 21 (2014).
16
trics for MOS devices, as well as how these are affected M. Higashiwaki, H. Murakami, Y. Kumagai, and A. Kuramata, Jpn. J.
by process/patterning and contacting conditions.754 This 17
Appl. Phys. 55, 1202A1 (2016).
S. Fujita, Jpn. J. Appl. Phys. 54, 030101 (2015).
holds for both conventional wafers and the membranes 18
M. Higashiwaki, A. Kuramata, H. Murakami, and Y. Kumagai, J. Phys.
lifted-off by mechanical exfoliation.754,755 D: Appl. Phys. 50, 333002 (2017).
19
(x) Understanding the apparent p-type (hole) conductivity S. Jin, X. Wang, X. Wang, M. Ju, S. Shen, W. Liang, Y. Zhao, Z. Feng, H.
observed at high temperatures, mentioned earlier, ten- Y. Playford, R. I. Walton, and C. Li, J. Phys. Chem. C 119, 18221 (2015).
20
See, for example, T. Miyata, T. Nakatani, and T. Minami, J. Lumin.
tatively ascribed to deep Ga vacancy acceptors pre- 87–89, 1183 (2000); Y. Li, T. Tokizono, M. Liao, M. Zhong, Y. Koide, I.
sent in the PLD material.757 Yamada, and J. J. Delaunay, Adv. Funct. Mater. 20, 3972 (2010).
21
M. Fleischer, W. Haurierder, and H. Meixner, Thin Solid Films 190, 93
(1990).
22
ACKNOWLEDGMENTS H. J. Lin, H. Gao, and P. X. Gao, Appl. Phys. Lett. 110, 043101 (2017).
23
H. J. Lin, J. P. Baltrus, H. Gao, Y. Ding, C. Y. Nam, P. Ohodnicki, and P.
The work at UF is partially supported by HDTRA1-17- X. Gao, ACS Appl. Mater. Interfaces 8, 8880–8887 (2016).
24
1-0011 (Jacob Calkins, monitor). The project or effort L. Mazeina, F. K. Perkins, V. M. Bermudez, S. P. Arnold, and S. M.
Prokes, Langmuir 26, 13722–13726 (2010).
depicted is sponsored by the Department of the Defense, 25
M. Ogita, K. Higo, Y. Nakanishi, and Y. Hatanaka, Appl. Surf. Sci. 175,
Defense Threat Reduction Agency. The content of the 721 (2001).
26
information does not necessarily reflect the position or the E. Aubay and D. Gourier, Phys. Rev. B 47, 15023 (1993).
27
policy of the federal government, and no official L. Binet and D. Gourier, J. Phys. Chem. 100, 17630 (1996).
28
N. D. Cuong, Y. W. Park, and S. G. Yoon, Sens. Actuators, B 140, 240
endorsement should be inferred. The work at NRL was (2009).
partially supported by DTRA Grant No. HDTRA1-17-1- 29
M. Marezio and J. P. Remeika, J. Chem. Phys. 46, 1862 (1967).
30
0011 and the Office of Naval Research. The work at Korea J. P. Remeika and M. Marezio, J. Am. Chem. Soc. 8, 87 (1966).
31
E. A. Albanesi, S. J. Sferco, I. Lefebvre, G. Allan, and G. Hollinger,
University was supported by the Korea Institute of Energy Phys. Rev. B 46, 13260 (1992).
Technology Evaluation and Planning (KETEP) and the 32
L. Binet, D. Gourier, and C. Minot, J. Solid State Chem. 113, 420 (1994).
33
Ministry of Trade, Industry and Energy (MOTIE) of Korea Z. Hajnal, J. Mir o, G. Kiss, F. Reti, P. Deak, R. C. Herndon, and J. M.
(Grant No. 20172010104830) and Space Core Technology Kuperberg, J. Appl. Phys. 86, 3792 (1999).
34
K. Yamaguchi, Solid State Commun. 131, 739 (2004).
Development Program (2017M1A3A3A02015033) through 35
L. J. Eckert and R. C. Bradi, J. Am. Ceram. Soc. 56, 229 (1973).
the National Research Foundation of Korea funded by the 36
F. Litmein, D. Rached, R. Khenata, and H. Balache, J. Alloys Compd.
Ministry of Science, ICT and Future Planning of Korea. The 488, 148 (2009).
37
H. Peelaers and C. G. Van de Walle, Phys. Status Solidi B 252, 828 (2015).
authors thank their many valued colleagues for discussions 38
J. B. Varley, A. Janotti, C. Franchini, and C. G. Van de Walle, Phys. Rev.
and input, including Akito Kuramata from Tamura B 85, 081109 (2012).
39
Corporation and Novel Crystal Technology, Leonid J. Furthmuller and F. Bechstedt, Phys. Rev. B 93, 115204 (2016).
40
Chernyak and Jonathan Lee from University of Central Y. Zhang, J. Yan, G. Zhao, and W. Xie, Phys. B: Phys. Condens. Matter
405, 3899 (2010).
Florida, Soohwan Jang from Dankook University, Kwang 41
W. Guo, Y. Guo, H. Dong, and X. Zhou, Phys. Chem. Chem. Phys. 17,
Baik from Hongik University, Rohit Khanna from 5817 (2015).
42
PlasmTherm, and Chip Eddy, Jr., Neeraj Nepal, and Virginia H. H. Tippins, Phys. Rev. A 140, 316 (1965).
43
M. R. Lorentz, J. F. Woods, and R. J. Gambino, J. Phys. Chem. Solids 28,
Wheeler from Naval Research Laboratory. 403 (1967).
44
T. Matsumoto, M. Aoki, A. Kinoshita, and T. Aono, Jpn. J. Appl. Phys.
1
R. Roy, V. G. Hill, and E. F. Osburn, J. Am. Chem. Soc. 74, 719 (1952). 13, 1578 (1974).
45 78
C. Janowitz, V. Scherer, M. Mohamed, A. Krapf, H. Dwelk, R. Manzke, W. Mi, C. Luan, Z. Li, C. Zhao, X. Feng, and J. Ma, Opt. Mater. 35, 2624
Z. Galazka, R. Uecker, K. Irmscher, R. Fornari, M. Michling, D. (2013).
79
Schmeißer, J. R. Weber, J. B. Varley, and C. G. Van de Walle, New J. L. Liu, M. K. Li, D. Q. Yu, J. Zhang, H. Zhang, C. Qian, and Z. Yang,
Phys. 13, 085014 (2011). Appl. Phys. A 98, 831 (2010).
46 80
T. C. Lovejoy, E. N. Yitamben, N. Shamir, J. Morales, E. G. Vıllora, K. M. J. Tadjer, N. A. Mahadik, V. D. Wheeler, E. R. Glaser, L. Ruppalt, A.
Shimamura, S. Zheng, F. S. Ohuchi, and M. A. Olmstead, Appl. Phys. D. Koehler, K. D. Hobart, C. R. Eddy, Jr., and F. J. Kub, ECS J. Solid
Lett. 94, 081906 (2009). State Sci. Technol. 5, 468 (2016).
47 81
M. Mohamed, C. Janowitz, I. Unger, R. Manzke, Z. Galazka, R. Uecker, E. G. Villora, S. Arjoca, K. Shimamura, D. Inomata, and K. Aoki, “Light-
R. Fornari, J. R. Weber, J. B. Varley, and C. G. Van de Walle, Appl. Emitting Diodes: Materials, Devices, and Applications for Solid State
Phys. Lett 97, 211903 (2010). Lighting XX,” Proc. SPIE 9768, 976805 (2016).
48 82
C.-H. Ho, C.-Y. Tseng, and L.-C. Tien, Opt. Express 18, 16360 (2010). H. Xie, L. Chen, Y. Liu, and K. Huang, Solid State Commun. 141, 12
49
L. Dong, R. Jia, B. Xin, B. Peng, and Y. Zhang, Sci. Rep. 7, 40160 (2007).
83
(2017). T. Tsuchiya, T. Matsuura, K. Shinoda, T. Nakajima, J. Akimoto, and Y.
50
Y. Zhuo, Z. Chen, W. Tu, X. Ma, Y. Pei, and G. Wang, Appl. Surf. Sci. Idemoto, Jpn. J. Appl. Phys. 53, 05FB14 (2014).
84
420, 802 (2017). K. Sasaki, Q. Tu Thieu, D. Wakimoto, Y. Koishikawa, A. Kuramata, and
51
J. B. Varley, J. R. Weber, A. Janotti, and C. G. Van de Walle, Appl. Phys. S. Yamakoshi, Appl. Phys. Express 10, 124201 (2017).
85
Lett. 97, 142106 (2010). K. Momma and F. Izumi, J. Appl. Crystallogr. 41, 653 (2008).
52 86
X. Ma, Y. Zhang, L. Dong, and R. Jia, Results Phys. 7, 1582 (2017). H. Wang, “Investigation of power semiconductor devices for high fre-
53
L. Dong, R. Jia, C. Li, B. Xin, and Y. Zhang, J. Alloys Compd. 712, 379 quency high density power converters,” Ph.D. thesis (Virgina Tech,
(2017). 2007).
54 87
M. J. Tadjer, M. A. Mastro, N. A. Mahadik, M. Currie, V. D. Wheeler, J. G. Jessen, K. Chabak, A. Green, J. McCandless, S. Tetlak, K. Leedy, R.
A. Freitas, Jr., J. D. Greenlee, J. K. Hite, K. D. Hobart, C. R. Eddy, Jr., Fitch, S. Mou, E. Heller, S. Badescu, A. Crespo, and N. Moser, “Toward
and F. J. Kub, J. Electron. Mater. 45, 2031 (2016). realization of Ga2O3 for power electronics applications,” in 75th IEEE
55
K. Irmscher, Z. Galazka, M. Pietsch, R. Uecker, and R. Fornari, J. Appl. Device Research Conference (DRC), 2017.
88
Phys. 110, 063720 (2011). V. M. Bermudez, Chem. Phys. 323, 193 (2006).
56 89
F. Mezzadri, G. Calestani, F. Boschi, D. Delmonte, M. Bosi, and R. Y. Tomm, P. Reiche, D. Klimm, and T. Fukuda, J. Cryst. Growth 220,
Fornari, Inorg. Chem. 55, 12079 (2016). 510 (2000).
57 90
O. Ueda, N. Ikenaga, K. Koshi, K. Iizuka, A. Kuramata, K. Hanada, T. E. G. Vıllora, Y. Murakami, T. Sugawara, T. Atou, M. Kikuchi, D.
Moribayashi, S. Yamakoshi, and M. Kasu, Jpn. J. Appl. Phys. 55, Shindo, and T. Fukuda, Mater. Res. Bull. 37, 769 (2002).
91
1202BD (2016). G. Schmitz, P. Gassmann, and R. Franchy, J. Appl. Phys. 83, 2533
58
S. Fujita, M. Oda, K. Kaneko, and T. Hitora, Jpn. J. Appl. Phys. 55, (1998).
92
1202A3 (2016). A. Segura, L. Art us, R. Cusc o, R. Goldhahn, and M. Feneberg, Phys.
59
Z. Galazka, R. Uecker, D. Klimm, K. Irmscher, M. Naumann, M. Pietsch, Rev. Mater. 1, 024604 (2017).
93
A. Kwasniewski, R. Bertram, S. Ganschow, and M. Bickermann, ECS J. A. D. Becke, J. Chem. Phys. 98, 5648 (1993).
94
Solid State Sci. Technol. 6, Q3007 (2017). H. Fr€ohlich, H. Pelzer, and S. Zienau, Philos. Mag. 41, 221 (1950).
60 95
H. Y. Playford, A. C. Hannon, M. G. Tucker, D. M. Dawson, S. E. T. Onuma, S. Saito, K. Sasaki, K. Goto, T. Masui, T. Yamaguchi, T.
Ashbrook, R. J. Kastiban, J. Sloan, and R. I. Walton, J. Phys. Chem. C Honda, A. Kuramata, and M. Higashiwaki, Appl. Phys. Lett. 108, 101904
118, 16188 (2014). (2016).
61 96
M. B. Maccioni and V. Fiorentini, Appl. Phys. Exp. 9, 041102 (2016). C. Sturm, J. Furthm€ uller, F. Bechstedt, R. Schmidt-Grund, and M.
62
D. Munima, B. Sahariah, and C. Y. Kadolkar, J. Phys.: Condens. Matter Grundmann, Appl. Phys. Lett. Mater. 3, 106106 (2015).
97
19, 156215 (2007). A. Ratnaparkhe and W. R. L. Lambrecht, Appl. Phys. Lett. 110, 132103
63
Y. Oshima, E. G. Villora, Y. Matsushita, S. Yamamoto, and K. (2017).
98
Shimamura, J. Appl. Phys. 118, 085301 (2015). B. Hoeneisen, C. A. Mead, and M. A. Nicolet, Solid-State Electron. 14,
64
X. Xia, Y. Chen, Q. Feng, H. Liang, P. Tao, M. Xu, and G. Du, Appl. 1057 (1971).
99
Phys. Lett. 108, 202103 (2016). H. Yusa, T. Tsuchiya, N. Sata, and Y. Ohishi, Phys. Rev. B 77, 064107
65
M. Slomski, N. Blumenschein, P. P. Paskov, J. F. Muth, and T. Paskova, (2008).
100
J. Appl. Phys. 121, 235104 (2017). M. Passlack, E. F. Schubert, W. S. Hobson, M. Hong, N. Moriya, S. N. G.
66
N. Ueda, H. Hosono, R. Waseda, and H. Kawazoe, Appl. Phys. Lett. 71, Chu, K. Konstadinidis, J. P. Mannaerts, M. L. Schnoes, and G. J. Zydzik,
933 (1997). J. Appl. Phys. 77, 686 (1995).
67 101
M. Schubert, R. Korlacki, S. Knight, T. Hofmann, S. Sch€ oche, V. M. Rebien, W. Henrion, M. Hong, J. P. Mannaerts, and M. Fleischer,
Darakchieva, E. Janzen, B. Monemar, D. Gogova, Q.-T. Thieu, R. Appl. Phys. Lett. 81, 250 (2002).
102
Togashi, H. Murakami, Y. Kumagai, K. Goto, A. Kuramata, S. I. Bhaumik, R. Bhatt, S. Ganesamoorthy, A. Saxena, A. K. Karnal, P. K.
Yamakoshi, and M. Higashiwaki, Phys. Rev. B 93, 125209 (2016). Gupta, A. K. Sinha, and S. K. Deb, Appl. Opt. 50, 6006 (2011).
68 103
M. Yan-Mei, C. Hai-Yong, Y. Kai-Feng, L. Min, C. Qi-Liang, L. Jing, CRC Handbook of Chemistry and Physics, 70th ed., edited by R. C.
and Z. Guang-Tian, Chin. Phys. Lett. 25, 1603 (2008). Weast (CRC, 1989), p. B-92.
69 104
A. Kuramata, K. Koshi, S. Watanabe, Y. Yamaoka, T. Masui, and S. T. C. Lovejoy, E. N. Yitamben, N. Shamir, J. Morales, E. G. Villora, K.
Yamakoshi, Jpn. J. Appl. Phys. Part 1 55, 1202A2 (2016). Shimamura, S. Zheng, F. S. Ohuchi, and M. A. Olmstead, Appl. Phys.
70
H. Aida, K. Nishiguchi, H. Takeda, N. Aota, K. Sunakawa, and Y. Lett. 94, 081906 (2009).
105
Yaguchi, Jpn. J. Appl. Phys. 47(11R), 8506 (2008). H. Peelaers, D. Steiauf, J. B. Varley, A. Janotti, and C. G. Van de Walle,
71
T. Onuma, S. Saito, K. Sasaki, T. Masui, T. Yamaguchi, T. Honda, and Phys. Rev. B 92(8), 085206 (2015).
106
M. Higashiwaki, Jpn. J. Appl. Phys. 54, 112601 (2015). T. J. Coutts, D. L. Young, and X. Li, MRS Bull. 25, 58 (2000).
72 107
M. Baldini, M. Albrecht, A. Fiedler, K. Irmscher, R. Schewski, and G. L. Binet and D. Gourier, J. Phys. Chem. Solids 59, 1241 (1998).
108
Wagner, ECS J. Solid State Sci. Technol. 6, Q3040 (2017). K. W. Chang and J. J. Wu, Adv. Mater. 16, 545 (2004).
73 109
R. Schewski, G. Wagner, M. Baldini, D. Gogova, Z. Galazka, T. Schulz, Z. Hajnal, P. Deak, Th. K€ ohler, R. Kaschner, and Th. Frauenheim, Solid
T. Remmele, T. Markurt, H. von Wenckstern, M. Grundmann, O. State Commun. 108, 93 (1998).
110
Bierwagen, P. Vogt, and M. Albrecht, Appl. Phys. Express 8, 011101 T. Zacherle, P. C. Schmidt, and M. Martin, Phys. Rev. B 87, 235206
(2015). (2013).
74 111
M. Choi and J. Son, Curr. Appl. Phys. 17, 713 (2017). S. Nakagomi, S. Kubo, and Y. Kokubun, J. Cryst. Growth 445, 73 (2016).
75 112
Z. Guo, A. Verma, X. Wu, F. Sun, A. Hickman, T. Masui, A. Kuramata, X. Cailleaux, M. del Carmen Marco de Lucas, O. Merdrignac-Conanec,
M. Higashiwaki, D. Jena, and T. Luo, Appl. Phys. Lett. 106, 111909 F. Tessier, K. Nagasaka, and S. Kikkawa, J. Phys. D: Appl. Phys. 42,
(2015). 045408 (2009).
76 113
S. Ohira and N. Arai, Phys. Status Solidi C 5, 3116 (2008). Y. Xiang, C. Zhou, and W. Wang, J. Alloys Compd. 699, 1192 (2017).
