# Fileset

[Materials issues and devices of α- and β-Ga2O3.pdf](https://mdr.nims.go.jp/filesets/d38436d3-3362-46de-869b-be30fd9b7638/download)

## Creator

Elaheh Ahmadi, [Yuichi Oshima](https://orcid.org/0000-0001-8293-4891)

## Rights

This article may be downloaded for personal use only. Any other use requires prior permission of the author and AIP Publishing. This article appeared in Elaheh Ahmadi et al., J. Appl. Phys. 126, 160901 (2019) and may be found at https://doi.org/10.1063/1.5123213.[In Copyright](http://rightsstatements.org/vocab/InC/1.0/)

## Other metadata

[Materials issues and devices of α- and β-Ga2O3](https://mdr.nims.go.jp/datasets/7e919553-33e9-4090-8ac3-f5d4cf89aa76)

## Fulltext

Materials issues and devices of α- and β-Ga2O3ViewOnlineExportCitationCrossMarkRESEARCH ARTICLE |  OCTOBER 30 2019Materials issues and devices of α- and β-Ga2O3 Elaheh Ahmadi  ; Yuichi Oshima J. Appl. Phys. 126, 160901 (2019)https://doi.org/10.1063/1.5123213 29 January 2024 02:24:51https://pubs.aip.org/aip/jap/article/126/16/160901/564784/Materials-issues-and-devices-of-and-Ga2O3https://pubs.aip.org/aip/jap/article/126/16/160901/564784/Materials-issues-and-devices-of-and-Ga2O3?pdfCoverIconEvent=citehttps://pubs.aip.org/aip/jap/article/126/16/160901/564784/Materials-issues-and-devices-of-and-Ga2O3?pdfCoverIconEvent=crossmarkjavascript:;https://orcid.org/0000-0002-8330-9366javascript:;https://orcid.org/0000-0001-8293-4891javascript:;https://doi.org/10.1063/1.5123213https://servedbyadbutler.com/redirect.spark?MID=176720&plid=2100974&setID=592934&channelID=0&CID=768787&banID=521069223&PID=0&textadID=0&tc=1&scheduleID=2025884&adSize=1640x440&data_keys=%7B%22%22%3A%22%22%7D&matches=%5B%22inurl%3A%5C%2Fjap%22%5D&mt=1706495091448477&spr=1&referrer=http%3A%2F%2Fpubs.aip.org%2Faip%2Fjap%2Farticle-pdf%2Fdoi%2F10.1063%2F1.5123213%2F15235269%2F160901_1_online.pdf&hc=4a64607a94f4ab33d3185b91909054153e93405f&location=Materials issues and devices of α- and β-Ga2O3Cite as: J. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213View Online Export Citation CrossMarkSubmitted: 7 August 2019 · Accepted: 3 October 2019 ·Published Online: 30 October 2019Elaheh Ahmadi1,a) and Yuichi Oshima2AFFILIATIONS1Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, Michigan 48109, USA2Optical Single Crystals Group, National Institute for Materials Science, Tsukuba, Ibaraki 3050044, Japana)Email: eahmadi@umich.eduABSTRACTGa2O3 is an ultrawide bandgap semiconductor with a bandgap energy of 4.5–5.3 eV (depending on its crystal structure), which ismuch greater than those of conventional wide bandgap semiconductors such as SiC and GaN (3.3 eV and 3.4 eV, respectively).Therefore, Ga2O3 is promising for future power device applications, and further high-performance is expected compared to those ofSiC or GaN power devices, which are currently in the development stage for commercial use. Ga2O3 crystallizes into variousstructures. Among them, promising results have already been reported for the most stable β-Ga2O3, and for α-Ga2O3, which has thelargest bandgap energy of 5.3 eV. In this article, we overview state-of-the-art technologies of β-Ga2O3 and α-Ga2O3 for future powerdevice applications. We will give a perspective on the advantages and disadvantages of these two phases in the context of comparing thetwo most promising polymorphs, concerning material properties, bulk crystal growth, epitaxial growth, device fabrication, and resultingdevice performance.Published under license by AIP Publishing. https://doi.org/10.1063/1.5123213I. INTRODUCTIONSemiconductor power devices are now widely used for virtu-ally every single electric power conversion device and, therefore,improving the efficiency of the power devices is of crucial impor-tance for global energy savings. About 13% of generated electricitytoday is wasted through switching and conversion. Today’s com-mercial high-power semiconductor devices are dominantlySi-based. However, in the past few decades, gallium nitride (GaN)and silicon carbide (SiC) have emerged as desirable candidates toreplace Si in high power applications for high-frequency operation,efficiency enhancement, compactness, and weight reduction. Thesematerials are characterized by a bandgap of approximately 3.3 eV,significantly wider than silicon’s bandgap of 1.12 eV, allowing theuse of high electric fields and thus high voltages with low associ-ated losses. In the case of SiC power devices, for example, it is esti-mated that 300 × 106 barrels of crude oil can be saved, and 83 × 106tons of CO2 gas emission can be suppressed per year by 2050.1The power device related market is expected to reach $22 billionby 2024,2 which would be a great benefit to the semiconductorindustry. Servicing the electrification of the automobile is alreadyone of the fastest growing segments of the electronics industry.Nonetheless, producing large scale, cost-effective, and high qualityGaN and (to a lesser extent) SiC substrates remains the majorchallenge in the development roadmap of power electronics basedon these wide bandgap materials.There is an urgent need for new semiconductor devices tobe operated in the multiple kilovolts range for power electronicsapplications in many advanced systems, including distributedgrid systems, electric vehicles, high-speed trains, and industrialautomation. Due to the limited availability of semiconductorswitches of ≥10 kV, transformers are currently used to stepdown high voltages (HVs), which will be switched and thenstepped back up to the desired HV. Recently, ultrawide bandgap(UWB) semiconductors including Ga2O3, diamond, high Alcontent AlGaN, and AlN have drawn a great deal of interest inthe research community for HV applications beyond those acces-sible with GaN and SiC. Figure 1 demonstrates the materialproperties of these semiconductors along with those for GaNand 4H-SiC.Considering only the material properties, diamond is the mostpromising semiconductor for future high-power applications due toits wide bandgap, high thermal conductivity, and high electron andhole mobility. However, it possesses a number of serious challengeswhich need to be addressed before diamond can be considered forany practical application. These challenges include unavailability oflarge-scale and dislocation free diamond substrates in addition toJournal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-1Published under license by AIP Publishing. 