77 114
T. Oshima, T. Okuno, N. Arai, Y. Kobayashi, and S. Fujita, Jpn. J. Appl. A. Kaya, H. Mao, J. Gao, R. V. Chopdekar, Y. Takamura, S. Chowdhury,
Phys. 48, 040208 (2009). and M. Saif Islam, IEEE Trans. Electron Devices 64, 2047 (2017).
115 151
A. K. Chandiran, N. Tetreault, R. Humphry-Baker, F. Kessler, E. S. Rafique, L. Han, M. J. Tadjer, J. A. Freitas, Jr., N. A. Mahadik, and H.
Baranoff, C. Yi, M. K. Nazeeruddin, and M. Gr€atzel, Nano Lett. 12, 3941 Zhao, Appl. Phys. Lett. 108, 182105 (2016).
152
(2012). G. Wagner, M. Baldini, D. Gogova, M. Schmidbauer, R. Schewski, M.
116
J.-B. Yang, T.-C. Chang, J.-J. Huang, S.-C. Chen, and M.-J. Tsai, Thin Albrecht, Z. Galazka, D. Klimm, and R. Fornari, Phys. Status Solidi A
Solid Films 529, 200 (2013). 211, 27 (2014).
117 153
T. C. Chang, F. Y. Jian, S. C. Chen, and Y. T. Tsai, Mater. Today 14, 608 H. W. Kim and N. H. Kim, Mater. Sci. Eng. B 110, 34 (2004).
154
(2011). J. Wager, Science 300(5623), 1245 (2003).
118 155
L. P. Ma, J. Liu, and Y. Yang, Appl. Phys. Lett. 80, 2997 (2002). N. M. Sbrockey, T. Salagaj, E. Coleman, G. S. Tompa, Y. Moon, and M.
119
Y. E. Syu, T. C. Chang, T. M. Tsai, Y. C. Hung, K. C. Chang, M. J. Tsai, S. Kim, J. Electron. Mater. 44, 1357 (2015).
156
M. J. Kao, and S. M. Sze, IEEE Electron Device Lett. 32, 545 (2011). R. Schewski, M. Baldini, K. Irmscher, A. Fiedler, T. Markurt, B.
120
L. W. Feng, Y. F. Chang, C. Y. Chang, T. C. Chang, S. Y. Wang, P. W. Neuschulz, T. Remmele, T. Schulz, G. Wagner, Z. Galazka, and M.
Chiang, C. C. Lin, S. C. Chen, and S. C. Chen, Thin Solid Films 519, Albrecht, J. Appl. Phys. 120, 225308 (2016).
157
1536 (2010). M. Baldini, M. Albrecht, A. Fiedler, K. Irmscher, D. Klimm, R.
121
H. Shima, F. Takano, H. Muramatsu, H. Akinaga, I. H. Inoue, and H. Schewski, and G. Wagner, J. Mater. Sci. 51, 3650 (2016).
158
Takagi, Appl. Phys. Lett. 92, 043510 (2008). M. Baldini, M. Albrecht, D. Gogova, R. Schewski, and G. Wagner,
122
K. M. Kim, B. J. Choi, and C. S. Hwang, Appl. Phys. Lett. 90, 242906 Semicond. Sci. Technol. 30, 024013 (2015).
159
(2007). L. M. Lin, Y. Luo, P. T. Lai, and K. M. Lau, Thin Solid Films 515, 2111
123
K. Jung, J. Choi, Y. Kim, H. Lm, S. Seo, R. Jung, D. C. Kim, J. S. Kim, (2006).
160
B. H. Park, and J. P. Hong, J. Appl. Phys. 103, 034504 (2008). Y. Chen, H. Liang, X. Xia, R. Shen, Y. Liu, Y. Luo, and G. Du, Appl.
124
T. Minami, H. Yamada, Y. Kubota, and T. Miyata, Jpn. J. Appl. Phys. 36, Surf. Sci. 325, 258 (2015).
161
L1191 (1997). W. Mi, C. Luan, Z. Li, C. Zhao, and J. Ma, Mater. Lett. 107, 83
125
T. Xiao, A. H. Kitai, G. Liu, A. Nakua, and J. Barbier, Appl. Phys. Lett. (2013).
162
72, 3356 (1998). Y. Lv, J. Ma, W. Mi, C. Luan, Z. Zhu, and H. Xiao, Vacuum 86, 1850
126
V. I. Vasyltsiv, Ya. I. Rym, and Ya. M. Zakharo, Phys. Stat. Sol. B 195, (2012).
163
653 (1996). X. J. Du, Z. Li, C. N. Luan, W. G. Wang, M. X. Wang, X. J. Feng, H. D.
127
Z. Galazka, K. Irmscher, R. Uecker, R. Bertram, M. Pietsch, A. Xiao, and J. Ma, J. Mater. Sci. 50, 3252 (2015).
164
Kwasniewski, M. Naumann, T. Schulz, R. Schewski, D. Klimm, and M. H. Murakami, K. Nomura, K. Goto, K. Sasaki, K. awara, Q. Tu Thieu, R.
Bickermann, J. Cryst. Growth 404, 184 (2014). Togashi, Y. Kumagai, M. Higashiwaki, A. Kuramata, S. Yamakoshi, B.
128
Z. Galazka, R. Uecker, K. Irmscher, M. Albrecht, D. Klimm, M. Pietsch, Monemar, and A. Koukitu, Appl. Phys. Express 8, 015503 (2015).
165
M. Br€utzam, R. Bertram, S. Ganschow, and R. Fornari, Cryst. Res. V. I. Nikolaev, A. I. Pechnikov, S. I. Stepanov, I. P. Nikitina, A. N.
Technol. 45, 1229 (2010). Smirnov, A. V. Chikiryaka, S. S. Sharofidinov, V. E. Bougrov, and A. E.
129
K. Irmscher, Z. Galazka, M. Pietsch, R. Uecker, and R. Fornari, J. Appl. Romanov, Mater. Sci. Semicond. Process 47, 16 (2016).
166
Phys. 110, 063720 (2011). L. Kong, J. Ma, C. Luan, and Z. Zhu, J. Solid State Chem. 184, 1946
130
E. G. Vıllora, Y. Morioka, T. Atou, T. Sugawara, M. Kikuchi, and T. (2011).
167
Fukuda, Phys. Status Solidi A193, 187 (2002). Y. Oshima, E. G. Vıllora, and K. Shimamura, J. Cryst. Growth 410, 53
131
J. Zhang, B. Li, C. Xia, G. Pei, Q. Deng, Z. Yang, W. Xu, H. Shi, F. Wu, (2015).
168
Y. Wu, and J. Xu, J. Phys. Chem. Solids 67, 2448 (2006). W. Mi, Z. Li, C. Luan, H. Xiao, and J. Ma, Ceram. Int. 41, 2572 (2015).
132 169
N. Suzuki, S. Ohira, M. Tanaka, T. Sugawara, K. Nakajima, and T. Y. Takiguchi and S. Miyajima, J. Cryst. Growth 468, 129 (2017).
170
Shishido, Phys. Status Solidi C 4, 2310 (2007). F. Mezzadri, G. Calestani, F. Boschi, D. Delmonte, M. Bosi, and R.
133
S. Ohira, N. Suzuki, N. Arai, M. Tanaka, T. Sugawara, K. Nakajima, and Fornari, Inorg. Chem. 55, 12079–12084 (2016).
171
T. Shishido, Thin Solid Films 516, 5763 (2008). B. Kramm, A. Laufer, D. Reppin, A. Kronenberger, P. Hering, A. Polity,
134
V. I. Vasyltsiv, Y. I. Rym, and Y. M. Zakharko, Phys. Status Solidi B and B. K. Meyer, Appl. Phys. Lett. 100, 094102 (2012).
172
195, 653 (1996). F. Boschi, M. Bosi, T. Berzina, E. Buffagni, C. Ferrari, and R. Fornari,
135
V. I. Nikolaev, V. Maslov, S. I. Stepanov, A. I. Pechnikov, and A. E. J. Cryst. Growth 443, 25 (2016).
173
Romanov, J. Cryst. Growth 457, 132 (2017). G. Sinha, K. Adhikary, and S. Chauduri, J. Cryst. Growth 276, 204
136
M. Mohamed, K. Irmscher, C. Janowitz, Z. Galazka, R. Manzke, and R. (2004).
174
Fornari, Appl. Phys. Lett. 101, 132106 (2012). H. Akazawa, Vacuum 123, 8 (2016).
137 175
K. Suzuki, T. Okamoto, and M. Takata, Ceram. Int. 30, 1679 (2004). H. W. Kim and N. H. Kim, J. Alloy. Compd. 389, 177 (2005).
138 176
J. Zhang, C. Xia, Q. Deng, W. Xu, H. Shi, F. Wu, and J. Xu, J. Phys. A. Navarro-Quezada, Z. Galazka, S. Alame, D. Skuridina, and N. Esser,
Chem. Solids 67, 1656 (2006). Appl. Surf. Sci. 349, 368 (2015).
139 177
L. G. Bloor, C. J. Carmall, and D. Pugh, Coord. Chem. Rev. 255, 1293 (2011). Y. Oshima, E. G. Vıllora, and K. Shimamura, Appl. Phys. Express 8,
140
E. G. Villora, K. Shimamura, Y. Yoshikawa, K. Aoki, and N. Ichinose, 055501 (2015).
178
J. Cryst. Growth 270, 420 (2004). T. Watahiki, Y. Yuda, A. Furukawa, M. Yamamuka, Y. Takiguchi, and S.
141
E. G. Villora, K. Shimamura, Y. Yoshikawa, T. Ujiie, and K. Aoki, Appl. Miyajima, Appl. Phys. Lett. 111, 222104 (2017).
179
Phys. Lett. 92, 202120 (2008). S. Jie, L. Tong, L. Jingming, Y. Jun, S. Guiying, B. Yongbiao, D.
142
E. G. Villora, K. Shimamura, T. Ujiie, and K. Aoki, Appl. Phys. Lett. 92, Zhiyuan, and Z. Youwen, J. Semicond. 37, 103004 (2016).
180
202118 (2008). H. Murakami, K. Nomura, K. Goto, K. Sasaki, K. Kawara, Q. T.
143
M. B€aumer, D. Cappus, H. Kuhlenbeck, G. Brodde, and H. Neddermeyer, Thieu, R. Togashi, Y. Kumagai, M. Higashiwaki, A. Kuramata, S.
Surf. Sci. 253, 116 (1991). Yamakoshi, B. Monemar, and A. Koukitu, Appl. Phys. Express 8,
144
K. Hoshikawa, E. Ohba, T. Kobayashi, J. Yanagisawa, C. Miyagava, and 015503 (2015).
181
Y. Nakamura, J. Cryst. Growth 447, 36 (2016). K. Nomura, K. Goto, R. Togashi, H. Murakami, and A. Koukitu, J. Cryst.
145
D. Klimm, S. Ganschow, D. Schulz, R. Bertram, R. Uecker, P. Reiche, Growth 405, 19 (2014).
182
and R. Fornari, J. Cryst. Growth 311, 534 (2009). M. Orita, H. Ohta, M. Hirano, and H. Hosono, Appl. Phys. Lett. 77, 4166
146
K. Hanada, T. Moribayashi, K. Koshi, K. Sasaki, A. Kuramata, O. Ueda, (2000).
183
and M. Kasu, Jpn. J. Appl. Phys. 55, 1202BG (2016). A. Petitmangin, B. Gallas, C. Hebert, J. Perrière, L. Binet, P. Barboux,
147
F. Zhang, K. Saito, T. Tanaka, M. Nishio, and Q. Guo, Solid State and X. Portier, Appl. Surf. Sci. 278, 153 (2013).
184
Commun. 186, 28 (2014). W. Seiler, M. Selmane, K. Abdelouhadi, and J. Perrière, Thin Solid Films
148
K. Nakai, T. Nagai, K. Noami, and T. Futagi, Jpn J. Appl. Phys. 54, 589, 556 (2015).
185
015201 (2015). S. Penner, M. St€ oger-Pollach, F. Klauser, H. Lorenz, B. Kl€ otzer, X. Liu,
149
F. Alema, B. Hertog, A. Osinsky, P. Mukhopadhyay, M. Toporkov, and B. Jenewein, and E. Bertel, Mater. Chem. Phys. 116, 175 (2009).
186
W. V. Schoenfeld, J. Cryst. Growth 475, 77 (2017). K. Matsuzaki, H. Hiramatsu, K. Nomura, H. Yanagi, T. Kamiya, M.
150
D. Gogova, G. Wagner, M. Baldini, M. Schmidbauer, K. Irmscher, R. Hirano, and H. Hosono, Thin Solid Films 496, 37 (2006).
187
Schewski, Z. Galazka, M. Albrecht, and R. Fornari, J. Cryst. Growth 401, M. Orita, H. Hiramatsu, H. Ohta, M. Hirano, and H. Hosono, Thin Solid
665 (2014). Films 411, 134 (2002).
188 226
L. M. Garten, A. Zakutayev, J. D. Perkins, B. P. Gorman, P. F. Ndione, H. Altuntas, I. Donmez, C. Ozgit-Akgun, and N. Biyikli, J. Alloys
and D. S. Ginley, MRS Commun. 6, 348 (2016). Compd. 593, 190 (2014).
189 227
F. B. Zhang, K. Saito, T. Tanaka, M. Nishio, and Q. X. Guo, J. Mater. P. C. Chang, Z. Fan, W. Y. Tseng, A. Rajagopal, and J. G. Lu, Appl.
Sci.: Mater. Electron. 26, 9624 (2015). Phys. Lett. 87, 222102 (2005).
190 228
S. M€uller, H. von Wenckstern, D. Splith, F. Schmidt, and M. Grundmann, R. S. Wagner and W. C. Ellis, Appl. Phys. Lett. 4, 89 (1964).
229
Phys. Status Solidi A 211, 34 (2014). K. K. Cho, G. B. Cho, K. W. Kim, and K. S. Ryu, Phys. Scr. T 2010,
191
F.-P. Yu, S.-L. Ou, and D.-S. Wuu, Opt. Mater. Exp. 5, 1240 (2015). 014079 (2010).
192 230
K. D. Leedy, K. D. Chabak, V. Vasilyev, D. C. Look, J. J. Boeckl, J. L. L.-W. Chang, T.-Y. Lu, Y.-L. Chen, J.-W. Yeh, and H. C. Shih, Mater.
Brown, S. E. Tetlak, A. J. Green, N. A. Moser, A. Crespo, D. B. Lett. 65, 2281 (2011).
231
Thomson, R. C. Fitch, J. P. McCandless, and G. H. Jessen, Appl. Phys. K. H. Choi, K. K. Cho, G. B. Cho, H. J. Ahn, and K. W. Kim, J. Cryst.
Lett. 111, 012103 (2017). Growth 311, 1195 (2009).
193 232
L. Nagarajan, R. A. De Souza, D. Samuelis, I. Valov, A. B€ orger, J. B. Fernandes, M. Hegde, P. C. Stanish, Z. L. Misković, and P. V.
Janek, K. D. Becker, P. C. Schmidt, and M. Martin, Nat. Mater. 7, 391 Radovanovic, Chem. Phys. Lett. 684, 135 (2017).
233
(2008). X. Liu, G. Qiu, Y. Zhao, N. Zhang, and R. Yi, J. Alloys Compd. 439, 275
194
C. Hebert, A. Petitmangin, J. Perrière, E. Millon, A. Petit, L. Binet, and (2007).
234
P. Barboux, Mater. Chem. Phys. 133, 135 (2012). G. Meligrana, W. Lueangchaichaweng, F. Col o, M. Destro, and C.
195
Z. Chen, X. Wang, S. Noda, K. Saito, and Q. Guo, Superlattices Gerbaldi, Electrochim. Acta 235, 143 (2017).
235
Microstruct. 90, 207 (2016). L. Feng, Y. Li, X. Su, S. Wang, H. Liu, J. ang, Z. Gong, W. Ding, Y.
196
T. Kawaharamura, Jpn. J. Appl. Phys. 53, 05FF08 (2014). Zhang, and F. Yun, Appl. Surf. Sci. 389, 205 (2016).
197 236
S-D. Lee, K. Kaneko, and S. Fujita, Jpn. J. Appl. Phys. 55, 1202B8 A. H. A. Makinudin, T. M. Bawazeer, N. Alsenany, M. S. Alsoufi, and A.
(2016). Supangat, Mater. Lett. 194, 53 (2017).
198 237
T. Kawaharamura, G. T. Dang, and M. Furuta, Jpn. J. Appl. Phys. 51, E. Auer, A. Lugstein, S. L€ offler, Y. J. Hyun, W. Brezna, E. Bertagnolli,
040207 (2012). and P. Pongratz, Nanotechnology 20, 434017 (2009).
199 238
G. T. Dang, T. Kawaharamura, M. Furuta, and M. W. Allen, IEEE F. Zhu, Z. X. Yang, W. M. Zhou, and Y. F. Zhang, Solid State Commun.
Electron Device Lett. 36, 463 (2015). 137, 177 (2006).
200 239
S. Fujita and K. Kaneko, J. Cryst. Growth 401, 588 (2014). Y. Huang, S. Yue, Z. Wang, Q. Wang, C. Shi, Z. Xu, X. D. Bai, C. Tang,
201
K. Akaiwa and S. Fujita, Jpn. J. Appl. Phys. 51, 070203 (2012). and C. Gu, Phys. Chem. B 110, 796 (2006).
202 240
D. Shinohara and S. Fujita, Jpn. J. Appl. Phys. 47, 7311 (2008). K. F. Cai, S. Shen, C. Yan, and S. Bateman, Curr. Appl. Phys. 8, 363
203
K. Kaneko, H. Kawanowa, H. Ito, and S. Fujita, Jpn. J. Appl. Phys. 51, (2008).
241
020201 (2012). S. P. Arnold, S. M. Prokes, F. K. Perkins, and M. E. Zaghloul, Appl.
204
D. Tahara, H. Nishinaka, S. Morimoto, and M. Yoshimoto, Jpn. J. Appl. Phys. Lett. 95, 103102 (2009).
242
Phys. 56, 078004 (2017). L. Mazeina, F. K. Perkins, V. M. Bermudez, S. P. Arnold, and S. M.
205
H. Nishinaka, D. Tahara, S. Morimoto, and M. Yoshimoto, Mater. Lett. Prokes, Langmuir 26, 13722 (2010).
243
205, 28 (2017). E. Nogales, B. Mendez, and J. Piqueras, Ultramicroscopy 111, 1037
206
T. Oshima, T. Nakazono, A. Mukai, and A. Ohtomo, J. Cryst. Growth (2011).
244
359, 60 (2012). L. Mazeina, Y. N. Picard, S. I. Maximenko, F. K. Perkins, E. R. Glaser,
207