29 January 2024 02:24:51https://doi.org/10.1063/1.5123213https://doi.org/10.1063/1.5123213https://www.scitation.org/action/showCitFormats?type=show&doi=10.1063/1.5123213http://crossmark.crossref.org/dialog/?doi=10.1063/1.5123213&domain=pdf&date_stamp=2019-10-30http://orcid.org/0000-0002-8330-9366http://orcid.org/0000-0001-8293-4891mailto:eahmadi@umich.eduhttps://doi.org/10.1063/1.5123213https://aip.scitation.org/journal/japhigh ionization energies of n-type (570 meV for phosphorous) andp-type (380 meV for boron) dopants in diamond.High Al-content AlGaN channel MESFETs on AlN substrates ortemplates have been proposed for high power applications. However,this material system suffers from low electron mobilities due to alloyscattering and low thermal conductivity, in addition to challenges withproducing large-scale AlN substrates. AlN-on-sapphire templatesare commercially available. However, similar to GaN-on-sapphiretemplates, they suffer from high threading dislocation density whichmakes them unsuitable for AlN-based vertical devices.Ga2O3, the target material of this article, is an ultrawidebandgap semiconductor with a bandgap energy of 4.5–5.3 eV(depending on its crystal structure), which is much greater thanthose of conventional wide bandgap semiconductors such as SiC andGaN (3.3 eV and 3.4 eV, respectively). Therefore, Ga2O3 is promisingfor future power device applications, and further high-performance isexpected compared to those of SiC or GaN power devices, which arecurrently in the development stage for commercial use. Ga2O3 crys-tallizes into various structures. Among them, promising results havealready been reported for the most stable β-Ga2O3 and for α-Ga2O3,which has the largest bandgap energy of 5.3 eV (The material proper-ties of these two different phases are listed in Table I). In this article,we overview state-of-the-art technologies of β-Ga2O3 and α-Ga2O3for future power device applications. We will give a perspective onadvantages and disadvantages of these two phases in the context ofcomparing the two most promising polymorphs, concerning materialproperties, bulk crystal growth, epitaxial growth, device fabrication,and resulting device performance.II. BULK CRYSTAL GROWTHAvailability of high-quality native substrates is a critical factorto grow high-quality epilayers, which are essential to realize high-performance semiconductor devices. Most of the wide bandgapsemiconductor crystals, such as GaN, SiC, diamond, etc., cannot begrown from the melt and, therefore, they were only heteroepitax-ially grown on largely lattice-mismatched foreign substrates at theearly stage of development in contrast to the conventional semicon-ductors such as Si and GaAs. However, the realization of bulk sub-strates by a vapor phase growth technique has enabled the growthof high-quality homoepitaxial layers and paved the way to the com-mercialization of GaN and SiC devices.3–5 Needless to say, theimportance of substrates also applies for Ga2O3 to fully exploit thesuperior potential. In this section, we overview the current statusand technical issues of substrates used to grow Ga2O3 epilayers.A. β-Ga2O3 substratesFortunately, as oppose to other wide bandgap semiconductors,β-Ga2O3 bulk crystals can be grown from the melt and high-qualitysingle crystal substrates are commercially available. This feature,in addition to the large bandgap, makes β-Ga2O3 a very attractivecandidate for power device applications. However, it was not easy toFIG. 1. Diagram comparing material properties of widebandgap semiconductors. Note that the melting point ofGaN is obtained from molecular dynamics simulations138and SiC is actually the decomposition temperature.Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-2Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/japproduce twin-free β-Ga2O3 wafers with a diameter large enough forpractical use because of the strong cleavage nature, and suchβ-Ga2O3 wafers have been realized only recently.6 In this part, weoverview the features, present status, and technical issues of the rep-resentative melt growth techniques of bulk β-Ga2O3 crystals.1. EFG methodAt present, edge-defined film fed growth (EFG) is the mostsuccessful bulk growth technique of β-Ga2O3. Currently, (010)wafers up to 25 × 25 mm2, (�201) wafers up to 2 in., and (001)wafers up to 4 in., both Sn-doped n-type conductive and Fe-dopedsemi-insulating, are commercially available from Novel CrystalTechnologies, Inc.In the EFG method, a board-shaped β-Ga2O3 crystal is grownon top of an iridium dye with narrow slits, through which theGa2O3 melt is supplied by capillarity [Fig. 2(a)].7,8 The melt areacan be much smaller than that of other melt growth techniques,and, therefore, it is possible to minimize the dissociation and evap-oration of the melt. The crystal shape is controlled by the shapeof the dye. The growth atmosphere needs to be controlled appro-priately (N2/O2 = 98%/2% for example) to minimize the iridiumloss by oxidation. The growth rate is up to a few dozen cm/day.The pulling direction is usually [010] in order to suppress twinningand seed blistering. Therefore, the area of EFG-grown (010) wafersis limited to approximately 25 × 25 mm2 although numerous prom-ising results have been reported on the substrates. In contrast,β-Ga2O3 wafers with a principal crystal plane in the [010] zone arescalable. Indeed, 6-in. (001) wafers have been demonstrated byNovel Crystal Technologies, Inc. Note that the β-phase has a mono-clinic crystal structure and, therefore, mutually orthogonal planesand directions do not have the same indices except for (010)and [010].n-type conductivity has been controlled mainly by Sndoping.8 Ge doping is difficult because of the high vapor pressureof GeO2. Although Nd−Na decreases by O2 annealing after thegrowth to remove residual strain, Nd−Na can be recovered to bevirtually identical to the donor concentration by N2 annealing.8The mechanism has not been clarified yet, but N2 annealing couldintroduce oxygen vacancies, and it is likely that an impurity–vacancy complex consisting of a dopant element and oxygen vacan-cies acts as a shallow donor.In general, dislocations in the semiconductor material candeteriorate the device performance. Experimental results suggestedthat dislocations in β-Ga2O3 can be current leakage paths.9Dislocation density in EFG-grown β-Ga2O3 has been reported to be∼103 cm−2,8 which is much lower than that in halide vapor phaseepitaxy (HVPE)-grown GaN substrates. It should be examined ifthis dislocation density is low enough for power device applica-tions. Note that the allowable dislocation density strongly dependson the device design and the drive conditions. In addition to dislo-cations, rod-shaped voids also exist in EFG-grown β-Ga2O3 crystalswith a density of ∼102 cm−2.10 The diameter of the void is around100 nm, and the length is 15 μm or more. These voids are believedto be formed by the evaporation of metallic impurities in thecrystal. The influence of these voids on the device performance hasnot been clarified yet, but the inner surface of such a void could bea current leakage path.As described above, the EFG method is very promising.However, the production cost needs to be reduced substantially inorder to put EFG-grown β-Ga2O3 wafers into practical use. A keytechnology for the cost reduction should be the suppression of theiridium loss, although the realization needs a breakthrough.2. Czhochralski methodThe Czhochralski (Cz) method is one of the most representa-tive melt-growth techniques, which is widely used for the massproduction of semiconductor crystals such as Si, Ge, GaAs, etc.11Figure 2(b) shows the principle of the Cz method. The growth of2-in. β-Ga2O3 bulk crystals has been demonstrated,12,13 and 1-in.semi-insulating (010) β-Ga2O3 wafers are commercially availablefrom Kyma Technologies and SYNOPTICS.In the Cz method, the melt surface area is much greater thanthat of EFG, therefore evaporation and dissociation of the Ga2O3melt are more significant. The amount of metal Ga produced bydissociation increases with increasing crucible diameter, beingenough to damage the iridium crucible by forming eutectic orintermetallic GaIr phases.14 The dissociation is suppressed byincreasing the oxygen partial pressure of the growth atmosphere;however, oxidization of the iridium crucible will be furtherenhanced. Thus, a breakthrough is required to realize the mass pro-duction of large-diameter (4–6 in.) β-Ga2O3 wafers by the Czmethod for power device applications.TABLE I. A comparison between the material properties and technological maturityof β- and α-Ga2O3.β αCrystal structure Monoclinic(β-gallia) CorundumEg 4.8 eV 5.3 eVElectron mobility(est.)200 cm2/V s 200 cm2/V sEbr (est.) 6.5 MV/cm 9.5 MV/cmε 10 10 (est.)BFOM 1231 3844Thermal conductivity 0.27W/cm K [010] No report0.11W/cm K [100]Substrate β-Ga2O3 (melt-grown)Available ∼ f2 in.,R&D ∼ f6 in.Currently expensiveSapphireAvailable ∼ f6 in.cheapEpitaxial growthtechniqueMBE, HVPE, MOCVD,Mist-CVD, PLD, etc.Mist-CVD, HVPE,MOCVDDislocation density <103 cm−2 Typical: 1010 cm−2ELO: <5 × 106 cm−2Surface roughness ofepilayersMacro step, need CMP Smooth under SEM(AlxGa1-x)2O3 x < 18% No limitationn-type doping Si, Ge, Sn Si, Ge, Snp-type doping N (deep acceptor 1 eV) No reportHetero-pn-junction NiO (FCC crystalstructure)α-Ir2O3, α-Rh2O3SI Fe No reportJournal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-3Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/jap3. Floating zone methodThe first demonstration of β-Ga2O3 wafers with a practicaldiameter was carried out using the floating zone (FZ) method.6 Inthe FZ method, the β-Ga2O3 single crystal rod is grown from theGa2O3 melt supported between the polycrystalline feed rod and thegrown single crystal by surface tension [Fig. 2(c)]. The FZ methoddoes not use a crucible, and, therefore, high-purity crystals can begrown. n-Type conductivity is controlled by intentional Si or Sndoping.15,16 The growth can be carried out under high oxygenpartial pressure without the crucible oxidization problem. However,the large-diameter growth is difficult because of the small hot zoneand high temperature gradient owing to lamp heating.4. Vertical Bridgman methodThe first application of the vertical Bridgman (VB) method toβ-Ga2O3 was demonstrated by Hoshikawa et al., and they reportedthe growth of 1-in. (100) β-Ga2O3 bulk crystals with a dislocationdensity as low as 103 cm−2.17 In the VB method, the β-Ga2O3crystal is grown by one-way solidification of the Ga2O3 melt in aPt/Rh alloy crucible slowly passing through a temperature gradientin a vertical furnace [Fig. 2(d)]. The shape of the bulk crystalfollows that of the crucible. The VB method can be cost-effectivebecause the growth is carried out in air and the crucible material isreusable. In principle, the VB method should be advantageous togrow large-scale bulk crystals since the growth is carried out in aFIG. 2. Schematic of (a) EFG, (b) CZ, (c) FZ, and (d) VB melt growth techniques suitable for the growth of bulk β-Ga2O3.Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-4Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/japgentle temperature gradient. Thus, the VB method is promising forthe mass production of large-diameter β-Ga2O3 wafers at a reason-able cost although the VB technology remains in its infancy andfurther development is required.B. Substrates for α-Ga2O3α-Ga2O3 is meta-stable at ambient pressure, and the meltgrowth is not possible just like SiC, GaN, diamond, etc. Accordingly,α-Ga2O3 films need to be grown heteroepitaxially. Currently, onlysapphire has been reported as a substrate for α-Ga2O3. Fortunately,large-diameter sapphire wafers up to 6 in. are mass-produced mainlyfor nitride LEDs and commercially available at a reasonable price.However, the dislocation density of an α-Ga2O3 epilayer is typicallyas high as 1010 cm−2 because of the large lattice mismatch, if nomeasure is taken.18As described in Sec. III B 2, the rapid growth of α-Ga2O3 ispossible by HVPE at a growth rate over 100 μm/h.19 It would,therefore, be possible to produce freestanding α-Ga2O3 wafers in asimilar way of producing freestanding GaN wafers, i.e., growing athick α-Ga2O3 layer by HVPE and removing the base substrateafter the growth.α-Ga2O3 is more stable than β-Ga2O3 at high pressures, andβ-Ga2O3 turns into α-Ga2O3 under 20–22GPa or higher at ambienttemperature.20 Therefore, it is expected to be possible to growα-Ga2O3 crystals from the NaOH flux under 4.4 GPa at 1000 °C usingβ-Ga2O3 powder as the raw material.21 In future, such a high-pressureflux method could be utilized to produce bulk α-Ga2O3 single crystals,similar to that now being used for diamond.III. EPITAXIAL GROWTHIn order to realize high-performance Ga2O3 power devices, itis essential to establish epitaxial growth techniques which producehigh-quality Ga2O3 films with controlled electrical conductivity at areasonably high growth rate. Although most of the epitaxial growthtechniques currently being investigated for Ga2O3 are based onthose proven to be effective for other conventional semiconductors,they need to be tailor-made to fit the nature of Ga2O3. In thissection, we overview the features, achievements, and technicalissues of epitaxial growth techniques for Ga2O3.A. Epitaxial growth of β-Ga2O3In most cases, β-Ga2O3 films for power device applicationsare grown homoepitaxially since high-quality native substrates areavailable as described in Sec. II A, (010) and (001) substrates aremost frequently used. On (010), relatively smooth epilayers can begrown at a rapid growth rate although the wafer size is limited. (001)wafers are scalable and the epi growth rate on (001) is reasonable,however, the surface morphology tends to be rough. Therefore, it isnot easy to grow an abrupt interface and the CMP process is requiredprior to the device fabrication. (100) wafers should also be scalable,but the epi growth rate is extremely low and twin boundaries areeasily introduced although twinning can be suppressed to somedegree by using miscut substrates.22 (�201) epilayers also suffer fromstacking fault. In conclusion, (010) and (001) are the realistic choicesat present. In the following part, we overview the most frequentlyreported growth techniques for β-Ga2O3, i.e., molecular beam epitaxy(MBE), HVPE, and metalorganic vapor phase epitaxy (MOVPE).1. MBEPlasma-assisted MBE and Ozone-MBE have been reported.Higher growth rates can be achieved in ozone-MBE comparedwith those in plasma-assisted MBE. Nevertheless, in ozone-MBE,the growth apparatus needs to be designed so that the ozoneinjection nozzle is close to the substrate surface and ozone gas canreach the surface during its short lifetime. The relationshipbetween the growth rate and crystal orientation has been studiedin detail. The growth rate is the fastest on (010) and much sloweron cleavage planes such as (100) and (001).23 The growth rate issuppressed and becomes even negative under Ga-rich growth con-ditions because of the formation of volatile sub-oxide Ga2O.24,25The negative growth (etching) can be utilized as a cleaning tech-nique of the β-Ga2O3 substrate surface prior to the film growth.26The β-(AlxGa1-x)2O3 solid solution can also be grown, althoughthe Al composition is limited because the stable structure of Al2O3 isnot the β-gallia structure. The solubility limit increases with increas-ing growth temperature.27 However, the Al composition is less than20%28 because the MBE growth temperature is usually limited to lessthan 700 °C to suppress the