H. Nishinaka, D. Tahara, and M. Yoshimoto, Jpn. J. Appl. Phys. 55, M. E. Twigg, J. A. Freitas, Jr., and S. M. Prokes, Cryst. Growth Des. 9,
1202BC (2016). 4471 (2009).
208 245
R. Jinno, T. Uchida, K. Kaneko, and S. Fujita, Appl. Phys. Express 9, L. Li, E. Auer, M. Y. Liao, X. S. Fang, T. Y. Zhai, U. K. Gautam, A.
071101 (2016). Lugstein, Y. Koide, Y. Bando, and D. Golberg, Nanoscale 3, 1120
209
See https://www.ut-ec.co.jp/english/portfolio/flosfia for information on (2011).
246
applications of Gallium Oxide rectifiers; See also https://spectrum.ieee. Q. T. Thieu, D. Wakimoto, Y. Koishikawa, K. Sasaki, K. Goto, K.
org/semiconductors/materials/gallium-oxide-power-electronics-cool-new-flavor Konishi, H. Murakami, A. Kuramata, Y. Kumagai, and S. Yamakoshi,
for information on power electronics applications. Jpn. J. Appl. Phys. 56, 110310 (2017).
210 247
K. Sasaki, A. Kuramata, T. Masui, E. G. Villora, K. Shimamura, and S. C. H. Hsieh, L. J. Chou, G. R. Lin, Y. Bando, and D. Golberg, Nano Lett.
Yamakoshi, Appl. Phys. Express 5, 035502 (2012). 8, 3081 (2008).
211 248
K. Sasaki, M. Higashiwaki, A. Kuramata, T. Masui, and S. Yamakoshi, E. Filippo, M. Tepore, F. Baldassarre, T. Siciliano, and A. Tepore, Appl.
J. Cryst. Growth 392, 30 (2014). Surf. Sci. 338, 69 (2015).
212 249
T. Oshima, T. Okuno, and S. Fujita, Jpn. J. Appl. Phys. 46, 7217 (2007). S. Kumar, G. Sarau, C. Tessarek, M. Y. Bashouti, A. H€ahnel, S.
213
K. Sasaki, M. Higashiwaki, A. Kuramata, T. Masui, and S. Yamakoshi, Christiansen, and R. Singh, J. Phys. D: Appl. Phys. 47, 435101 (2014).
250
J. Cryst. Growth 378, 591 (2013). M. Kumar, S. Kumar, N. Chauhan, D. S. Kumar, V. Kumar, and R.
214
H. Okumura, M. Kita, K. Sasaki, A. Kuramata, M. Higashiwaki, and J. S. Singh, Semicond. Sci. Technol. 32, 085012 (2017).
251
Speck, Appl. Phys. Express 7, 095501 (2014). Y. Guo, J. Zhang, F. Zhu, Z. X. Yang, and J. Yu, Appl. Surf. Sci. 254,
215
T. Oshima, N. Arai, N. Suzuki, S. Ohira, and S. Fujita, Thin Solid Films 5124 (2008).
252
516, 5768– 5771 (2008). H. Z. Zhang, Y. C. Kong, Y. Z. Wang, X. Due, Z. G. Bai, J. J. Wang, D.
216
A. S. Pratiyush, S. Krishnamoorthy, S. Vishnu Solanke, Z. Xia, R. P. Yu, Y. Ding, Q. L. Hang, and S. Q. Feng, Solid State Commun. 109,
Muralidharan, S. Rajan, and D. N. Nath, Appl. Phys. Lett. 110, 221107 677 (1999).
253
(2017). Y. C. Choi, W. S. Kim, Y. S. Park, S. M. Lee, D. J. Bae, Y. H. Lee, G.-S.
217
D-w. Choi, K.-B. Chung, and J.-S. Park, Thin Solid Films 546, 31 Park, W. B. Choi, N. S. Lee, and J. M. Kim, Adv. Mater. 12, 746 (2000).
254
(2013). J. Q. Hu, Q. Li, X. M. Meng, C. S. Lee, and S. T. Lee, J. Phys. Chem. B
218
C. L. Dezelah, IV, J. Niinisto, K. Arstila, L. Niinisto, and C. H. Winter, 106, 9536 (2002).
255
Chem. Mater. 18, 471 (2006). C. H. Liang, G. W. Meng, G. Z. Wang, Y. W. Wang, and L. D. Zhang,
219
T. Hisatomi, J. Brillet, M. Cornuz, F. LeFormal, N. Tetreault, K. Sivula, Appl. Phys. Lett. 78, 3202 (2001).
256
and M. Gratzel, Faraday Discuss. 155, 223 (2012). M. C. Johnson, S. Aloni, D. E. McCready, and E. D. Bourret-Courchesne,
220
G. X. Liu, F. K. Shan, J. J. Park, W. J. Lee, G. H. Lee, I. S. Kim, B. C. Cryst. Growth Des. 6, 1936 (2006).
257
Shin, and S. G. Yoon, J. Electroceram. 17, 145 (2006). A. K. Narayana Swamy, E. Shafirovich, and C. V. Ramana, Ceram. Int.
221
G. A. Battiston, R. Gerbasi, M. Porchia, R. Bertoncello, and F. 39, 7223 (2013).
258
Caccavale, Thin Solid Films 279, 115 (1996). J. Zhang, Z. Liu, C. Lin, and J. Lin, J. Cryst. Growth 280, 99 (2005).
222 259
M. Valet and D. M. Hoffman, Chem. Mater. 13, 2135 (2001). W.-S. Jung, H. U. Joo, B.-K. Min, K. W. Chang, and J. J. Wu, Physica E
223
R. Binions, C. J. Carmalt, I. P. Parkin, K. F. E. Pratt, and G. A. Shaw, 36, 226 (2007).
260
Chem. Mater. 16, 2489 (2004). N. D. Cuong, Y. W. Park, and S. G. Yoon, Sens. Actuators B 140, 240
224
M. Valet and D. M. Hoffman, Chem. Mater. 13, 2135 (2001). (2009).
225 261
V. Gottschalch, K. Mergenthaler, G. Wagner, J. Bauer, H. Paetzelt, C. J.-Y. Jung, W.-S. Cho, J.-H. Kim, K.-T. Hwang, E. T. Kang, and K.-S.
Sturm, and U. Teschner, Phys. Status Solidi A 206, 243 (2009). Han, Ceram. Int. 42, 2582 (2016).
262 304
K. W. Chang and J. J. Wu, Phys. Chem. B 108, 1838 (2004). N. Boukhari, B. Abidri, S. Hiadsi, D. Rached, M. Rabah, H. Rached, and
263
M. Baldini, M. Albrecht, A. Fiedler, K. Irmscher, D. Klimm, R. A. Abdellaoui, Comput. Mater. Sci. 50(7), 1965–1972 (2011).
305
Schewski, and G. Wagner, J. Materials Sci. 51, 3650–3656 (2016). J. E. Medvedeva, Europhys. Lett. 78, 57004 (2007).
264 306
G. Martınez-Criado, J. Segura-Ruiz, M. H. Chu, R. Tucoulou, I. Lopez, C. Vigreux, L. Binet, D. Gourier, and B. Piriou, J. Solid State Chemistry
E. Nogales, B. Mendez, and J. Piqueras, Nano Lett. 14, 5479 (2014). 157(1), 94–101 (2001).
265
Y. Luo, Z. Hou, J. Gao, D. Jin, and X. Zheng, Mater. Sci. Eng: B 140, 307 Gali, J. Miro, and P. Deak, Phys. Status Solidi A 171(2),
Z. Hajnal, A.
123 (2007). R5–R8 (1999).
266 308
T. Terasako, Y. Kawasaki, and M. Yagi, Thin Solid Films 620, 23 K. A. Klinedinst and D. A. Stevenson, J. Chem. Thermodyn. 4, 565
(2016). (1972).
267 309
C. L. Kuo and M. H. Huang, Nanotechnology 19, 155604 (2008). T. Onuma, S. Fujioka, T. Yamaguchi, Y. Itoh, and T. Honda, J. Cryst.
268
Z. R. Dai, Z. W. Pan, and Z. L. Wang, J. Phys. Chem. B 106, 902 (2002). Growth 401, 330 (2014).
269 310
Z. W. Pan, Z. R. Dai, and Z. L. Wang, Science 291, 1947 (2001). Y. Usui, T. Oya, G. Okada, N. Kawaguchi, and T. Yanagida, Mater. Res.
270
J.-S. Hwang, T.-Y. Liu, S. Chattopadhyay, G.-M. Hsu, A. M. Basilio, H.- Bull. 90, 266 (2017).
311
W. Chen, Y.-K. Hsu, W.-H. Tu, Y.-G. Lin, K.-H. Chen, C.-C. Li, S.-B. S. Rafique, L. Han, S. Mou, and H. Zhao, Opt. Mater. Express 7, 3561
Wang, H.-Y. Chen, and L.-C. Chen, Nanotechnology 24, 055401 (2013). (2017).
271 312
G. Sinha, A. Dutta, S. K. Panda, P. G. Chavan, M. A. More, D. S. Joag, Y. S. Choi, J. W. Kang, D. K. Hwang, and S. J. Park, IEEE Trans.
and A. Patra, J. Phys. D: Appl. Phys. 42, 185409 (2009). Electron Devices 57, 26 (2010).
272 313
C. Li, C. Gu, Z. Liu, J. Mi, and Y. Yang, Chem. Phys. Lett. 411, 198 S. J. Pearton and F. Ren, Int. Mater. Rev. 59, 61 (2014).
314
(2005). S. J. Pearton and F. Ren, Curr. Opin. Chem. Eng. 3, 51 (2014).
273 315
Z. Li, B. Zhao, P. Liu, and Y. Zhang, Microelectron. Eng. 85, 1613 (2008). E. M. C. Fortunato, P. M. C. Barquinha, and R. F. P. Martins, Adv.
274
Z. Zhu, T. L. Chen, Y. Gu, J. Warren, and R. M. Osgood, Chem. Mater. Mater. 24, 2945 (2012).
316
17, 4227 (2005). S.-Y. Sung, J. H. Choi, U. B. Han, K. C. Lee, J.-H. Lee, J.-J. Kim, W.
275
J. L. Johnson, Y. Choi, and A. Ural, J. Vac. Sci. Technol. B 26, 1841 (2008). Lim, S. J. Pearton, D. P. Norton, and Y.-W. Heo, Appl. Phys. Lett. 96,
276
M. Kumar, V. Kumar, and R. Singh, Nano Res. Lett. 12, 184 (2017). 102107 (2010).
277 317
S. Kumar and R. Singh, Phys. Status Solidi. Rapid Res. Lett. 7, 781 Y.-L. Wang, H. S. Kim, D. P. Norton, S. J. Pearton, and F. Ren, Appl.
(2013). Phys. Lett. 92, 112101 (2008).
278 318
S. Kumar, C. Tessarek, S. Christiansen, and R. Singh, J. Alloys Compd. Y. L. Wang, F. Ren, H. S. Kim, S. J. Pearton, and D. P. Norton, Appl.
587, 812 (2014). Phys. Lett. 90, 092116 (2007).
279 319
M. Lopez, M. Alonso-Orts, E. Nogales, B. Mendez, and J. Piqueras, Y. L. Wang, H. S. Kim, D. P. Norton, S. J. Pearton, and F. Ren,
Semicond. Sci. Technol. 31, 115003 (2016). Electrochem. Solid-State Lett. 11, H88-H91 (2008).
280 320
L. Mazeina, Y. N. Picard, and S. M. Prokes, Cryst. Growth Des. 9, 1164 Y. Q. Wu, S. Gao, and H. Huang, Mater. Sci. Semicond. Proc. 71, 321
(2009). (2017).
281 321
M. Suh, M. Meyyappan, and S. Ju, Nanotechnology 23, 305203 (2012). J. Y. Tsao, S. Chowdhury, M. A. Hollis, D. Jena, N. M. Johnson, K. A.
282
S. Phumying, S. Labauyai, W. Chareonboon, S. Phokha, and S. Maensiri, Jones, R. J. Kaplar, S. Rajan, C. G. Van de Walle, E. Bellotti, C. L.
Jpn. J. Appl. Phys. 54, 06FJ13 (2015). Chua, R. Collazo, M. E. Coltrin, J. A. Cooper, K. R. Evans, S. Graham,
283
I. D. Hosein, M. Hegde, P. D. Jones, V. Chirmanov, and P. V. T. A. Grotjohn, E. R. Heller, M. Higashiwaki, M. S. Islam, P. W.
Radovanovic, J. Cryst. Growth 396, 24 (2014). Juodawlkis, M. A. Khan, A. D. Koehler, J. H. Leach, U. K. Mishra, R.
284
X. Xu, M. Lei, K. Huang, C. Liang, and W. H. Tang, J. Alloys Compd. J. Nemanich, R. C. N. Pilawa-Podgurski, J. B. Shealy, M. J. Tadjer, A.
623, 24 (2015). F. Witulski, M. Wraback, and J. A. Simmons, “Adv. Electron. Mater.
285
G. Gundiah, A. Govindaraj, and C. N. R. Rao, Chem. Phys. Lett. 351, (published online).
322
189 (2002). B. Bayraktaroglu, “Assessment of gallium oxide technology,” Air Force
286
J. Jie, C. Wu, Y. Yu, L. Wang, and Z. Hu, Nanotechnology 20, 075602 Research Laboratory Report No. AFRL-RY-WP-TR-2017-0167, 2017.
323
(2009). K. D. Chabak, J. K. Gillespie, V. Miller, A. Crespo, J. Roussos, M. Trejo,
287
M. Kumar, G. Sarau, M. Heilmann, S. Christiansen, V. Kumar, and R. D. E. Walker, Jr., G. D. Via, G. H. Jessen, and J. Wasserbauer, IEEE
Singh, J. Phys. D: Appl. Phys. 50, 035302 (2017). Electron Device Lett. 31, 99 (2010).
288 324
E. Nogales, J. A. Garcia, B. Mendez, and J. Piqueras, J. Appl. Phys. 101, M. J. Tadjer, T. J. Anderson, K. D. Hobart, T. I. Feygelson, J. D.
033517 (2007). Caldwell, C. R. Eddy, Jr., F. J. Kub, J. E. Butler, B. Pate, and J.
289
E. Nogales, B. Mendez, J. Piqueras, and J. A. Garcia, Nanotechnology Melngailis, IEEE Electron Device Lett. 33, 23 (2012).
325
20, 115201 (2009). P.-C. Chao, K. Chu, C. Creamer, J. Diaz, T. Yurovchak, M. Shur, R.
290
I. Lopez, E. Nogales, B. Mendez, J. Piqueras, A. Peche, J. Ramırez- Kallaher, C. McGray, G. D. Via, and J. D. Blevins, IEEE Trans. Electron
Castellanos, and J. M. Gonzalez-Calbet, J. Phys. Chem. C 117, 3036 Device 62, 3658 (2015).
326
(2013). M. R. Lorenz, J. F. Woods, and R. J. Gambino, J. Phys. Chem. Solids 28,
291
X.-S. Wang, W.-S. Li, J.-Q. Situ, X.-Y. Ying, H. Chen, Y. Jin, and Y.-Z. 403 (1967).
327
Du, RSC Adv. 5, 12886 (2015). K. Nakai, T. Nagai, K. Noami, and T. Futagi, Jpn. J. Appl. Phys. 54,
292
E. Nogales, J. A. Garcia, B. Mendez, and J. Piqueras, Appl. Phys. Lett. 051103 (2015).
328
91, 133108 (2007). M. Grundmann, F. Kl€ upfel, R. Karsthof, P. Schlupp, F.-L. Schein, D.
293
E. Nogales, P. Hidalgo, K. Lorenz, B. Mendez, J. Piqueras, and E. Alves, Splith, C. Yang, S. Bitter, and H. von Wenckstern, J. Phys. D: Appl.
Nanotechnology 22, 285706 (2011). Phys. 49, 213001 (2016).
294 329
I. Lopez, E. Nogales, B. Mendez, and J. Piqueras, Appl. Phys. Lett. 100, K. H. L. Zhang, K. Xi, M. G. Blamire, and R. G. Egdell, J. Phys.:
261910 (2012). Condens. Matter 28, 383002 (2016).
295 330
C. Ronning, C. Borschel, S. Geburt, and R. Niepelt, Mater. Sci. Eng. R T. Oshima, K. Kaminaga, A. Mukai, K. Sasaki, T. Maui, A. Kuramata, S.
70, 30 (2010). Yamakoshi, S. Fujita, and A. Ohtomo, Jpn. J. Appl. Phys. 52, 051101
296
J. B. Goodenough and A. Manthiram, MRS Commun. 4, 135 (2014). (2013).
297 331
T. Harwig and F. Kellendonk, J. Solid State Chem. 24, 255 (1978). G. A. Battiston, R. Gerbasi, M. Porchia, R. Bertoncello, and F.
298
M. A. Blanco, M. B. Sahariah, H. Jiang, A. Costales, and R. Pandey, Caccavale, Thin Solid Films 279, 115 (1996).
332
Phys. Rev. B 72, 184103 (2005). Y. Li, C. Yang, L. Wu, and R. Zhang, Mod. Phys. Lett. B 31, 1750172
299
G. Sinha, D. Ganguli, and S. Chaudhuri, J. Phys.: Condens. Matter 18, (2017).
333
11167 (2006). Y. Kang, K. Krishnaswamy, H. Peelaers, and C. G. Van de Walle,
300
J. Wang, H. Zhuang, X. Zhang, S. Zhang, and J. Li, Vacuum 85, 802 J. Phys.: Condens. Matter 29, 234001 (2017).
334
(2011). E. Ahmadi, O. S. Koksaldi, S. W. Kaun, Y. Oshima, D. B. Short, U. K.
301
T. Shao, P. Zhang, L. Jin, and Z. Li, Appl. Catal. B: Environ. 142/143, Mishra, and J. S. Speck, Appl. Phys. Express 10, 041102 (2017).
335
654 (2013). X. J. Du, W. Mi, M. Jin et al. J. Cryst. Growth 404, 75–79 (2014).
302 336
A. Singhal and I. Lieberwirth, Mater. Lett. 161, 112 (2015). K. Lorenz, M. Peres, M. Felizardo, J. G. Correia, L. C. Alves, E. Alves, I.
303
S. Kumar and R. Singh, Phys. Status Solidi Rapid Res. Lett. 7, 781 (2013). Lopez, E. Nogales, B. Mendez, J. Piqueras, M. B. Barbosa, J. P. Araujo,
368
J. N. Gonçalves, J. Rodrigues, L. Rino, T. Monteiro, E. G. Villora, and K. H. S. Oon and K. Y. Cheong, Mater. Sci. Semicond. Process 16, 1217
Shimamura, Proc. SPIE 8987, 89870M (2014). (2013).
337 369
K. Sasaki, M. Higishiwaki, A. Kuramata, T. Matsui, and S. Yamakoshi, Y.-L. Chiou, L.-H. Huang, and C.-T. Lee, Semicond. Sci. Technol. 25,
Appl. Phys. Express 6, 086502 (2013). 045020 (2010).
338 370
M. Higashiwaki, K. Sasaki, A. Kuramata, T. Masui, and S. Yamakoshi, C.-T. Lee, H.-W. Chen, and H.-Y. Lee, Appl. Phys. Lett. 82, 4304 (2003).
371
Appl. Phys. Lett. 100, 013504 (2012). C.-T. Lee, H.-Y. Lee, and H.-W. Chen, IEEE Electron Device Lett. 24,
339
K. Takakura, S. Funasaki, I. Tsunoda, H. Ohyama, D. Takeuchi, T. 54 (2003).
372
Nakashima, and C. Claeys, Physica B 407, 2900 (2012). S. Pal, R. Mahapatra, S. K. Ray, B. R. Chakraborty, S. M. Shivaprasad, S.
340
M. Higashiwaki, K. Sasaki, T. Kamimura, M. H. Wong, D. K. Lahiri, and D. N. Bose, Thin Solid Films 425, 20 (2003).
373
Krishnamurthy, A. Kuramata, T. Masui, and S. Yamakoshi, Appl. Phys. Y. Zhou, C. Ahyi, T. Isaacs-Smith, M. Bozack, C.-C. Tin, J. Williams, M.
Lett. 103, 123511 (2013). Park, A.-J. Cheng, J.-H. Park, D.-J. Kim, D. Wang, E. A. Preble, A.
341
N. A. Moser, J. P. McCandless, A. Crespo, K. D. Leedy, A. J. Green, E. Hanser, and K. Evans, Solid-State Electron. 52, 756 (2008).
374
R. Heller, K. D. Chabak, N. Peixoto, and G. H. Jessen, Appl. Phys. Lett. S. D. Wolter, B. P. Luther, D. L. Waltemyer, C. Onneby, S. E. Mohney,
110, 143505 (2017). and R. J. Molnar, Appl. Phys. Lett. 70, 2156 (1997).
342 375
M. H. Wong, K. Sasaki, A. Kuramata, S. Tamakoshi, and M. S. D. Wolter, S. E. Mohney, H. Venugopalan, A. E. Wickenden, and D.
Higashiwaki, IEEE Electron Device Lett. 37, 212 (2016). D. Koleske, J. Electrochem. Soc. 145, 629 (1998).
343 376
G. Sinha, K. Adhikary, and S. Chaudhuri, J. Cryst. Growth 276, 204 H. Kim, S. Park, and H. Hwang, J. Vac. Sci. Technol. B 19, 579 (2001).
377
(2005). Y. Nakano and T. Jimbo, Appl. Phys. Lett. 82, 218 (2003).
344 378
E. Korhonen, F. Tuomisto, D. Gogova, G. Wagner, M. Baldini, Z. Y. Nakano, T. Kachi, and T. Jimbo, Appl. Phys. Lett. 82, 2443 (2003).
379
Galazka, R. Schewski, and M. Albrecht, Appl. Phys. Lett. 106, 242103 R. Korbutowicz, J. Prazmowska, Z. Wagrowski, A. Szyszka, and M.
(2015). Tlaczala, in International Conference on Advanced Semiconductor
345
K. D. Chabak, N. Moser, A. J. Green, D. E. Walker, Jr., S. E. Tetlak, E. Devices and Microsystems (2008), pp. 163–166.
380
Heller, A. Crespo, R. Fitch, J. P. McCandless, K. Leedy, M. Baldini, G. J. Brendt, D. Samuelis, T. E. Weirich, and M. Martin, Phys. Chem.
Wagner, Z. Galazka, X. Li, and G. Jessen, Appl. Phys. Lett. 109, 213501 Chem. Phys. 11, 3127 (2009).
381
(2016). P. Chen, R. Zhang, X. F. Xu, Y. G. Zhou, Z. Z. Chen, S. Y. Xie, W. P. Li,
346
A. Green, K. Chabak, E. Heller, R. Fitch, M. Baldini, A. Fiedler, K. and Y. D. Zheng, Appl. Phys. A 71, 191 (2000).
382
Irmscher, G. Wagner, Z. Galazka, S. Tetlak, A. Crespo, K. Leedy, and G. T. G. Allen and A. Cuevas, Phys. Status Solidi. Rapid Res. Lett. 9,
Jessen, IEEE Electron Device Lett. 37, 902 (2016). 220–224 (2015).
347 383
N. T. Son, K. Goto, K. Nomura, Q. T. Thieu, R. Togashi, H. Murakami, T. G. Allen and A. Cuevas, Appl. Phys. Lett. 105, 031601 (2014).
384
Y. Kumagai, A. Kuramata, M. Higashiwaki, A. Koukitu, S. Yamakoshi, S. S. Kumar, E. J. Rubio, M. Noor-A-Alam, G. Martinez, S. Manandhar,
B. Monemar, and E. Janzen, J. Appl. Phys. 120, 235703 (2016). V. Shutthanandan, S. Thevuthasan, and C. V. Ramana, J. Phys. Chem. C
348
F. B. Zhang, K. Saito, T. Tanaka, M. Nishio, and Q. X. Guo, J. Cryst. 117, 4194 (2013).
385
Growth 387, 96 (2014). M. Gr€atzel, Acc. Chem. Res. 42, 1788 (2009).
349 386
S.-D. Lee, Y. Ito, K. Kaneko, and S. Fujita, Jpn. J. Appl. Phys. 54, Beta-Gallium Oxide (b-Ga2O3) Handbook of Optical Constants of
030301 (2015). Solids, edited by D. F. Edwards (Elsevier, London, 1997), Vol. III, pp.
350
K. Akaiwa, K. Kaneko, K. Ichino, and S. Fujita, Jpn. J. Appl. Phys. 55, 753–760.
387
1202BA (2016). Y. Kokubun, S. Kubo, and S. Nakagomiet, Appl. Phys, Express 9,
351
T. Harwig, G. J. Wubs, and G. J. Dirksen, Solid State Commun. 18, 1223 091101 (2016).
388
(1976). D. Liu, Y. Guo, L. Lin, and J. Robertson, J. Appl. Phys. 114, 083704
352
A. T. Neal, S. Mou, R. Lopez, J. V. Li, D. B. Thomson, K. D. Chabak, (2013).
389
and G. H. Jessen, http://arxiv.org:443/find/cond-mat/1/au:6Mou_S/0/1/0/ F. Zhang, K. Saito, T. Tanaka, M. Nishio, M. Arita, and Q. Guo, Appl.
all/0/1 for Effects of Incomplete Ionization on Beta - Ga2O3 Power Phys. Lett. 105, 162107 (2014).
390
Devices: Unintentional Donor with Energy 110 meV. W.-Z. Xiao, L.-L. Wang, L. Xu, Q. Wan, and A.-L. Pan, Solid State
353
N. Ma, N. Tanen, A. Verma, Z. Guo, T. Luo, H. G. Xing, and D. Jena, Commun. 150, 852 (2010).
391
Appl. Phys. Lett. 109, 212101 (2016). K. Matsunaga, T. Tanaka, T. Yamamoto, and Y. Ikuhara, Phys. Rev. B
354
E. Korhonen, F. Tuomisto, O. Bierwagen, J. S. Speck, and Z. Galazka, 68, 214102 (2003).
392
Phys. Rev. B 90, 245307 (2014). Z. Yinn€u and Y. Jinliang, J. Semicond. 36, 082004 (2015).
355 393
B. E. Kananen, L. E. Halliburton, E. M. Scherrer, K. T. Stevens, G. K. C. H. Park, S. B. Zhang, and S. H. Wei, Phys. Rev. B. 66, 073202 (2002).
394
Foundos, K. B. Chang, and N. C. Giles, Appl. Phys. Lett. 111, 072102 J. L. Lyons, A. Janotti, and C. G. Van de Walle, Appl. Phys. Lett. 95,
(2017). 252105 (2009).
356 395
Y. Nakano, ECS J. Solid State Sci. Technol. 6, P615 (2017). L. Vines, C. Bhoodoo, H. von Wenckstern, and M. Grundmann, J.
357
V. Wang, W. Xiao, L.-J. Kang, R.-J. Liu, H. Mizuseki, and Y. Kawazoe, Physics: Condensed Matter (to be published).
396
J. Phys. D: Appl. Phys. 48, 015101 (2015). K. Ghosh and U. Singisetti, Appl. Phys. Lett. 109, 072102 (2016).
358 397
M. Peres, K. Lorenz, E. Alves, E. Nogales, B. Mendez, X. Biquard, B. K. Ghosh and U. Singisetti, J. Appl. Phys. 122, 035702 (2017).
398
Daudin, E. G. Vıllora, and K. Shimamura, J. Phys. D: Appl. Phys. 50, A. Parisini and R. Fornari, Semicond. Sci. Technol. 31, 035023 (2016).
399
325101 (2017). F. Ricci, F. Boschi, A. Baraldi, A. Filippetti, M. Higashiwaki, A.
359
N. F. Santos, J. Rodrigues, A. J. S. Fernandes, L. C. Alves, E. Alves, F. Kuramata, V. Fiorentini, and R. Fornari, J. Phys. Condens. Matter 28,
M. Costa, and T. Monteiro, Appl. Surf. Sci. 258, 9157 (2012). 224005 (2016).
360 400
T. Onuma, S. Fujioka, T. Yamaguchi, M. Higashiwaki, K. Sasaki, T. Z. Guo, A. Verma, X. Wu, F. Sun, A. Hickman, T. Masui, A. Kuramata,
Masui, and T. Honda, Appl. Phys. Lett. 103, 041910 (2013). M. Higashiwaki, D. Jena, and T. Luo, Appl. Phys. Lett. 106, 111909
361
T. Harwig, G. J. Wubs, and G. J. Dirksen, Solid State Commun. 18, 1223 (2015).
401
(1976). T. Omata, N. Ueda, N. Hikuma, K. Ueda, H. Mizoguchi, T. Hashimoto,
362
Y. J. Park, C. S. Oh, T. H. Yeom, and Y. M. Yu, J. Cryst. Growth 264, 1 and H. Kawazoe, Appl. Phys. Lett. 62, 499 (1993).
402
(2004). H. Peelaers and C. G. Van de Walle, Phys. Rev. B 94, 195203 (2016).
363 403
H. Zhu, R. Li, W. Luo, and X. Chen, Phys. Chem. Chem. Phys. 13, 4411 J. L. Lyons, D. Steiauf, A. Janotti, and C. G. Van de Walle, Phys. Rev.
(2011). Appl. 2, 064005 (2014).
364 404
K. Nishihagi, Z. Chen, K. Saito, T. Tanaka, and Q. Guo, Mater. Res. Bull. Y. S. Lee, D. Chua, R. E. Brandt, S. C. Sia, J. V. Li, J. P. Mailoa, S. W.
94, 170 (2017). Lee, R. G. Gordon, and T. Buonassisi, Adv. Mater. 26, 4704 (2014).
365 405
J. M. Johnson, S. Krishnamoorthy, S. Rajan, and J. Hwang, Microsc. G. Schmitz, P. Gassmann, and R. Franchy, J. Appl. Phys. 83, 2533
Microanal. 23(S1), 1454 (2017). (1998).
366 406
J. M. Johnson, S. Im, W. Windl, and J. Hwang, Ultramicroscopy 172, 17 S. W. Kaun, F. Wu, and J. S. Speck, J. Vac. Sci. Technol. A 33, 041508
(2017). (2015).
367 407
J. Hwang, J. Y. Zhang, A. J. D’Alfonso, L. J. Allen, and S. Stemmer, E. Ahmadi, Y. Oshim, F. Wu, and J. S. Speck, Semicond. Sci. Technol.
Phys. Rev. Lett. 111, 266101 (2013). 32, 035004 (2017).
408 444
Y. Oshima, E. Ahmadi, S. C. Badescu, F. Wu, and J. S. Speck, Appl. P. D. C. King and T. D. Veal, J. Phys. Condens. Matter 23, 334214
Phys. Express 9, 61102 (2016). (2011).
409 445
H. Peelaers, D. Steiauf, J. B. Varley, A. Janotti, and C. G. Van de Walle, P. D. C. King, R. L. Lichti, Y. G. Celebi, J. M. Gil, R. C. Vil~ao, H. V.
Phys. Rev. B 92, 085206 (2015). Alberto, J. P. Durte, D. J. Payne, R. B. Egdell, I. McKenzie, C. F.
410
M. B. Maccioni, F. Ricci, and V. Fiorentini, J. Phys.: Condens. Matter 28, McConville, S. F. J. Cox, and T. D. Veal, Phys. Rev. B 80, 081201
224001 (2016). (2009).
411 446
K. A. Mengle, G. Shi, D. Bayerl, and E. Kioupakisa, Appl. Phys. Lett. T. Koida, H. Fujiwara, and M. Kondo, Jpn. J. Appl. Phys. 46, L685
109, 212104 (2016). (2007).
412 447
T. Harwig, F. Kellendonk, and S. Slappendel, J. Phys. Chem. Solids 39, P. Weiser, Y. Qin, W. Yin, M. Stavola, W. Beall Fowler, and L. A.
675 (1978). Boatnerless, Appl. Phys. Lett. 109, 202105 (2016).
413 448
C. Sturm, J. Furthm€ uller, F. Bechstedt, R. Schmidt-Grund, and M. W. Yin, K. Smithe, P. Weiser, M. Stavola, W. B. Fowler, L. Boatner, S.
Grundmann, APL Mater. 3, 106106 (2015). J. Pearton, D. C. Hays, and S. G. Koch, Phys. Rev. B 91, 075208 (2015).
414 449
G.-L. Li, F. Zhang, Y.-T. Cui, H. Oji, J.-Y. Son, and Q. Guo, Appl. Phys. J. B. Bates and R. A. Perkins, Phys. Rev. B 16, 3713 (1977).
450
Lett. 107, 022109 (2015). E. Herklotz, E. V. Lavrov, and J. Weber, Phys. Rev. B 83, 235202
415
S. Ghose, M. S. Rahman, J. S. Rojas-Ramirez, M. Caro, R. Droopad, A. (2011).
451
Arias, and N. Nedev, J. Vac. Sci. Technol. B 34, 02L109 (2016). S. Yang and L. E. Halliburton, Phys. Rev. B 81, 035204 (2010).
416 452
M. Yamaga, E. G. Villora, K. Shimamura, N. Ichinose, and M. Honda, A. T. Brant, S. Yang, N. C. Giles, and L. E. Halliburton, J. Appl. Phys.
Phys. Rev. B 68, 155207 (2003). 110, 053714 (2011).
417 453
G. J. Dirksen, A. N. J. M. Hoffman, T. P. van de Bout, M. P. G. Laudy, S. Yang, A. T. Brant, N. C. Giles, and L. E. Halliburton, Phys. Rev. B 87,
and G. Blasse, J. Mater. Chem. 1, 1001 (1991). 125201 (2013).
418 454
T. Harwig, F. Kellendonk, and S. Slappendel, J. Phys. Chem. Solids 39, A. T. Brant, N. C. Giles, and L. E. Halliburton, J. Appl. Phys. 113,
675 (1978). 053712 (2013).
419 455
B. E. Kananen, L. E. Halliburton, K. T. Stevens, G. K. Foundos, and N. F. Bekisli, W. B. Fowler, and M. Stavola, Phys. Rev. B 86, 155208
C. Giles, Appl. Phys. Lett. 110, 202104 (2017). (2012).
420 456
M. Yamaga, H. Tsuzuki, S. Takano, E. G. Villora, and K. Shimamura, S. Ahn, F. Ren, E. Patrick, M. E. Law, S. J. Pearton, and A. Kuramata,
J. Non-Cryst. Solids 358, 2458–2461 (2012). Appl. Phys. Lett. 109, 242108 (2016).
421 457
S. J. Pearton, J. W. Corbett, and M. Stavola, Hydrogen in Crystalline S. Ahn, F. Ren, E. Patrick, M. E. Law, and S. J. Pearton, ECS J. Solid
Semiconductors (Springer, Berlin, 1992). State Sci. Technol. 6, Q3026 (2017).
422 458
J. B. Varley, H. Peelaers, A. Janotti, and C. G. van de Walle, J. Phys.: G. Greco, F. Iucolano, and F. Roccaforte, Appl. Surf. Sci. 383, 324
Condens. Matter 23, 334212 (2011). (2016).
423 459
S. J. Pearton, C. R. Abernathy, C. B. Vartuli, R. J. Shul, R. G. Wilson, S. N. Mohammad, J. Appl. Phys. 95, 7940 (2004).
460
and J. M. Zavada, Electron. Lett. 31(10), 836–837 (1995). D. K. Schroder, Semiconductor Material and Device Characterization
424
S. J. Pearton, R. J. Shul, R. G. Wilson, F. Ren, J. Zavada, C. R. (John Wiley and Sons, New York, 1998).
461
Abernathy, C. Vartuli, J. W. Lee, J. R. Mileham, and J. D. MacKenzie, T. K. Zywietz, J. Neugebauer, and M. Scheffler, Appl. Phys. Lett. 74,
J. Electron. Mater. 25, 845 (1996). 1695 (1999).
425 462
K. Ip, M. E. Overberg, Y. W. Heo, D. P. Norton, S. J. Pearton, S. O. J. D. Guo, C. I. Lin, M. S. Feng, F. M. Pan, G. C. Chi, and C. T. Lee,
Kucheyev, C. Jagadish, J. S. Williams, R. G. Wilson, and J. M. Zavada, Appl. Phys. Lett. 68, 235 (1996).
463
Appl. Phys. Lett. 81, 3996 (2002). J. S. Foresi and T. D. Moustakas, Appl. Phys. Lett. 62, 2859 (1993).
426 464
S. J. Pearton, S. Bendi, K. S. Jones, V. Krishnamoorthy, R. G. Wilson, F. Y. Oshima, E. Ahmadi, S. Kaun, F. Wu, and J. S. Speck, Semicond. Sci.
Ren, R. Karlicek, and R. A. Stall, Appl. Phys. Lett. 69, 1879 (1996). Technol. 33, 015013 (2018).