decomposition of β-Ga2O3 in a vacuum.The residual carrier concentration of an Uunintentionally doped(UID) β-Ga2O3 epilayer is reported to be less than ∼7 × 1015 cm−3.29The primary residual donor is believed to be Si coming from anambient environment and the quartz parts used in the plasmasource. Intentional n-type conductivity control is possible usingmainly Sn- or Ge-doping.30 β-(AlxGa1-x)2O3/β-Ga2O3 modulation-doped FETs (MODFETs) have already been demonstrated,26,31,32 andthe 2DEG channel mobility has been reported to be as high as143 cm2/V s at RT.32The biggest drawback of MBE for commercial applications isthat the growth rate is very low and, therefore, not suitable for thegrowth of thick drift layers of vertical power devices. MBE could bestill useful for lateral devices.2. HVPECurrently, HVPE currently gives the highest growth rate ofβ-Ga2O3 as a vapor phase growth technique and it is possible togrow over a few dozen μm/h under atmospheric pressure.33,34In addition, the influence of carbon impurity can be minimizedupon n− doping control since carbon-free precursors are used.Thus, HVPE is promising for growing thick drift layers with highproductivity. At present, (001) β-Ga2O3 homoepitaxial wafers with ann− layer are commercially available from Novel Crystal Technologies.Figure 3(a) shows the principle of HVPE. GaClx is used asthe Ga precursor in the HVPE of β-Ga2O3. CaClx is produced by thechemical reaction between metal Ga and HCl gas upstream in thereactor. O2 is used as the oxygen precursor in many cases while H2Owas also reported.35 In any case, the equilibrium constant of thechemical reaction to produce Ga2O3 is much greater than that ofGaN-HVPE. Therefore, special attention should be paid to sup-press the parasitic gas-phase reaction when the precursors aresupplied at high partial pressures upon rapid growth. The homoe-pitaxy of β-Ga2O3 has been reported mainly on (001) andJournal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-5Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/jap(010).33,35–37 On both (001) and (010), macrosteps developstrongly when a thick film is grown.35,36In general, HVPE-grown crystals tend to have relatively highincorporation of Si impurity from the quartz reactor becauseSi-included gas is released when hydrogen-included gas is incontact with quartz at high temperatures. On the other hand, theresidual Si concentration needs to be low enough for precise con-ductivity control. In order to remove hydrogen-included gas fromthe system, it is effective to use Cl2 gas instead of HCl gas forGaClx synthesis and to use O2 gas rather than H2O as the oxygenprecursor. As a result, the residual carrier concentration can besuppressed to less than 1013 cm−3.36 Other than Si, Cl impurity isalso detected at a concentration of ∼1016 cm−3.35 Experimentally,Cl has been reported to be electrically inactive in β-Ga2O3,36 althoughCl is predicted to be a shallow donor.38 However, careful investigationshould be continued to check whether Cl could affect the electricalproperties under specific conditions. Intentional Si doping has beencarried out using SiCl4 as a dopant source and it is possible tocontrol the net carrier concentration in the 1016 cm-3 range.39In order to mass-produce β-Ga2O3 epi wafers for power deviceapplications by HVPE, it is essential to develop a large-scale multiple-wafer apparatus. A planetary reactor or a close coupled showerheadreactor should be employed to secure in-wafer/wafer-to-wafer unifor-mity because the growth rate decay along the gas flow direction isfast in HVPE of β-Ga2O3. Reducing the reactor pressure is also aneffective option.3. MOVPEMOVPE is the most widely used epitaxial growth techniquefor the mass production of III-V compound semiconductor epiwafers. MOVPE of β-Ga2O3 has been investigated using TMGa,TEGa, and Ga(DPM)3 as gallium precursors, and O2, and H2Oas oxygen precursors. Equilibrium constants of chemical reac-tions to produce Ga2O3 by MOVPE are much greater thanthose of GaN, and the parasitic gas-phase reaction easily occursespecially upon rapid growth. Accordingly, the MOVPE ofβ-Ga2O3 is carried out at a reduced reactor pressure in manycases. A close coupled reactor is also effective to minimize theparasitic reaction and high growth rates up to 10 μm/h havebeen demonstrated.40In general, MOVPE-grown epilayers tend to include carbonimpurities originated from the metalorganic compound precursors.Carbon is predicted to be a shallow donor in β-Ga2O3,41 and theconcentration should be reduced for precise n− control in driftlayers. Promisingly, the carbon concentration in a MOVPE-grownUID β-Ga2O3 film was reported to be less than the detectionlimit of SIMS (∼2 × 1016 cm−2).42 Intentional n-type conductivitycontrol by Si doping is possible and carrier concentrations between1015 ∼ 1020 cm−3 are reported.42–46 In a recent report, a high-puritySi-doped β-Ga2O3 film with a net carrier concentration of2.5 × 1016 cm−3 has been demonstrated, and the bulk electronmobility was as high as 184 cm2/V s at RT.42 Note that the growthrate of the β-Ga2O3 film was relatively low, less than 1 μm/h. Itshould be assessed whether such high purity and large mobility canbe maintained at higher growth rates.The MOVPE growth temperature of β-Ga2O3 can be higher thanthat of MBE because the reactor pressure is much higher. The highgrowth temperature raises the solubility limit of β-(AlxGa1-x)2O3, andthe growth of β-(AlxGa1− x)2O3 with x > 40% has been demonstratedat >800 °C.43 The β-(AlxGa1− x)2O3/β-Ga2O3 super lattice with abruptinterfaces can also be grown.43As described above, the combination of high growth rates(much higher than that possible by MBE) and the possibility ofgrowing abrupt heterointerfaces, which is difficult by HVPE, makesMOVPE very promising for mass production of β-(AlxGa1-x)2O3FIG. 3. Schematic of (a) HVPE and (b) mist-CVDsystems used for the epitaxial growth of α- and β-Ga2O3.Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-6Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/japepiwafers. Although the residual carrier concentration is higherthan that of the HVPE-grown film at present, the bulk electronmobility is excellent. Scalability of the growth apparatus should bebetter than that of MBE and HVPE. MOVPE has great potential toplay a leading role in the mass production of β-Ga2O3 epiwafers forpower device applications in future.B. Epitaxial growth of α-Ga2O3As explained in Sec. II B, α-Ga2O3 is needed to be grown het-eroepitaxially. Sapphire can be utilized as the substrate, which leadsto very high dislocation densities due to the large lattice mismatch,if no measure is taken. The epitaxial growth has been reportedmainly on c-plane substrates, but the growth on other crystalplanes such as a, m, and r is also possible. In this section, wereview the current status of mist-CVD, HVPE, MOVPE, and MBE,which are mainly reported as the growth techniques for α-Ga2O3.1. Mist-CVDIn the mist-CVD technique, the aqueous solution of precur-sors is ultrasonically atomized and the mist is transferred togetherwith a carrier gas onto the heated substrate surface to grow Ga2O3[Fig. 3(b)].47 The growth is usually conducted under atmosphericpressure. Gallium (III) acetyl acetonate or GaCl3 is used as the Gasource, and H2O is the primary oxygen source. HCl is sometimesadded to the aqueous solution to improve the precursor solubility.Isomorphic α-Ga2O3 is obtained by growing at sufficiently lowtemperatures in the range of 500–600 °C on a sapphire substrate.The use of the GaCl3 precursor enables a larger growth rate as highas 4.5 μm/h.48The α-(AlxInyGa1− x-y)2O3 solid solution can be grown byadding aluminum(III) acetyl acetonate and/or indium(III) acetylacetonate in the aqueous solution.49 α-(AlxGa1− x)2O3 can begrown with no Al composition limitation, and wide range bandgapcontrol (5.3–8.8 eV) has been demonstrated.49 Although the lowthermal stability of α-Ga2O3 (usually ∼500 °C) is a drawback, alloy-ing with α-Al2O3 dramatically improves the thermal stability.