427 465
M. G. Weinstein, C. Song, M. Stavola, S. J. Pearton, R. G. Wilson, R. J. R. Suzuki, S. Nakagomi, and Y. Kokubun, Appl. Phys. Lett. 98, 131114
Shul, K. P. Killeen, and M. J. Ludowise, Appl. Phys. Lett. 72, 1703 (1998). (2011).
428 466
S. J. Pearton, C. R. Abernathy, R. G. Wilson, J. M. Zavada, C. Y. Song, P. Vogt and O. Bierwagen, Appl. Phys. Lett. 108, 072101 (2016).
467
M. G. Weinstein, M. Stavola, H. Han, and R. J. Shul, Nucl. Instrum. A. Ratnaparkhe and W. R. L. Lambrecht, Appl. Phys. Lett. 110, 132103
Methods Phys. Res. B147, 171 (1999). (2017).
429 468
J. Zavada, R. Wilson, F. Ren, S. J. Pearton, and R. F. Davis, Solid State K. Sasaki, M. Higashiwaki, A. Kuramata, T. Masui, and S. Yamakoshi,
Electron. 41, 677 (1997). IEEE Electron Device Lett. 34, 493 (2013).
430 469
A. Y. Polyakov, N. B. Smirnov, A. Gorokov, K. Ip, M. E. Overberg, Y. Y. Sawada, N. Weda, K. Konishi, and Y. Kumagai, “Thermodynamic
Heo, D. P. Norton, S. J. Pearton, B. Luo, F. Ren, and J. M. Zavada, analysis on molecular beam epitaxy of Ga2O3,” The 78th JSAP Autumn
J. Appl. Phys. 94, 400 (2003). Meeting, 7p-C17-9 (2017).
431 470
K. Ip, M. Overberg, Y. Heo, D. Norton, S. J. Pearton, C. E. Stutz, B. D. Splith, S. M€uller, F. Schmidt, H. von Wenckstern, J. J. van Rensburg,
Luo, F. Ren, D. C. Look, and J. M. Zavada, Appl. Phys. Lett. 82, 385 W. E. Meyer, and M. Grundmann, Phys. Status Solidi A 211, 40 (2014).
471
(2003). T. Oshima, R. Wakabayashi, M. Hattori, A. Hashiguchi, N. Kawano, K.
432
R. G. Wilson, S. J. Pearton, C. R. Abernathy, and J. M. Zavada, J. Vac. Sasaki, T. Masui, A. Kuramata, S. Yamakoshi, K. Yoshimatsu, A.
Sci. Technol. A 13, 719 (1995). Ohtomo, T. Oishi, and M. Kasu, Jpn. J. Appl. Phys. 55, 1202B7 (2016).
433 472
A. Janotti and C. G. Van de Walle, Nat. Mater. 6, 44 (2007). M. Higashiwaki, S. Sasaki, M. H. Wong, T. Kamimura, D.
434
M. A. Blanco, M. B. Sahariah, H. Jiang, A. Costales, and R. Pandey, Krishnamurthy, A. Kuramata, T. Masui, and S. Yamakoshi, in IEDM
Phys. Rev. B 72, 184103 (2005). Technical Digest (2013), p. 28.7.1.
435 473
H. Li and J. Robertson, J. Appl. Phys. 115, 203708 (2014). M. H. Wong, Y. Nakata, A. Kuramata, S. Yamakoshi, and M.
436
W. M. H. Oo, S. Tabatabaei, M. D. McCluskey, J. B. Varley, A. Janotti, Higashiwaki, Appl. Phys. Express 10, 041101 (2017).
474
and C. G. Van de Walle, Phys. Rev. B 82, 193201 (2010). K. Zeng, J. S. Wallace, C. Heimburger, K. Sasaki, A. Kuramata, T.
437
W. Beall Fowler, M. Stavola, and F. Bekisli, AIP Conf. Proc. 1583, 359 Matsui, J. A. Gardella, and U. Singisetti, IEEE Electron Device Lett. 38,
(2014). 513 (2017).
438 475
M. Stavola, F. Bekisli, W. Yin, K. Smithe, W. Beall Fowler, and L. A. K. Chabak, A. Green, N. Moser, S. Tetlak, J. McCandless, K. Leedy, R.
Boatnerless, J. Appl. Phys. 115, 012001 (2014). Fitch, A. Crespo, and G. Jessen, in DRC (2017).
439 476
F. Bekisli, M. Stavola, W. B. Fowler, L. Boatner, E. Spahr, and G. Lupke, Y. Yao, R. F. Davis, and L. M. Porter, J. Electron. Mater. 46, 2053 (2017).
477
Phys. Rev. B 84, 035213 (2011). P. H. Carey, J. Yang, F. Ren, D. C. Hays, S. J. Pearton, S. Jang, A.
440
F. Bekisli, W. B. Fowler, M. Stavola, L. A. Boatner, E. Spahr, and G. Kuramata, and I. I. Kravchenko, J. Vac. Sci. Technol. B 35, 061201 (2017).
478
L€upke, Phys. Rev. B 85, 205202 (2012). P. H. Carey, J. Yang, F. Ren, D. C. Hays, S. J. Pearton, S. Jang, A.
441
G. Alvin Shi, M. Saboktakin, M. Stavola, and S. J. Pearton, Appl. Phys. Kuramata, and I. I. Kravchenko, AIP Adv. 7, 095313 (2017).
479
Lett. 85, 5601 (2004). T. Oshima, T. Okuno, N. Arai, S. Suzuki, S. Ohira, and S. Fujita, Appl.
442
M. Stavola, AIP Conf. Proc. 671, 21 (2003). Phys. Express 1, 011202 (2008).
443 480
P. D. C. King, I. McKenzie, and T. D. Veal, Appl. Phys. Lett. 96, 062110 A. A. Mostofi, J. R. Yates, Y.-S. Lee, I. Souza, D. Vanderbilt, and N.
(2010). Marzari, Comput. Phys. Commun. 178, 685 (2008).
481 515
S. Oh, Y. Jung, M. Mastro, J. K. Hite, C. R. Eddy, and J. Kim, Opt. E. Farzana, Z. Zhang, P. K. Paul, A. R. Arehart, and S. A. Ringel, Appl.
Express 23, 28300 (2015). Phys. Lett. 110, 202102 (2017).
482 516
S. M€uller, H. von Wenckstern, F. Schmidt, D. Splith, F. L. Schein, H. M. Alonso, R. Cimino, and K. Horn, Phys. Rev. Lett. 64, 1947 (1990).
517
Frenzel, and M. Grundmann, Appl. Phys. Express 8, 121102 (2015). H. Mao, B. Alhalaili, A. Kaya, D. M. Dryden, J. M. Woodall, and M. S.
483
T. Oshima, T. Okuno, N. Arai, N. Suzuki, H. Hino, and S. Fujita, Jpn. J. Islam, Proc. SPIE 10381, 103810B (2017).
518
Appl. Phys. 48, 011605 (2009). H. Nishinaka, D. Tahara, S. Morimoto, and M. Yoshimoto, Mater. Lett.
484
R. Suzuki, S. Nakagomi, Y. Kokubun, N. Arai, and S. Ohira, Appl. Phys. 205, 28–31 (2017).
519
Lett. 94, 222102 (2009). M. J. Tadjer, V. D. Wheeler, D. I. Shahin, C. R. Eddy, Jr., and F. J. Kub,
485
H. Zhou, M. Si, S. Alghamdi, G. Qiu, L. Yang, and P. D. Ye, IEEE ECS J. Solid State Sci. Technol. 6, P165 (2017).
520
Electron Device Lett. 38, 103 (2017). K. Konishi, K. Goto, H. Murakami, Y. Kumagai, A. Kuramata, S.
486
M. Si, L. Yang, H. Zhou, and P. D. Ye, ACS Omega. 2(10), 7136–7140 Yamakosji, and M. Higashiwaki, Appl. Phys. Lett. 110, 103506 (2017).
521
(2017). M. W. Allen, R. J. Mendelsberg, R. Reeves, and S. M. Durbin, Appl.
487
K. Sasaki, “Study of homoepitaxial growth and n-type doping of b- Phys. Lett. 94, 103508 (2009).
522
Ga2O3” Ph.D. Thesis, Graduate School of Engineering, Kyoto University T. Matsumoto, M. Aoki, A. Kinoshita, and T. Aono, Jap. Appl. Phys. 13,
(2016). 1578–1582 (1974).
488 523
M. H. Wong, K. Sasaki, A. Kuramata, S. Yamakoshi, and M. S. H. M€ uller, H. von Wenckstern, F. Schmidt, D. Splith, R. Heinhold, M.
Higashiwaki, Jpn. J. Appl. Phys. 55, 1202B9 (2016). Allen, and M. Grundmann, J. Appl. Phys. 116, 194506 (2014).
489 524
K. Zeng, K. Sasaki, A. Kuramata, T. Masui, and U. Singisetti, in 74th T. Matsumoto, M. Aoki, A. Kinoshita, and T. Aono, Jpn. J. Appl. Phys.
Annual Device Research Conference (DRC), 2016 (IEEE, 2016). 13, 1578–1582 (1974).
490 525
H. Zhou, S. Alghamdi, M. W. Si, and P. D. Ye, IEEE Electron Device M. Passlack, M. Hong, J. P. Mannaerts, R. L. Opila, S. N. G. Chu, N.
Lett. 37, 1411 (2016). Moriya, F. Ren, and J. R. Kwo, IEEE Trans. Electron Devices 44,
491
D. Y. Guo, Z. P. Wu, Y. H. An, X. C. Guo, X. L. Chu, C. L. Sun, L. H. 214–225 (1997).
526
Li, P. G. Li, and W. H. Tangless, Appl. Phys. Lett. 105, 023507 (2014). F. Ren, M. Hong, J. P. Mannaerts, J. R. Lothian, and A. Y. Cho,
492
D. Y. Guo, Z. P. Wu, P. G. Li, Y. H. An, H. Liu, X. C. Guo, H. Yan, G. J. Electrochem. Soc. 144, L239 (1997).
527
F. Wang, C. L. Sun, L. H. Li, and W. H. Tang, Opt. Mater. Express 4, D. Zhuang and J. H. Edgar, Mater. Sci. Eng. R 48, 1 (2005).
528
1067 (2014). D. P. Norton, Mat. Sci. Eng R 43, 139 (2004).
493 529
K. H. Baik, S. Jang, F. Ren, and S. J. Pearton, J. Alloys Compd. 731, 118 K. H. Baik, H.-Y. Song, S.-M. Hwang, Y. Jung, J. Ahn, and J. Kim,
(2018). J. Electrochem. Soc. 158, D196 (2011).
494 530
S. M. Sze and K. K. Ng, Physics of Semiconductor Devices, 3rd ed. (John Y. Jung, J. Ahn, K. H. Baik, D. Kim, S. J. Pearton, F. Ren, and J. Kim,
Wiley & Sons, Hoboken, 2007), Chap. 3. J. Electrochem. Soc. 159, H117 (2012).
495 531
M. H. Hecht, Phys. Rev. B 41, 7918 (1990). S. E. Krasavin, Semiconductors 46, 598 (2012).
496 532
H. J. K€ummerer and G. Denninger, Magn. Reson. Chem. 43, S145 G. Blasse and A. Bril, Solid State Commun 7, 3 (1969).
533
(2005). B. E. Kananen, N. C. Giles, L. E. Halliburton, G. K. Foundos, K. B.
497
N. A. Moser, R. C. Fitch, D. E. Walker, A. J. Green, D. Chabak, Kelson, Chang, and K. T. Stevens, J. Appl. Phys. 122, 215703 (2017).
534
E. Heller, J. P. McCandless, E. Tetlak, E. Stephen, A. Crespo,K. D. J. Ahman, G. Svensson, and J. Albertsson, Acta Cryst. C 52, 1336 (1996).
535
Leedy, and G. H. Jessen, presented at U.S. Air Force Review, March S. Geller, J. Chem. Phys. 33, 676 (1960).
536
2017, Wright Paatterson AFB, Dayton OH. Available at http://www.dtic. S. Schulz, G. Bendt, W. Assenmacher, D. Sager, and G. Bacher,
mil/get-tr-doc/pdf?AD¼AD1041825 Chemical Vapor Deposition 19, 347 (2013).
498 537
M.-Y. Tsai, O. Bierwagen, M. E. White, and J. S. Speck, J. Vac. Sci. S.-L. Ou, D.-S. Wuu, Y.-C. Fu, S.-P. Liu, and Z.-C. Feng, Mater. Chem.
Technol. A: Vacuum, Surfaces, Film. 28, 354 (2010). Phys. 133, 700 (2012).
499 538
T. Oishi, Y. Koga, K. Harada, and M. Kasu, Appl. Phys. Express 8, G. T. Dang, T. Kawaharamura, M. Furuta, and M. W. Allen, IEEE Trans.
031101 (2015). Electron Devices 62, 3640 (2015).
500 539
Y. Yao, R. Gangireddy, J. Kim, K. K. Das, R. F. Davis, and L. M. Porter, Handbook of Advanced Plasma Processing Techniques, edited by R. J.
J. Vac. Sci. Technol. B 35, 03D113 (2017). Shul and S. J. Pearton (Springer, Berlin, 2000).
501 540
M. Higashiwaki, K. Konishi, K. Sasaki, K. Goto, K. Nomura, Q. T. A. P. Shah and A. Bhattacharya, J. Vac. Sci. Technol. A. 35, 041301
Thieu, R. Togashi, H. Murakami, Y. Kumagai, B. Monemar, A. Koukitu, (2017).
541
A. Kuramata, and S. Yamakoshi, Appl. Phys. Lett. 108, 133503 (2016). H. Liang, Y. Chen, X. Xia, C. Zhang, R. Shen, Y. Liu, Y. Luo, and G.
502
S. Ahn, F. Ren, L. Yuan, S. J. Pearton, and A. Kuramata, ECS J. Solid Du, Mater. Sci. Semicond. Proc. 39, 582 (2015).
542
State Sci. Technol. 6, P68 (2017). J. E. Hogan, S. W. Kaun, E. Ahmadi, Y. Oshima, and J. S. Speck,
503
A. M. Armstrong, M. H. Crawford, A. Jayawardena, A. Ahyi, and S. Semicond. Sci. Technol. 31, 065006 (2016).
543
Dhar, J. Appl. Phys. 119, 103102 (2016). L. Zhang, A. Verma, H. Grace Xing, and D. Jena, Jpn. J. Appl. Phys. 56,
504
S. Oh, G. Yang, and J. Kim, ECS J. Solid State Sci. Technol. 6, Q3022 030304 (2017).
544
(2017). J. Yang, S. Ahn, F. Ren, S. J. Pearton, R. Khanna, K. Bevlin, D.
505
Z. Zhang, E. Farzana, A. R. Arehart, and S. A. Ringel, Appl. Phys. Lett. Geerpuram, and A. Kuramata, J. Vac. Sci. Technol. B35, 031205
108, 052105 (2016). (2017).
506 545
W. Monch, J. Mater. Sci: Mater. Electron. 27, 1444 (2016). X. Cao, H. Cho, S. J. Pearton, G. T. Dang, A. P. Zhang, F. Ren, R. J.
507
T. C. Lovejoy, R. Chen, X. Zheng, E. G. Villora, K. Shimamura, H. Shul, L. Zhang, R. Hickman, and J. M. Van Hove, Appl. Phys. Lett. 75,
Yoshikawa, Y. Yamashita, S. Ueda, K. Kobayashi, S. T. Dunham, F. S. 232 (1999).
546
Ohuchi, and M. A. Olmstead, Appl. Phys. Lett. 100, 181602 (2012). X. A. Cao, S. J. Pearton, G. T. Dang, A. P. Zhang, F. Ren, and J. M. Van
508
R. Sharma, E. Patrick, M. E. Law, S. Ahn, F. Ren, S. J. Pearton, and A. Hove, IEEE Trans. Electron Devices 47, 1320 (2000).
547
Kuramata, ECS J. Solid State Sci. Technol. 6(12), P794–P797 (2017). K. Ip, K. Baik, M. E. Overberg, E. S. Lambers, Y. W. Heo, D. P. Norton,
509
J. Bae, H.-Y. Kim, and J. Kim, ECS J. Solid State Sci. Technol. 6, Q3045 S. J. Pearton, F. Ren, and J. M. Zavada, Appl. Phys. Lett. 81, 3546
(2017). (2002).
510 548
S. M€uller, H. Wenckstern, F. Schmidt, D. Splith, F.-L. Schein, H. J. Yang, S. Ahn, F. Ren, R. Khanna, K. Bevlin, D. Geerpuram, S. J.
Frenzel, and M. Grundmann, Appl. Phys. Express 8, 1211102 (2015). Pearton, and A. Kuramata, Appl. Phys. Lett. 110, 142101 (2017).
511 549
S. M€uller, H. Wenckstern, F. Schmidt, D. Splith, H. Frenzel, and M. Y. Kwon, G. Lee, S. Oh, J. Kim, S. J. Pearton, and F. Ren, Appl. Phys.
Grundmann, Semicond. Sci. Technol. 32, 065013 (2017). Lett. 110, 131901 (2017).
512 550
A. Jayawardena, A. C. Ahyi, and D. Sarit, Semicond. Sci. Technol. 31, J. Yang, F. Ren, R. Khanna, K. Bevlin, D. Geerpuram, L.-C. Tung, J. Lin,
115002 (2016). H. Jiang, J. Lee, E. Flitsiyan, L. Chernyak, S. J. Pearton, and A.
513
Q. He, W. Mu, H. Dong, S. Long, Z. Jia, H. Lv, Q. Liu, M. Tang, X. Tao, Kuramata, J. Vac. Sci. Technol. B 35, 051201 (2017).
551
and M. Liu, Appl. Phys. Lett. 110, 093503 (2017). J. Robertson and C. W. Chen, Appl. Phys. Lett. 74, 1168 (1999).
514 552
D. Zhang, W. Zheng, R. C. Lin, T. T. Li, Z. J. Zhang, and F. Huang, J. J. Robertson, J. Vac. Sci. Tech. B 18, 1785 (2000).
553
Alloys and Compounds 735, 150–154 (published online). J. Robertson, MRS Bull. 27, 217 (2002).
554 590
E. A. Kraut, R. W. Grant, J. R. Waldrop, and S. P. Kowalczyk, Phys. J. Yang, F. Ren, S. J. Pearton, G. Yang, J. Kim, and A. Kuramata, J. Vac.
Rev. Lett. 44, 1620 (1980). Sci. Technol. B 35, 031208 (2017).
555 591
D. A. Shirley, Phys. Rev. B 5, 4709 (1972). J. Lee, E. Flitsiyan, L. Chernyak, S. Ahn, F. Ren, L. Yuna, S. J. Pearton,
556
R. S. List and W. E. Spicer, J. Vac. Sci. Technol. B 6, 1228 (2016). J. Kim, B. Meyler, and J. Salzman, ECS J. Solid State Sci. Technol. 6,
557
S. Tougaard, Surf. Sci. 216, 343 (1989). Q3049 (2017).
558 592
J. C. H. Spence, Rep. Prog. Phys. 69, 725 (2006). S. Ahn, Y.-H. Lin, F. Ren, S. Oh, Y. Jung, G. Yang, J. Kim, M. A.
559
R. Carli and C. L. Bianchi, Appl. Surf. Sci. 74, 99 (1994). Mastro, J. K. Hite, C. R. Eddy, Jr., and S. J. Pearton, J. Vac. Sci. Technol.
560
D. C. Hays, B. P. Gila, S. J. Pearton, and F. Ren, Appl. Phys. Rev. 4, B 34, 041213 (2016).
593
021301 (2017). E. Wendler, E. Treiber, J. Baldauf, S. Wolf, and C. Ronnig, Nucl.
561
P. H. Carey, F. Ren, D. C. Hays, B. P. Gila, S. J. Pearton, S. Jang, and A. Instrum. Methods Phys. Res. B379, 85–90 (2016).
594
Kuramata, Appl. Surf. Sci. 422, 179 (2017). M. H. W. Akinori, T. Takahiro, M. Takeshi, O. Kohei, S. Akito, K. S.
562
P. H. Carey IV, F. Ren, D. C. Hays, B. P. Gila, S. J. Pearton, S. Jang, and Yamakoshi, and M. Higashiwaki, in Device Research Conference, South
A. Kuramata, Vacuum 141, 103 (2017). Bend, IN, June 2017.
563 595
P. H. Carey, F. Ren, D. C. Hays, B. P. Gila, S. J. Pearton, S. Jang, and A. G. Yang, S. Jang, F. Ren, S. J. Pearton, and J. Kim, ACS Appl. Mater.
Kuramata, J. Vac. Sci. Technol. B 35, 041201 (2017). Interfaces 9, 40471 (2017).
564 596
P. Carey, F. Ren, D. C. Hays, B. P. Gila, S. J. Pearton, S. Jang, and A. A. S. Pratiyush, S. Krishnamoorthy, S. V. Solanke, Z. Xia, R.
Kuramata, Vacuum 142, 52 (2017). Muralidharan, S. Rajan, and D. N. Nathless, Appl. Phys. Lett. 110,
565
P. Carey, F. Ren, D. C. Hays, B. P. Gila, S. J. Pearton, S. Jang, and A. 221107 (2017).
597
Kuramata, Jpn. J. Appl. Phys. 56, 071101 (2017). Y. L. Wu, S.-J. Chang, W. Y. Weng, C. H. Liu, T. Y. Tsai, C. L. Hsu, and
566
J. Gao, A. Kaya, R. V. Chopdekar, D. M. Dryden, Y. Takamura, M. Saif K. C. Chen, IEEE Sens. J. 13, 2368 (2013).
598
Islam, and S. Chowdhury, in Device Research Conference, South Bend, V. V. Kuryatkov, B. A. Borisov, S. A. Nikishin, Y. Kudryavtsev, R.
IN, June 2017. Asomoza, V. I. Kuchinskii, G. S. Sokolovskii, D. Y. Song, and M. Holtz,
567
K. Konishi, T. Kamimura, M. H. Wong, K. Sasaki, A. Kuramata, S. J. Appl. Phys. 100, 096104 (2006).
599
Yamakoshi, and M. Higashiwaki, Phys. Status Solidi B 253, 623 (2016). R. Wakabayashi, K. Yoshimatsu, M. Hattori, and A. Ohtomo, Appl. Phys.
568
Y. Jia, K. Zheng, J. S. Wallace, J. A. Gardella, and U. Singisetti, Appl. Lett. 111, 162101 (2017).
600
Phys. Lett. 106, 102107 (2015). R. Zou, Z. Zhang, Q. Liu, J. Hu, L. Sang, M. Liao, and W. Zhang, Small
569
T. Kamimura, K. Sasaki, M. H. Wong, D. Krishnamurthy, A. Kuramata, 10, 1848 (2014).
601
T. Masui, S. Yamakoshi, and M. Higashiwaki, Appl. Phys. Lett. 104, T.-C. Wei, D.-S. Tsai, P. Ravadgar, J.-J. Ke, M.-L. Tsai, D.-H. Lien, C.-
192104 (2014). Y. Huang, R.-H. Horng, and J.-H. He, IEEE J. Sel. Top. Quantum
570
M. Hattori, T. Oshima, R. Wakabayashi, K. Yoshimatsu, K. Sasaki, T. Electron. 20(6), 3802006 (2014).
602
Masui, A. Kuramata, S. Yamakoshi, K. Horiba, H. Kumigashira, and A. F.-P. Yu, S.-L. Ou, and D.-S. Wuu, Opt. Mater. Express 5, 1240 (2015).
603
Ohtomo, Jpn. J. Appl. Phys. 55, 1202B6 (2016). M. Ai, D. Guo, Y. Qu, W. Cui, Z. Wu, P. Li, L. Li, and W. Tang,
571
Z. Chen, K. Hishihagi, X. Wang, K. Saito, T. Tanaka, M. Nishio, M. J. Alloys Compd. 692, 634 (2017).
604
Arita, and Q. Guo, Appl. Phys. Lett. 109, 102106 (2016). X. C. Guo, N. H. Hao, D. Y. Guo, Z. P. Wu, Y. H. An, X. L. Chu, L. H.
572
W. Wei, Z. Qin, S. Fan, Z. Li, K. Shi, Q. Zhu, and G. Zhang, Nanoscale Li, P. G. Li, M. Lei, and W. H. Tang, J. Alloys Compd. 660, 136
Res. Lett. 7, 562 (2012). (2016).
573 605
S. H. Chang, Z. Z. Chen, W. Huang, X. C. Liu, B. Y. Chen, Z. Z. Li, and S. Nakagomi, T. Momo, S. Takahashi, and Y. Kokubun, Appl. Phys. Lett.
E. W. Shi, Chin. Phys. B 20, 116101 (2011). 103, 072105 (2013).
574 606
V. D. Wheeler, D. I. Shahin, M. J. Tadjer, and C. R. Eddy, Jr., ECS J. Q. Feng, L. Huang, G. Han, F. Li, X. Li, L. Fang, X. Xing, J. Zhang, W.
Solid State Sci. Technol. 6, Q3052 (2017). Mu, Z. Jia, D. Guo, W. Tang, X. Tao, and Y. Hao, IEEE Trans. Electron
575
J. Nord, K. Nordlund, and J. Keinonen, Phys. Rev. B 68, 184104 Device 63, 3578 (2016).
607
(2003). D. Patil-Chaudhari, M. Ombaba, J. Y. Oh, H. Mao, K. H. Montgomery,
576
J. R. Srour and J. W. Palko, IEEE Trans. Nucl. Sci. 53, 3610 (2006). A. Lange, S. Mahajan, J. M. Woodall, and M. Saif Islam, IEEE Photonics
577
S. J. Pearton, Y.-S. Hwang, and F. Ren, J. Mater. 67, 1601 (2015). J. 9, 1–7 (2017).
578 608
A. P. Karmarkar, B. D. White, D. Buttari, D. M. Fleetwood, R. D. Y. H. An, D. Y. Guo, S. Y. Li, Z. P. Wu, and Y. Q. Huang, J. Phys. D:
Schrimpf, R. A. Weller, L. J. Brillson, and U. K. Mishra, IEEE Trans. Appl. Phys. 49, 285111 (2016).
609
Nucl. Sci. 52, 2239 (2005). Q. Feng, X. Li, G. Han, L. Huang, F. Li, W. Tang, J. Zhang, and Y. Hao,
579
J. Chen, Y. S. Puzyrev, C. X. Zhang, E. X. Zhang, M. W. McCurdy, D. Opt. Mater. Express 7, 1240 (2017).
610
M. Fleetwood, R. D. Schrimpf, S. T. Pantelides, S. W. Kaun, E. C. H. D. Guo, X. Qin, M. Lv, H. Shi, Y. Su, G. Yao, S. Wang, C. Li, P. Li, and
Kyle, and J. S. Speck, IEEE Trans. Nucl. Sci. 60(6), 4080 - 4086 W. Tang, Electron. Mater. Lett. 13, 483–488 (2017).
611
(2013). X. Z. Liu, P. Guo, T. Sheng, L. X. Qian, and Y. R. Li, Opt. Mater. 51,
580
A. Y. Polyakov, S. J. Pearton, P. Frenzer, F. Ren, L. Liu, and J. Kim, 203 (2016).
612
J. Mater. Chem. C 1, 877 (2013). Z. Ji, J. Du, J. Fan, and W. Wang, Opt. Mater. 28, 415 (2006).
581 613
S. J. Pearton, R. Deist, F. Ren, L. Liu, A. Y. Polyakov, and J. Kim, S. Cui, Z. Mei, Y. Zhang, H. Liang, and X. Du, Adv. Opt. Mater. 5,
J. Vac. Sci. Technol. A 31, 050801 (2013). 1700454 (2017).
582 614
T. J. Anderson, A. D. Koehler, J. D. Greenlee, B. D. Weaver, M. A. F. Alema, B. Hertog, O. Ledyaev, D. Volovik, G. Thoma, R. Miller, A.
Mastro, J. K. Hite, C. R. Eddy, F. J. Kub, and K. D. Hobart, IEEE Osinsky, P. Mukhopadhyay, S. Bakhshi, H. Ali, and W. V. Schoenfeld,
Electron Dev. Lett. 35, 826 (2014). Phys. Status Solidi A 214, 1770127 (2017).
583 615
J. D. Greenlee, P. Specht, T. J. Anderson, A. D. Koehler, B. D. Weaver, Z. Liu, C. Yue, C. T. Xia, and W. L. Zhang, Chin. Phys. B 25, 017201
M. Luysberg, O. D. Dubon, F. J. Kub, T. R. Weatherford, and K. D. (2016).
616
Hobart, Appl. Phys. Lett. 107, 083504 (2015). L. Huang, Q. Feng, G. Han, F. Li, X. Li, L. Fang, X. Xing, J. Zhang, and
584
K.-H. Cho, Y.-H. Choi, J. Lim, and M.-K. Han, IEEE Trans Electron Y. Hao, IEEE Photonics J. 9, 6803708 (2017).
617
Devices 56, 365 (2009). S. Nakagomi, T.-A. Sato, Y. Takahashi, and Y. Kokubun, Sens. Actuators
585
A. Akturk, J. M. McGarrity, S. Potbhare, and N. Goldsman, IEEE Trans. A: Phys. 232, 208 (2015).
618
Nucl. Sci. 59(6), 3258–3264 (2012). W. E. Mahmoud, Sol. Energy Mater. Sol. Cells 152, 65 (2016).
586 619
N. E. Ives, J. Chen, A. F. Witulski, R. D. Schrimpf, D. M. Fleetwood, R. W. Y. Weng, T. J. Hsueh, S. J. Chang, G. J. Huang, and H. T. Hsueh,
W. Bruce, M. W. McCurdy, E. X. Zhang, and L. W. Massengill, IEEE IEEE Sens. J. 11, 999 (2011).
620
Trans. Nucl. Sci. 62, 2417 (2015). K. Zeng, K. Sasaki, A. Kuramata, T. Masui, and U. Singisetti, in
587
S. J. Pearton, F. Ren, E. Patrick, M. E. Law, and A. Y. Polyakov, ECS J. 74th IEEE Device Research Conference (Technical Digest, 2016),
Solid State Sci. Technol. 5, Q35 (2016). p. 105.
588 621
E. Patrick, M. Choudhury, F. Ren, S. J. Pearton, and M. Law, ECS J. S. Oh, J. Kim, F. Ren, S. J. Pearton, and J. Kim, J. Mater. Chem. C 4,
Solid State Sci. Technol. 4, Q21 (2015). 9245 (2016).
589 622
D. Szalkai, Z. Galazka, K. Irmscher, P. T€ utt}
o, A. Klix, and D. Gehre, S. Ahn, F. Ren, S. Oh, Y. Jung, M. A. Mastro, J. K. Hite, C. R. Eddy, Jr.,
IEEE Trans. Nucl. Sci. 64, 1574 (2017). J. Kim, and S. J. Pearton, J. Vac. Sci. Technol. B 34, 041207 (2016).
623 655
S. Oh, M. A. Mastro, M. J. Tadjer, and J. Kim, ECS J. Solid State Sci. M. Oda, J. Kikawa, A. Takatsuka, R. Tokuda, T. Sasaki, K. Kaneko, S.
Technol. 6, Q79 (2017). Fujita, and T. Hitora, in 73rd Annual Device Research Conference (DRC)
624
J. Kim, M. A. Mastro, M. J. Tadjer, and J. Kim, ACS Appl. Mater. (2015), pp 137–138.
656
Interfaces 9, 21322 (2017). K. Hanada, T. Moribayashi, T. Uematsu, S. Masuya, K. Koshi, K. Sasaki,
625
I. Lopez, A. Castaldini, A. Cavallini, E. Nogales, B. Mendez, and J. A. Kuramata, O. Ueda, and M. Kasu, Jpn. J. Appl. Phys. 55, 030303
Piqueras, J. Phys. D: Appl. Phys. 47, 415101 (2014). (2016).
626 657
L. X. Qian, X. Z. Liu, T. Sheng, W. L. Zhang, Y. R. Li, and P. T. Lai, Y. Tomm, J. M. Ko, A. Yoshikawa, and T. Fukuda, Sol. Energy Mater.
AIP Adv. 6, 045009 (2016). Sol. Cells 66, 369 (2001).
627 658
Y. H. An, D. Y. Guo, Z. M. Li, Z. P. Wu, Y. S. Zhi, W. Cui, X. L. Zhao, M. Kasu, K. Hanada, T. Moribayashi, A. Hashiguchi, T. Oshima, T.
P. G. Li, and W. H. Tang, RSC Adv. 6, 66924 (2016). Oishi, K. Koshi, K. Sasaki, A. Kuramata, and O. Ueda, Jpn. J. Appl.
628
B. Janjua, H. Sun, C. Zhao, D. H. Anjum, D. Priante, A. A. Alhamoud, F. Phys. 55, 1202BB (2016).
659
Wu, X. Li, A. M. Albadri, A. Y. Alyamani, M. M. El-Desouki, T. K. Ng, T. Oshima, A. Hashiguchi, T. Moribayashi, K. Koshi, K. Sasaki, A.
and B. S. Ooi, Opt. Express 25, 1381 (2017). Kuramata, O. Ueda, T. Oishi, and M. Kasu, Jpn. J. Appl. Phys. 56,
629
G. C. Hu, C. X. Shan, N. Zhang, M. M. Jiang, S. P. Wang, and D. Z. 086501 (2017).
660
Shen, Opt. Express 23, 13554 (2015). S. Rafique, L. Han, A. T. Neal, S. Mou, J. Boeckl, and H. Zhao, Physica
630
Z. Cheng, M. Hanke, P. Vogt, O. Bierwagen, and A. Trampertless, Appl. Status Solidi A (to be published).
661
Phys. Lett. 111, 162104 (2017). M. Kasu, T. Oshima, K. Hanada, T. Moribayashi, A. Hashiguchi, T.
631
K. H. Baik, Y. Irokawa, F. Ren, S. J. Pearton, S. S. Park, and Y. J. Park, Oishi, K. Koshi, K. Sasaki, A. Kuramata, and O. Ueda, Jpn. J. Appl.
Solid-State Electron 47, 975 (2003). Phys. 56, 091101 (2017).
632 662
K. H. Baik, F. Ren, Y. Irokawa, S. J. Pearton, S. S. Park, and S. K. Lee, F. J. Bruni, C.-M. Liu, and J. Stone-Sundberg, Acta Phys. Pol. A 124, 213
“Design and fabrication of GaN high power rectifiers,” in Optoelectronic (2013).
663
Devices: III Nitrides (Elsevier, Berlin, 2005), Chap. 12, pp. 323–350. E. Ohba, T. Kobayashi, M. Kado, and K. Hoshikawa, Jpn. J. Appl. Phys.
633
X. A. Cao, H. Lu, S. F. LeBoeuf, C. Cowen, S. D. Arthur, and W. 55, 1202BF (2016).
664
Wangless, Appl. Phys. Lett. 87, 053503 (2005). A. D. Koehler, T. J. Anderson, M. J. Tadjer, A. Nath, B. N. Feigelson, D.
634
K. H. Baik, Y. Irokawa, J. Kim, J. R. LaRoche, F. Ren, S. S. Park, Y. J. I. Shahin, K. D. Hobart, and F. J. Kub, ECS J. Solid State Sci. Technol. 6,
Park, and S. J. Pearton, Appl. Phys. Lett. 83, 3192 (2003). Q10 (2017).
635 665
Y. Zhang, M. Sun, Z. Liu, D. Piedra, J. Hu, X. Gao, and T. Palacios, M. Kumar, S. Kumar, N. Chauhan, D. Sakthi Kumar, V. Kumar, and R.
Appl. Phys. Lett. 110, 193506 (2017). Singh, Semicond. Sci. Technol. 32, 085012 (2017).
636 666
Y. Cao, R. Chu, R. Li, M. Chen, R. Chang, and B. Hughes, Appl. Phys. S. Nakagomi, K. Hiratsuka, Y. Kakuda, and K. Yoshihiro, ECS J. Solid
Lett. 108, 062103 (2016). State Sci. Technol. 6, Q3030 (2017).
637 667
Y. Zhang, Z. Liu, M. J. Tadjer, M. Sun, D. Piedra, C. Hatem, T. J. R. Schewski, M. Baldini, K. Irmscher, A. Fiedler, T. Markurt, B.
Anderson, L. E. Luna, A. Nath, A. D. Koehler, H. Okumura, J. Hu, X. Neuschulz, T. Remmele, T. Schulz, G. Wagner, Z. Galazka, and M.
Zhang, X. Gao, B. N. Feigelson, K. D. Hobart, and T. Palacios, IEEE Albrecht, J. Appl. Phys. 120, 225308 (2016).
668
Electron Device Lett. 38, 1097 (2017). S. Krishnamoorthy, Z. Xia, C. Joishi, Y. Zhang, J. McGlone, J. Johnson,
638
T.-T. Kao, J. Kim, Y.-C. Lee, A. F. M. S. Haq, M.-H. Ji, T. Detchprohm, M. Brenner, A. R. Arehart, J. Hwang, S. Lodha, and S. Rajanless, Appl.
R. D. Dupuis, and S.-C. Shen, IEEE Trans. Electron Devices 62, 2679 Phys. Lett. 111, 023502 (2017).
669
(2015). X. Zhao, Z. Wu, Y. Zhi, Y. An, W. Cui, L. Li, and W. Tang, J. Phys. D:
639