α-(AlxGa1− x)2O3 with only x < 1% was reported to be stable up to800 °C.50 α-(InyGa1− y)2O3 can be grown without phase separationwhen y = 0% –8% and y = 67% –70%.49Mosaicity of α-Ga2O3 grown on (0001) sapphire by mist-CVDis characterized by a very narrow tilt angle and a broad twist angle.FWHMs of x-ray rocking curves of 0006 and 10�12 diffractionsmeasured in symmetric and skew-symmetric geometries were20–60 arcsec and 1000–2000 arcsec, respectively,47,51 suggestingthat the dislocations having an edge component are dominant.The plan-view TEM observation revealed that the total dislocationdensity is ∼1010 cm−2, which is much higher than that in a GaNfilm grown on the sapphire substrate (typically ∼109 cm−2).The dislocation density can be reduced to ∼ 6 × 108 cm−2 bybending the dislocations using the interfacial strain of theα-(AlxGa1-x)2O3 super lattice.48 The epitaxal lateral overgrowth(ELO) is also effective, and the dislocation density in the laterallygrown wing region was reported to be less than 1 × 107 cm−2.Impurity concentrations in mist-CVD-grown α-Ga2O3 were veryhigh when the method was first applied to α-Ga2O3. For example,[H] = 2 × 1019 cm−3, [C] = 1 × 1019 cm−3, [Si] = 9 × 1018 cm−3, etc.51At present, the purity has been much improved [for example,[H] = 2 × 1017 cm−3, [C] < D. L. (7 × 1016 cm−3)]52 mainly byimproving the precursor purity. Si concentration and residualcarrier concentration have not been disclosed recently. In amist-CVD reactor, the quartz surface is exposed to high partialpressure of H2O which leads to Si contamination in the growthenvironment. Therefore, the mist-CVD-grown α-Ga2O3 films typi-cally have high unintentional Si concentration, which is needed tobe reduced to enable the precise control of carrier concentration.n-type doping control is conducted by adding a dopant sourcesuch as tin(II) chloride dehydrate in aqueous solution, and thecarrier concentrations in the range of 1 × 1017–3 × 1019 cm−3 havebeen reported.18,51,53 Carrier mobility in Sn-doped c- and m-planefilms has been reported to be 24 cm2/V s (n = 2 × 1018 cm−3) and60 cm2/V s (n = 1 × 1018 cm−3), respectively.Although it is difficult to achieve p-type conductivity inα-Ga2O3, corundum-structured p-type α-Ir2O3 and α-Rh2O3 areavailable.50 α-Ir2O3 was reported to exhibit clear p-type conductiv-ity with μ = 2.3 cm2/V s when p = 1 × 1021 cm−3. The small latticemismatch between α-Ir2O3 and α-Ga2O3 (∼ 0.3%) would beadvantageous to form a hetero pn-junction and to growth ofα-(IrxGa1-x)2O3 solid solutions for bandgap tuning. Normally-offMOSFETs using a novel p-type corundum semiconductor weredemonstrated in 2018 by FLOSFIA and Kyoto University.As described above, the mist-CVD technique has achievedpromising results and is cost-effective due to the simple growthapparatus. However, mist-CVD is a relatively new technique andhas not been applied to mass-production of semiconductor epiwa-fers. In order to utilize the mist-CVD for commercial production,it is necessary to realize the multiwafer large-scale growth appara-tus, and the realization requires a steady effort to accumulate basictechnical knowledge and know-hows.2. HVPEHVPE can be utilized to grow not only β-Ga2O3 but also high-purity α-Ga2O3 at high growth rates. The growth principle forα-Ga2O3 is similar to that of β-Ga2O3. Isomorphic α-Ga2O3 can begrown on a c-plane sapphire substrate when the growth tempera-ture is sufficiently low (typically 500–600 °C).19 The growth rateincreases with increasing growth temperature under fixed precursorsupply, which suggests that the growth occurs under the influenceof the chemical reaction rate. Nonetheless, the growth rate increaseswith increasing precursors supply, reaching over 100 μm/h withmaintaining a specular surface.The structural quality of HVPE-grown α-Ga2O3 is similar tothat of a mist-CVD-grown film, but the tilt angle tends to be larger(∼100 arcsec) probably because of the un-optimized nucleationprocess. The dislocation density measured by plan-view TEM istypically ∼1010 cm−2. ELO can also be done by HVPE to reducethe dislocation density down to less than 5 × 106 cm−2 in laterallygrown wing regions.54 It is also possible to perform facet-initiatedELO (FIELO), which can reduce the dislocation density not only inwing regions but also above mask windows by bending the disloca-tions by inclined facets.54 The rapid growth by HVPE is advanta-geous to achieve island coalescence in the ELO process especiallywhen a small fill factor mask (wide mask and small windows) isJournal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-7Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/japused and, therefore, the dislocation density can be reducedeffectively.The residual Si concentration in an HVPE-grown UIDα-Ga2O3 film can be less than the detection limit of SIMS(3 × 1015 cm−3, for example),55 in spite of the existence of H2 gas asthe by-product of the chemical reaction between Ga metal and HClgas to synthesize GaCl, probably because of the low growth temper-ature. Cl impurity has also been detected, and the concentration istypically ∼1 × 1016 cm−3.55 The residual carrier concentration inUID α-Ga2O3 films has not been reported yet, but the film is virtu-ally semi-insulating.n-type conductivity control is possible by doping donor impu-rities, such as Ge.54 Ge is expected to substitute Ga sites causingsmaller strain than other donor impurities such as Si and Sn withlarger ionic radius. The GeCl4 melt (the melting point is ∼−50 °C)is used as a dopant source and the doping is performed by intro-ducing the vapor into the HVPE reactor by means ofthe conventional bubbling technique. High Ge concentrations([Ge] = 2.5 × 1019 cm−3, for example) can be achieved without thebroadening of XRC-FWHM. The electron mobility of a Ge-dopedα-Ga2O3 film has been reported to be as high as 28 cm2/V s whenn = 3 × 1019 cm−3, which is much higher than that of Sn-dopedmist-CVD-grown films (4 cm2/V s when n = 2 × 1019 cm−3, forexample). The resistivity of the Ge-doped film was 8.6 mΩ cm,which is much smaller than that of commercially available conduc-tive SiC wafers.The significance of the HVPE technique which can effectivelyimprove the crystal quality by ELO combined with the rapidgrowth, is quite large for meta-stable α-Ga2O3, which needs to begrown on largely lattice-mismatched substrates. The realization offreestanding α-Ga2O3 wafers by HVPE can also be expected, as ithappened in GaN industry. Although a large-scale reactor needs tobe developed for commercial use, there is a great deal of flexibilityin the reactor design; cost-effective high-throughput systems