A. Barkley, M. Schupbach, B. Agrawal, and S. Allen, in 2017 IEEE Appl. Phys. 50, 085102 (2017).
670
Applied Power Electronics Conference and Exposition (APEC) (2017), O. Takayoshi, K. Yuji, K. Naoto, K. Akito, Y. Shigenobu, F. Shizuo, O.
pp. 2165–2172. Toshiyuk, and K. Makoto, Appl. Phys. Express 10, 035701 (2017).
640 671
A. Q. Huang, Proc. IEEE 105(11), 2019–2047 (2017). F. Zhang, K. Saito, T. Tanaka, M. Nishio, M. Arita, and Q. Guo, Appl.
641
X. She, A. Q. Huang, O. Lucıa, and B. Ozpineci, IEEE Trans. Ind. Phys. Lett. 105, 162107 (2014).
672
Electron 64, 8193 (2017). C. Kranert, M. Jenderka, J. Lenzner, M. Lorenz, H. von Wenckstern,
642
H. Sun, C. G. Torres Castanedo, K. Liu, K.-H. Li, W. Guo, R. Lin, X. R. Schmidt-Grund, and M. Grundmann, J. Appl. Phys. 117, 125703
Liu, J. Li, and X. Li, Appl. Phys. Lett. 111, 162105 (2017). (2015).
643 673
J. Kim, M. A. Mastro, M. J. Tadjer, and J. Kim, ACS Appl. Mater. T. Oshima, T. Okuno, N. Arai, Y. Kobayashi, and S. Fujita, Jpn. J. Appl.
Interfaces 9(25), 21322–21327 (2017). Phys. 48, 070202 (2009).
644 674
D. Hu, Y. Wang, S. Zhuang, X. Dong, Y. Zhang, J. Yin, B. Zhang, Y. A Paskaleva, D Spassov, and P Terziyska, J. Phys.: Conf. Series 794,
Lv, Z. Feng, and G. Du, Ceram. Int. 44(3) 3122–3127 (published 012017 (2017).
675
online). H. Ito, K. Kaneko, and S. Fujita, Jpn. J. Appl. Phys. 51, 100207 (2012).
645 676
Y. P. Qian, D. Y. Guo, X. L. Chu, H. Z. Shi, W. K. Zhu, K. Wang, X. K. J. B. Varley and A. Schleife, Semicond. Sci. Technol. 30, 024010
Huang, H. Wang, S. L. Wang, P. G. Li, X. H. Zhang, and W. H. Tang, (2015).
677
Mater. Lett. 209, 558–561 (2017). G. Kresse and J. Furthm€ uller, Phys. Rev. B 54, 11169 (1996).
646 678
M. Pavesi, F. Fabbri, F. Boschi, G. Piacentini, A. Baraldi, M. Bosi, E. M. J. Tadjer, N. A. Mahadik, V. D. Wheeler, E. R. Glaser, L. Ruppalt, A.
Gombia, A. Parisini, and R. Fornari, Mater. Chem. Phys. 205, 502–507 D. Koehler, K. D. Hobart, C. R. Eddy, Jr., and F. J. Kub, ECS J. Solid
(published online). State Sci. Technol. 5, 468 (2016).
647 679
Y. Q. Huang, Y. H. An, Z. P. Wu, D. Y. Guo, Y. S. Zhi, W. Cui, X. L. A. Fiedler, R. Schewski, M. Baldini, Z. Galazka, G. Wagner, M.
Zhao, and W. H. Tang, J. Alloys Compd. 717, 8–13 (2017). Albrecht, and K. Irmscher, J. Appl. Phys. 122, 165701 (2017).
648 680
J. Yang, S. Ahn, F. Ren, S. J. Pearton, S. Jang, J. Kim, and A. Kuramata, N. Moser, J. McCandless, A. Crespo, K. Leedy, A. Green, A. Neal, S.
Appl. Phys. Lett. 110, 192101 (2017). Mou, E. Ahmadi, J. Speck, K. Chabak, N. Peixoto, and G. Jessen, IEEE
649
J. Yang, S. Ahn, F. Ren, S. Jang, S. J. Pearton, and A. Kuramata, IEEE Electron Device Lett. 38, 775 (2017).
681
Electron Device Lett. 38, 906 (2017). Y. Zhu, Q. K. Yu, G. Q. Ding, X. G. Xu, T. R. Wu, Q. Gong, N. Y. Yuan,
650
M. Oda, R. Tokuda, H. Kambara, T. Tanikawa, T. Sasaki, and T. Hitora, J. N. Ding, S. M. Wang, X. M. Xie, and M. H. Jiang, Nanoscale Res. Lett.
Appl. Phys. Express 9, 021101 (2016). 9, 48 (2014).
651 682
Q. Liu, D. Guo, K. Chen, Y. Su, S. Wang, P. Li, and W. Tang, J. Alloys K. Zeng and U. Singisetti, Appl. Phys. Lett. 111, 122108 (2017).
683
Compd. 731, 1225–1229 (published online). N. Alfaraj, S. Mitra, F. Wu, I. Ajia, B. Janjua, A. Prabaswara, R. A.
652
T. Oishi, K. Harada, Y. Koga, and M. Kasu, Jpn. J. Appl. Phys. 55, Aljefri, H. Sun, T. K. Ng, B. S. Ooi, I. S. Roqan, and X. Li, Appl. Phys.
030305 (2016). Lett. 110, 161110 (2017).
653 684
E. Ohba, T. Kobayashi, M. Kado, and K. Hoshikawa, Jpn. J. Appl. Phys. Y. H. An, Y. S. Zhi, W. Cui, X. L. Zhao, Z. P. Wu, D. Y. Guo, P. G. Li,
55, 1202BF (2016). and W. H. Tang, J. Nanosci. Nanotechnol. 17(12), 9091–9094 (2017).
654 685
K. Sasaki, D. Wakimoto, Q. T. Thieu, Y. Koishikawa, A. Kuramata, M. M. Baldini, Z. Galazka, and G. Wagner, “Recent progress in the growth
Higashiwaki, and S. Yamakoshi, in Demonstration of Ga2O3 Trench of b-Ga2O3 for power electronics applications,” Mater. Sci. Semicond.
MOS-Type Schottky Barrier Diodes, Device Research Conference, South Process. (published online).
686
Bend, IN, June 2017. T. D. Moustakas and R. Paiella, Rep. Prog. Phys. 80, 106501 (2017).
687 722
A. J. Green, K. D. Chabak, M. Baldini, N. Moser, R. Gilbert, R. C. Fitch, F. Li, X. Han, and S. Liu, Biosens. Bioelectron. 26, 2619 (2011).
723
G. Wagner, Z. Galazka, J. Mccandless, A. Crespo, K. Leedy, and G. H. A. K. Cheng, B. Ge, and H. Z. Yu, Anal. Chem. 79, 5158 (2007).
724
Jessen, IEEE Electron Device Lett. 38, 790 (2017). A. Trinchi, W. Wlodarski, and Y. X. Li, Sens. Actuators B 100, 94
688
K. Chabak, A. Green, N. Moser, S. Tetlak, J. McCandless, K. Leedy, R. (2004).
725
Fitch, A. Crespo, and G. Jessen, “Gate-recessed, laterally-scaled b-Ga2O3 Y. Yang and P. Zhang, Phys. Lett. A 374, 4169 (2010).
726
MOSFETs with high-voltage enhancement-mode operation,” in IEEE W. Jochum, S. Penner, K. F€ ottinger, R. Kramer, and B. Kl€
otzer, J. Catal.
Device Research Conference (DRC), South Bend, IN, 2017. 256, 268 (2008).
689 727
M. H. Wong, K. Goto, A. Kuramata, S. Yamakoshi, H. Murakami, Y. V. Stambouli, M. Labeau, I. Matko, B. Chenevier, O. Renault, C.
Kumagai, and M. Higashiwaki, “First demonstration of vertical Ga2O3 Guiducci, P. Chaudo€ u, H. Roussel, D. Nibkin, and E. Dupuisa, Sens.
MOSFET: Planar structure with a current aperture,” in IEEE Device Actuators B 113, 1025 (2006).
Research Conference, South Bend IN, 2017. 728
A. Zebda, V. Stambouli, M. Labeau, C. Guiducci, J.-P. Diard, and B. Le
690
Z. Hu, K. Nomoto, W. Li, L. J. Zhang, J.-H. Shin, N. Tanen, T. Gorrec, Biosens. Bioelectron. 22, 178 (2006).
Nakamura, D. Jena, and H. G. Xing, “Vertical fin Ga2O3 power field- 729
V. Stambouli, A. Zebda, E. Appert, C. Guiducci, M. Labeau, J.-P. Diard, B.
effect transistors with on/off ratio > 109,” in IEEE Device Research Le Gorrec, N. Brack, and P. J. Pigram, Electrochim. Acta 51, 5206 (2006).
Conference, South Bend IN, 2017. 730
R. Pandeeswari and B. G. Jeyaprakash, Sens. Actuators B 195, 206
691
S. Krishnamoorthy, Z. Xia, S. Bajaj, M. Brenner, and S. Rajan, Appl. (2014).
Phys. Express 10, 051102 (2017). 731
J. Zeng and U. J. Krull, Chim. Oggi 21, 48 (2003).
692
M. Peres, K. Lorenz, E. Alves, E. Nogales, B. Mendez, X. Biquard, B. 732
J. Xu, J. J. Zhu, Q. Huang, and H. Y. Chenal, Electrochem. Commun. 3,
Daudin, E. G. Vıllora, and K. Shimamura, J. Phys. D: Appl. Phys. 50, 665 (2001).
325101 (2017). 733
I. V. Yang and H. H. Thorp, Anal. Chem. 73, 5316 (2001).
693
W. S. Hwang, A. Verma, H. Peelaers, V. Protasenko, S. Rouvimov, H. 734
M. Hai-Lin and F. Duo-Wang, Chin. Phys. Lett. 26, 117302 (2009).
(Grace) Xing, A. Seabaugh, W. Haensch, C. Van de Walle, Z. Galazka, 735
H. Kim, C. Jin, S. An, and C. Lee, Ceram. Int. 38, 3563 (2012).
M. Albrecht, R. Fornari, and D. Jena, Appl. Phys. Lett. 104, 203111 736
G. Korotcenkov and B. K. Cho, Sens. Actuators B 244, 182 (2017).
(2014). 737
W. Cui, X. L. Zhao, Y. H. An, D. Y. Guo, X. Y. Qing, Z. P. Wu, P. G. Li,
694
S. Ahn, F. Ren, J. Kim, S. Oh, J. Kim, M. A. Mastro, and S. J. Pearton, L. H. Li, C. Cui, and W. H. Tang, J. Phys. D: Appl. Phys. 50, 135109
Appl. Phys. Lett. 109, 062102 (2016). (2017).
695
J. Kim, S. Oh, M. Mastro, and J. Kim, Phys. Chem. Chem. Phys. 18, 738
M. Fleischer, J. Giber, and H. Meixner, Appl. Phys. A 54, 560 (1992).
15760 (2016). 739
F. Becker, C. Krummel, A. Freiling, M. Fleischer, and C. Kohl, Fresenius
696
J. D. Caldwell, T. J. Anderson, J. C. Culbertson, G. G. Jernigan, K. D. J. Anal. Chem. 358, 187 (1997).
Hobart, F. J. Kub, M. J. Tadjer, J. L. Tedesco, J. K. Hite, M. A. Mastro, 740
T. Schwebel, M. Fleischer, H. Meixner, and C.-D. Kohl, Sens. Actuators
R. L. Myers-Ward, C. R. Eddy, P. M. Campbell, and D. K. Gaskill, ACS B 49, 46 (1998).
Nano 4, 1108 (2010). 741
M. H. Lin, F. D. Wang, and N. X. Shan, Chin. Phys. B 19, 076102 (2010).
697
H. Peelaers and C. G. Van de Walle, Phys. Rev. B 96, 081409 (2017). 742
C. Baban, Y. Toyoda, and M. Ogita, J. Optoelectron. Adv. Mater. 7, 891
698
M. Ogita, K. Higo, Y. Nakanishi, and Y. Hatanaka, Appl. Surf. Sci. 175/ (2005); available at https://joam.inoe.ro/index.php.
176, 721 (2001). 743
T. Schwebel, M. Fleischer, and H. Meixner, Sens. Actuators B 65, 176
699
M. Ogita, N. Saika, Y. Nakanishi, and Y. Hatanaka, Appl. Surf. Sci. 142, (2000).
188 (1999). 744
M. Handwerg, R. Mitdank, Z. Galazka, and S. F. Fischer, Semicond. Sci.
700
C. Baban, Y. Toyoda, and M. Ogita, Thin Solid Films 484, 369 (2005). Technol. 31, 125006 (2016).
701
C. Baban, Y. Toyoda, and M. Ogita, Jpn. J. Appl. Phys. 43, 7213 (2004). 745
I. E. Demin and A. G. Kozlov, J. Phys.: Conf. Ser. 858, 012009 (2017).
702
M. Fleisher and H. Meixner, J. Mater. Sci. Lett. 11, 1728 (1992). 746
D. Wang, Y. Lou, R. Wang, P. Wang, and N. Jiang, Ceram. Int. 41,
703
M. Fleischer and H. Meixner, Sens. Actuators B 6, 257 (1992).
704 14790 (2015).
J. Frank, M. Fleischer, and H. Meixner, Sens. Actuators B 34, 373 747
B. S. Kang, H. T. Wang, F. Ren, and S. J. Pearton, J. Appl. Phys. 104,
(1996).
705 031101 (2008).
M. Fleischer, L. H€ollbauer, and H. Meixner, Sens. Actuators B 18, 119 748
E. M. Dede, F. Zhou, and S. N. Joshi, “Concepts for embedded cooling of
(1994).
706 vertical current wide band-gap semiconductor devices,” in 2017 16th
F. Reti, M. Fleischer, H. Meixner, and J. Giber, Sens. Actuators B 19,
IEEE Intersociety Conference on Thermal and Thermomechanical
573 (1994).
707 Phenomena in Electronic Systems (ITherm) (2017), pp. 508–515.
M. Fleischer and H. Meixner, Sens. Actuators B 52, 179 (1998). 749
708 D. H. Altman, A. Gupta, and M. Tyhach, “Development of a diamond
M. Fleischer and H. Meixner, Sens. Actuators B 4, 437 (1991).
709
M. Fleischer and H. Meixner, Sens. Actuators B 13, 259 (1993). microfluidics-based intra-chip cooling technology for GaN,” in 2015
710
U. Lampe, M. Fleischer, and H. Meixner, Sens. Actuators B 17, 187 Proceedings of the InterPACK Conference (2015, ), Paper No.
(1994). IPACK2015-48179, p. V003T04A006.
750
711
Z. Liu, T. Yamazaki, Y. Shen, T. Kikuta, N. Nakatani, and L. Yongxiang, G. Campbell, H. Eppich, K. Lang, C. Creamer, T. Yurovchak, K. Chu, A.
Sens. Actuators B 129, 666 (2008). Kassinos, M. Ohadi, A. Shooshtari, and S. Dessiatoun, “Advanced cool-
712
Y. Li, A. Trinchi, W. Wlodarski, K. Galatsis, and K. Kalantar-zadeh, ing designs for GaN-on-diamond MMICs,” 2015 Proceedings of the
Sens. Actuators B 93, 431 (2003). InterPACK Conference (2015), Paper No. IPACK2015–48429, p.
713
J. Frank, M. Fleischer, H. Meixner, and A. Feltz, Sens. Actuators B 49, V003T04A007.
751
110 (1998). A. Bar-Cohen, J. J. Maurer, and A. Sivananthan, MRS Adv. 1, 181 (2016).
752
714
M. Bartic, Y. Toyoda, C.-I. Baban, and M. Ogita, Jpn. J. Appl. Phys. 45, K. Bloschock and A. Bar-Cohen, “Advanced thermal management tech-
5186 (2006). nologies for defense electronics,” Proc. SPIE. 8405, 840501 (2012).
753
715
S. Nakagomi, T. Sai, and Y. Kokubun, Sens. Actuators B 187, 413 D. C. Dumka, T. M. Chou, F. Faili, D. Francis, and F. Ejeckam, Electron.
(2013). Lett. 49(20), 1298–1299 (2013).
754
716
T. Rahman, T. Masui, and T. Ichiki, Jpn. J. Appl. Phys. 54, 04DL08 M. Kim, J.-H. Seo, U. Singisetti, and Z. Ma, J. Mater. Chem. C 5, 8338
(2015). (2017).
717 755
D. Gonçalves, D. M. F. Prazeres, V. Chu, and J. P. Conde, Biosens. S. Rafique, L. Han, and H. Zhao, ECS Trans. 80, 203 (2017).
756
Bioelectron. 24, 545 (2008). H. Sun, C. G. Torres Castanedo, K. Liu, K.-H. Li, W. Guo, R. Lin, X.
718 Liu, J. Li, and X. Li, Appl. Phys. Lett. 111, 162105 (2017).
K.-H. Lee, J. O. Lee, S. Choi, J. B. Yoon, and G. H. Cho, Biosens.
757
Bioelectron. 31, 343 (2012). E. Chikoidze, A. Fellous, A. Perez-Tomas, G. Sauthier, T. Tchelidze, C.
719 Ton-That, T. T. Huynh, M. Phillips, S. Russelle, M. Jennings, B. Berini,
Y. Maruyama, S. Terao, and K. Sawada, Biosens. Bioelectron. 24,
3108–3112 (2009). F. Jomard, and Y. Dumont, “P-type b-gallium oxide: A new perspective
720
T. G. Drummond, M. G. Hill, and J. K. Barton, Nat. Biotechnol. 21, 1192 for power and optoelectronic devices,” Mater. Today Phys. 3, 118 (2017).
758
(2003). P. Weiser, M. Stavola, W. Beall Fowler, Y. Qin, and S. J. Pearton,
721
R. Lao, S. Song, H. Wu, L. Wang, Z. Zhang, L. He, and C. Fan, Anal. “Structure and properties of OH centers in Ga2O3,” Appl. Phys. Lett.
Chem. 77, 6475 (2005). (submitted).