can bebuilt since the growth temperature for α-Ga2O3 is relatively low.3. MOVPE and MBEPromising results have been reported for β-Ga2O3 usingMOVPE and MBE, however, the growth of α-Ga2O3 by thesegrowth techniques is still at the stage of investigating growth condi-tions to grow pure-phase α-Ga2O3.In the case of MOVPE, it was reported that the polymorph ofGa2O3 on sapphire can be changed from β-Ga2O3 to ε-Ga2O3, andthen to a mixture of ε-Ga2O3 and α-Ga2O3 by increasing the addi-tion of HCl gas. However, isomorphic α-Ga2O3 has not beenachieved yet.56In the case of MBE, thickness of isomorphic α-Ga2O3 on sap-phire is limited. The maximum thickness has been reported to beapproximately 0.3 nm on the c-plane, 14 nm on the a-plane, and200 nm on the m-plane.57,58It is not clear at present why the growth window of MOVPEand MBE is very narrow. The mechanism of the phase selection inthese growth techniques needs to be clarified experimentally andtheoretically, and the growth conditions should be improvedaccordingly.IV. DEVICE FABRICATION AND CHARACTERIZATIONAlthough promising results have been achieved on theepitaxial growth of α-Ga2O3 on sapphire via, mainly, mist-CVDand HVPE, there are only a few reports52,59,60 on α-Ga2O3-basedelectronic devices. FLOSFIA demonstrated the first vertical Schottkybarrier diodes (SBDs) on thin-film corundum-structured galliumoxide (α-Ga2O3) grown on sapphire substrates by the MISTEPITAXY technique.52 Taking advantage of the misfit dislocationsand strain accumulated between the α-Ga2O3 layer and the sap-phire substrate, the epistructure was then lifted off the sapphiresubstrate and mounted on a heat sink (Fig. 4). The ability to peeloff α-Ga2O3 from sapphire not only simplifies back-side Ohmiccontact deposition but is also of crucial importance to efficientlyremove the heat from the device. The latter is of particular interestin the Ga2O3 community as this material suffers from very lowthermal conductivity. The SBDs exhibited an on-resistance andbreakdown voltage of 0.1 mΩ cm2 and 531 V (SBD1) or 0.4mΩ cm2and 855 V (SBD2), respectively. FLOSFIA also reported the first nor-mally off Ga2O3 MOSFETs in 2018 fabricated on the α-Ga2O3 filmgrown on sapphire.61 They demonstrated a gate threshold voltageof 7.9 V using a novel p-type corundum semiconductor whichfunctions as an inversion layer. Dang et al.59 fabricated MESFETson α-Ga2O3 thin films grown on sapphire by mist-CVD. Theyused AgOx as a Schottky gate contact. The rectification ratio andreverse breakdown voltage of typical SDs were 6×106 and 19.6 V,FIG. 4. Schematic demonstrating fabrication steps of the Schottky diode on α-Ga2O3. Reproduced with permission from Oda et al. Appl. Phys. Express 9 021101 (2016),Copyright 2016 The Japan Society of Applied Physics.Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-8Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/japrespectively. The ON/OFF ratio of the corresponding transistorswas 2×107.On the other hand, electronic devices based on β-Ga2O3 haveshown marked progress, including MOSFETs,62–64 MODFETs,65–67FinFETs,68–71 and SBDs72–74 on β-Ga2O3 substrates facilitated bythe availability of substrates. In the following, we will discuss theprogress that has been made so far to develop some of the neces-sary processing steps for device fabrication, including ion implanta-tion, Ohmic contact, Schottky contact, and dry etching. We willthen overview various device structures and their characteristics.A. Building blocks for device fabrication1. Ohmic contactBorrowing from GaN technology, Ti/Au has been commonlyused as an Ohmic contact in Ga2O3-based electronic devices.75–79Yao et al.80 investigated different metals (Ti, In, Ag, Sn, W, Mo, Sc,Zn, and Zr) as electrical contacts with n-type single-crystalβ-Ga2O3 (�201) substrates and studied the impact of annealing tem-perature up to 800̊ C (in flowing Ar). Their studies confirmed that,among all the studied metals, Ti contacts with an Au capping layerwere ohmic with the lowest resistance after annealing at 400̊ C for1 min. However, the contacts degraded when annealed above 500̊ C.Recently, interfacial reactions and interdiffusion of Ti/Au ohmic con-tacts with a tin-doped single-crystal β-Ga2O3 (010) substrate havebeen investigated.81 These studies concluded that the ohmic proper-ties of the Ti/Au metal stack on Ga2O3 are attributed to the inter-diffusion of Ti and Au and the resulting thin Ti-TiOx layer, whichhelps band alignment.2. Schottky contactNi and Pt have been commonly used as the Schottky contactin Ga2O3-based devices,29,73,74,82–84 although other metals includ-ing Cu, W, and Ir have been studied as well.85,86 Studying variouswet chemical treatments on the electrical behavior (e.g., barrierheights and ideality factors) of Ni-β-Ga2O3 Schottky diodes con-cluded that the pretreatment with organic solvents followed byHCl, H2O2, and DI water results in the most favorable characteris-tics. Schottky diodes with five different Schottky metals (W, Cu, Ni,Ir, and Pt) were fabricated and little dependence of Schottkybarrier height on the metal work function was observed.85Ahmadi et al.29 studied the Schottky contact of Ni on β-(Al,Ga)2O3 upon varying the Al content from 1 to 13%. The Schottkybarrier height and ideality factor were extracted from the I-V mea-surements. They observed that the apparent Schottky barrier heighthas similar values for different compositions of β-(AlxGa1−x)2O3for the samples studied. This behavior was attributed to the lateralfluctuation in the alloy’s composition, with the barrier height deter-mined by the lowest Al composition material.3. Ion implantationDevelopment of Si ion implantation in β-Ga2O3 has enabledlow-resistance ohmic contacts.71,74,87–89 Sasaki et al.90 studiedSi-ion implantation doping with concentrations ranging from1 × 1019 cm−3 to 1 × 1020 cm−3 in β-Ga2O3. The impact of anneal-ing temperature on the activation of implanted Si atoms wasinvestigated and shown that for concentrations below5 × 1019 cm−3, a high electrical activation efficiency above 60% wasobtained after annealing at a relatively low temperature in the rangeof 900–1000°C. Contact resistance and resistivity as low as4.6 × 10−6 Ω cm2 and 1.4 mΩ cm, respectively, were achieved forthe sample with an implanted-Si concentration of 5 × 1019 cm−3.Implantation of Mg and N of β-Ga2O3 as deep acceptordopants has been also explored.91 A significant diffusion of Mgatoms was observed when annealed at temperatures above 900 °C,whereas N diffusion was negligible when annealed up to 1100 °C. Itwas also shown that Mg implantation induces more damage in theβ-Ga2O3 crystal structure in comparison with N implantation.Though holes were not observed with Mg and N implantation,carrier compensation in bulk n-type substrates implanted with Mgor N was observed as evidenced by current blocking in n-“p”-nstructures, where “p” represents the implanted region.Mg/N implantation was used to form a current blocking layer(CBL) in planar-gate current aperture vertical Ga2O3 MOSFETs.89,92In these device structures, Si-implantation was used under source/drain to reduce the contact resistance. Mg implantation was used inthe first-generation devices,92 which led to poor off-state device char-acteristics as well as a reduced peak extrinsic transconductance.Moreover, the device could not