A Review of Ga2O3 Materials, Processing, and Devices

Uploaded by

Copyright:

Available Formats

A Review of Ga2O3 Materials, Processing, and Devices

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Review of Ga2O3 Materials, Processing, and Devices

Uploaded by

Copyright:

Available Formats

A review of Ga2O3 materials, processing, and devices

Citation: Applied Physics Reviews 5, 011301 (2018);

APPLIED PHYSICS REVIEWS

A review of Ga2O3 materials, processing, and devices

TABLE OF CONTENTS PROPERTIES AND PROCESSING OF b-Ga2O3 . . . 13

1931-9401/2018/5(1)/011301/56/$30.00 5, 011301-1 Published by AIP Publishing.

TABLE I. Summary of the properties of Ga2O3 polymorphs.

Lattice Optical Volume

phase if quenched to room temperature.5,6 The b phase is the

Materials parameters Si GaAs 4H-SiC GaN Diamond b-Ga2O3 Comments

transistors (MESFETs) and metal-oxide-semiconductor field

FIG. 8. Cross-sectional TEM analysis for 391-nm-thick Ga2O3 sample with

bending.176 In concert with this, there is a recent work that

lack of control of doping, especially for lightly doped layers

PULSED LASER DEPOSITION (PLD)

HALIDE VAPOR PHASE EPITAXY (HVPE)

MIST-CVD MBE OF Ga2O3

growing very high quality Ga2O3.211–216 It is a scalable, uni-

Altuntas et al.226 deposited Ga2O3 dielectric thin films

modes compared to an undoped bulk Ga2O3 single crystal

PROPERTIES AND PROCESSING OF b-Ga2O3

along the [0 1 0] direction and the lattice spacing perpendic-

FIG. 17. Carrier concentration dependence of bandgap for b-Ga2O3 grown

point and a valence band maximum on the E line. The aniso-

vacancies and interstitials.391 Interstitial Ga and O vacancies

In2O3 and ZnO and the anisotropy in mobility (with mobili-

BAND STRUCTURE orientations are needed to deduce accurate band-gap values

of incidence light parallel to the c-axis (E//c), a*-axis (E//

103 cm1, but increased rapidly with increasing transition

TABLE III. Summary of Ohmic contact properties on b-Ga2O3.

TABLE IV. Table of reported Schottky barrier contacts to b-Ga2O3.

Ni 1.25 1.01 Evaporation, I-V Thermionic emission Oishi et al.499

are generation–recombination in the depletion region, WET ETCHING OF Ga2O3

Ga2 O3 þ 6HF ! 2GaF3 þ 3H2 O:

The etch rate in acids and bases strongly depends on the

TABLE V. Summary of wet etch results for b-Ga2O3.

aluminum in Cl2 gas (when there is no native oxide to inhibit

TABLE VI. Summary of dry etch results for b-Ga2O3.

Plasma Maximum etch

attached to the dangling surface bonds or lattice atoms. A

were obtained using chlorine-based plasmas.539 The study

SiO2 (single crystal) PECVD 3.1(60.2) 1.0(60.2) I Konishi et al.567

THIN FILM SOLAR-BLIND UV DETECTORS

drift layers grown by the established epi methods.638–640

with novel ALD TiN (65 nm thick) contacts on (201)

with a smaller diameter and 250 V for the largest diame-

POWER MOSFETs AND MESFETs of enhancement (e-mode) operation.676–691 The channels

extrinsic cutoff frequency (fT) and maximum oscillation fre-

FIG. 46. (a) DC output (IDS-VDS) characteristics, (b) DC input (IDS-VGS) as

Moser et al.680 reported MOSFETs on Ge-doped

nano-belts transferred to a SiO2/Si substrate. The d-mode

TABLE VIII. Summary of Ga2O3 gas sensors and performance.

Detection Operating Detection

FIG. 52. Schematics of concepts for embedded cooling of vertical current

You might also like