be fully pinched-off. These nonideal-ities were attributed to the diffusion of Mg at high temperatures. Inthe next generation,89 N-implantation was utilized to form the CBL.Well behaved devices with an on-current density of 0.42 kA/cm2, anRon,sp of 31.5m cm2, an ID on/off ratio larger than 108, and small IDdispersion were demonstrated.4. Dry etchingChlorine-based dry etching of β-Ga2O3 has been studiedby several groups.93–98 Hogan et al.99 compared Chlorine-basedReactive ion etching (RIE) and inductively coupled plasma (ICP)etching techniques. The impact of RF power and chamber pressureon the etch rate and surface roughness for three crystallographicplanes, i.e., (100); (010); and (�201) was investigated by RIE, andmoderate etch rates (<20 nm/min) were achieved. Higher etch rateswith smoother surface morphology were demonstrated using ICP,perhaps due to the much higher plasma densities and uniformitiespossible with plasma powers beyond those realized by RIE. They alsoutilized ICP to study the etch rate of β-Ga2O3 (010) with differentgas chemistries, including BCl3, BCl3/SF6, CF4/O2, and BCl3/O2,and a maximum etch rate of 43.0 nm/min was achieved by BCl3.Zhang et al.96 also studied the ICP (ICP-RIE) etching of β-Ga2O3(�201) in BCl3/Ar chemistry and demonstrated etch rates above150 nm/min using an RIE/ICP power combination of 60W/900W.They showed that adding Ar to BCl3 did not change the etch rate sig-nificantly till it reaches a BCl3/Ar flow rate of 25/15 sccm. Furtherincrease in the Ar flow rate reduced the etch rate, which was unex-pected. Ar+ ions can help to remove the etch products and createactive sites for chemical etching, which should lead to higher etchrates. However, since BCl3 produces heavy BCl2+ and BCl3+ ions thatparticipate in physical etching, meaning BCl3 by itself provides bothchemical and physical etch components in the plasma and, therefore,the addition of Ar may not result in an improvement of the etch rateand merely dilutes the plasma chemistry.Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-9Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/japThe impact of BCl3/Ar ICP etching of β-Ga2O3 (�201) on thequality of Ni Schottky contacts was also studied.93,94,97,98 It wasshown that electrically active damage introduced during etchingnegatively affected the Schottky barrier height and diode idealityfactor. The low power etching conditions (150W of 2MHz ICPsource power and 15W rf of 13.56MHz chuck power) whichresulted in an etch rate of 12 nm/min, had minimal impact on thereverse breakdown voltage (6% reduction) and the diode idealityfactor (increased from 1.00 to 1.06). The barrier height reduced from1.2 eV to 1.01 eV. In contrast, high power etch conditions (400WICP and 200W rf) resulting in an etch rate of 70 nm/min, signifi-cantly damaged the surface and reduced the reverse breakdownvoltage by 35%. In addition, the high power etch reduced the barrierheight to 0.86 eV and increased the ideality to 1.2. The same grouplater investigated the effect of annealing before and after gate metaldeposition on reducing the damage induced by dry etching.97 Theyshowed that annealing at 450 C for 1 min in Ar ambient before gatemetal deposition was very effective to reduce the surface damage andrecover the Schottky diode characteristics.Very recently, Chlorine-based dry etching of α-Ga2O3 wasstudied.100 The effects of the BCl3/Cl2/Ar gas ratio, bias, and plasmapowers and chamber pressure on the etch rate, surface roughness,and mask selectivity were investigated. In contrast to previous dryetching studies on GaN, and similar to those on β-Ga2O3, BCl3 wasfound to be more effective than Cl2 to etch α-Ga2O3. A high etchrate of 65 nm/min, and nearly vertical sidewalls and smooth etchedsurfaces were achieved. It was shown that the Si3N4 hard mask hasthe highest resistance to BCl3-based etching compared to SiO2 andthe photoresist.B. Schottky diodes (SDs)SDs are attractive because of their fast switching speed due tothe absence of minority carriers. β-Ga2O3 SDs with a breakdownvoltage above 1 kV101–108 have been demonstrated by severalgroups. Different techniques such as field plates,108,109 guard rings(N-implanted rings),101 and trench Schottky barriers102,104 havebeen utilized to improve the breakdown voltage.Although these are very promising results, when designing forpractical applications, the voltage variations (e. g. brownout andstartup conditions) require the ability for the diodes to operateunder high current surge in the circuit. This means if the inputvoltage is dropped, the current must increase to maintain a cons-tant output power. A conductivity-modulation concept is typicallyused when designing a SiC Schottky diode to accommodate highercurrent when needed during the surge. As demonstrated in Fig. 5,p-type regions are implanted in the n-type drift layer. Therefore,the diode has two turn on voltages: (i) Schottky junction turn-onvoltage (Von,SJ) and (ii) p-n junction turn-on voltage (Von,PN). Thediode is biased to operate at low voltages slightly above the Von,SJ.However, in rare occasions that surge occurs in the circuit, there-fore, the voltage on the Schottky diode is increased to the secondturn-on voltage (Von,PN) to accommodate the surge current andmaintain a low voltage across the diode because of conductivitymodulation.In β-Ga2O3 SDs, however, it is difficult to adopt a conductivity-modulation concept due to the unavailability of conductive p-typeβ-Ga2O3 which can provide holes in the device. Therefore, devel-oping technologies in β-Ga2O3 that can accommodate surge isnecessary.C. Field effect transistors (FETs)The first β-Ga2O3 FET was reported by Higashiwaki et al. andwas fabricated on the Sn-doped β-Ga2O3 (010) film grown homoe-pitaxially on a Mg-doped β-Ga2O3 substrate by MBE.110 At thistime the device fabrication techniques were immature and thereforea circular FET pattern was used to produce a device without theneed for device isolation. Though it possessed poor ohmic contactsand high gate leakage current, a promising breakdown voltage ofmore than 250 V and an on/off drain current ratio of ∼104 wereobtained. To improve the ohmic contacts, Si-implantation in thesource/drain region area was employed followed by an activationannealing at 925 °C in an N2 environment.62,111 An atomic layerdeposition (ALD) Al2O3 dielectric was employed to reduce the highgate current leakage previously observed on their devices. The sup-pressed gate leakage resulted in a high drain current on/off ratio ofover ten orders of magnitude because of an extremely low off-statedrain leakage of a few pA/mm, and the breakdown voltage increasedto 370 V in the off-state. Moreover, stable transistor operation wassustained at temperatures up to 250 °C, although the drain leakagecurrent increased six orders of magnitude at this high temperature.FIG. 5. Schematic of the SD using conductivity-modulation to accommodate forcurrent surge in the circuit.Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-10Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/japThis group changed their device fabrication approach goingforward, using ion implantation for the active region and theohmic contacts. First a 1.2 μm UID Ga2O3 epilayer was grown onan Fe-doped semi-insulating β-Ga2O3 (010) substrate by ozoneMBE.63 Selective area Si-implantation was used to define 300 nmdeep channels with a uniform concentration of 3 × 1017 cm−3.Rectangular devices with 2 μm gate length were then fabricated.Room temperature drift mobilities of 90–100 cm2 V−1 s−1 weremeasured in the channel on these devices.63 To increase thebreakdown voltage, the gate-connected field plates were later usedand a breakdown voltage up to 750 V in Ga2O3 devices was demon-strated.112 Pulsed measurements showed a power added efficiency of12%, drain efficiency of 22.4%, output power density of 0.13W/mm,and maximum gain up to 4.8 dB at 1 GHz.113 These devices alsoshowed excellent Gamma-ray tolerance after exposure to the highestdose (230 kGy).114 Radiation-induced degradations in the gate insu-lation and surface passivation limited the overall radiation resistanceof these devices.The development of chlorine-based dry etching of Ga2O3 hasenabled device isolation through mesa etch.115,116 The group atAFRL have fabricated MOSFETs on Sn-doped β-Ga2O3 channelsgrown by MBE117 and MOCVD64 on Fe-doped β-Ga2O3 (010) andMg-doped β-Ga2O3 (100) substrates, respectively. An averagegate-to-drain electric field of 3.8 MV/cm, which is the highest onereported for any transistor and surpassing bulk GaN and SiC theo-retical limits, was obtained. Alternate dielectrics and channeldopants have also been explored. Moser et al.117 studied high-kALD HfO2 dielectrics for gates in MOSFETs with Sn-doped chan-nels and a pulsed current density of >450 mA/mm was observedupon applying a gate voltage of 4 V. The same group demonstratedthe first MOSFET with the Ge-doped channel grown by MBE on(010) Fe-doped semi-insulating substrates.64 The Ge-dopedchannel devices performed similarly to previously reported deviceswith Sn- and Si-doped channels with the drain current ON/OFFratios >108 and the saturated drain current >75 mA/mm atVG = 0 V. Hall effect measurements showed a high carrier mobilityof 111 cm2/(V s) with 4 × 1017 cm−3 active carriers. They were alsofirst to report the RF performance of β-Ga2O3 MOSFETs.118,119A gate recess design was used to scale the gate length and a highlydoped cap layer was employed under source and drain to reduce theohmic resistance. Current density, transconductance, fT, and fMAX of150mA/mm, 21.2mS/mm, 3.3 GHz and 12.9 GHz were measured onthese devices, respectively. A maximum output power of 0.23W/mmwas demonstrated with a maximum power added efficiency of 6.3%.An output current of 20 mAmm−1 and an on-resistance of520mΩ cm2 were achieved on these devices when a highly-Si-dopedGa2O3 cap layer was used under the source and drain.120Joishi et al.121 have investigated the impact of Fe-related trap statesin the UID buffer due to the diffusion of Fe from the Fe-dopedsubstrate, on MOSFET characteristics. They showed that increasingthe buffer layer thickness from 100 nm to 600 nm improves thecharge density/electron mobility from 1.4 × 1013 cm−2/65 cm2/V sto 1.7 × 1013 cm−2/105 cm2/V s. Additionally, the growth of athicker UID buffer layer improved the DC-RF dispersion.In power switching applications, high voltage and highcurrent devices are required. MOSFETs with a breakdown voltageof ∼2 kV have been demonstrated using a 400-nm thick compositefield plate oxide, with a combination of atomic layer deposited andplasma enhanced chemical vapor deposited SiO2 layers.122 E-modetransistors are typically more desirable over D-mode devices due tosafety considerations and simplicity of gate-drive circuitry. TheE-mode Ga2O3 MOSFETs have been realized by depleting thechannel via a wrap-gate fin-array field-effect transistor (finFET)structure,69 an unintentionally doped Ga2O3 film with low carrierconcentration as the channel,88 and a gate-recessed structure withALD SiO2 as the gate dielectric.123 Chabak et al.69 were the first todemonstrate E-mode MOSFETs in Sn-doped Ga2O3 wrap-gatefinFETs on a native semi-insulating Mg-doped (100) β-Ga2O3 sub-strate. These finFETs demonstrated normally off operation with athreshold voltage between 0 and +1 V during high-voltage opera-tion and an ION/IOFF ratio of greater than 105 which were mainlylimited by a reduced ION due to the high on-resistance. The samegroup recently reported E-mode β-Ga2O3 transistors grown homoe-pitaxially by MBE, utilizing a recessed-gate process, which depletedthe channel under the gate followed by the deposition of ALD SiO2as the gate dielectric.124 Wong et al.88 also demonstrated E-modeMOSFETs with low series resistance using Si-ion implantation ofthe source/drain contacts and access region. In these devices, thechannel was formed by an unintentionally doped Ga2O3 film withlow background carrier concentration. They achieved a positivethreshold voltage without additional constraints on the channeldimensions or device architecture. The devices suffered from noni-dealities associated with the Al2O3 gate dielectric, which causedlarge hysteresis.The primary drawback of Ga2O3-based MOSFETs is that thelow channel mobilities in the channel lead to very low current den-sities compared with their GaN counterparts. To address this issue,modulation-doped field effect transistors (MODFETs) have beeninvestigated.26,125–127 Ahmadi et al.26 and Zhang et al.65 have sepa-rately reported (AlxGa1-x)2O3-Ga2O3 modulation-doped FETsusing Ge and Si as n-type dopants in the barrier, respectively.Though Ge is an attractive dopant it has been shown that Ge incor-poration in β-Ga2O3 films reduces as the substrate temperatureincreases.30 Unfortunately, this is in conflict with the higher sub-strate temperatures required to grow (AlxGa1-x)2O3 films with alarger Al content.28 Therefore, the Ge-doped MODFET structuressuffer from a low Al content (AlxGa1-x)2O3 barrier. In contrast, Siincorporation does not depend on the substrate temperature, whichallows high growth temperatures for (AlxGa1-x)2O3 films.128It was expected that the introduction of β-(AlxGa1-x)2O3and formation of two-dimensional electron gas (2DEG) will leadto an enhancement of electron mobility in these structures.However, the highest room-temperature electron mobility reported,so far, in β-(AlxGa1-x)2O3-Ga2O3 heterostructures has been only180 cm2/V s.129–131 The electron mobility increased to 2790 cm2/V s at50 K. This high electron mobility allowed for the observation ofShubnikov-de-Haas oscillations from which an electron effectivemass of 0.33 me was extracted, establishing that phonon scatteringcoupled with the high electron effective mass limited the electronmobility. A maximum drain current of IDS = 257 mA/mm, a peakgm of 39 mS/mm and a pinch off voltage of −7 V have been mea-sured on MODFETs.132 Very recently, the same group increasedthe breakdown voltage of the MODFET from 475 V for a devicewith a gate-drain spacing (LGD) of 1.55 μm to a high breakdownJournal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-11Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/japvoltage of 1.37 kV for a gate-to-drain separation (LGD) of 16 μmhaving a specific ON-resistance of 120.1 mΩ cm2 using SiNx as thepassivation dielectric.133For high voltage and high-power applications, vertical topolo-gies are preferred to enhance the packing density of devices andsuppress the sensitivity of surface effects. Two types of verticalGa2O3-based devices have been reported to date, including currentaperture vertical transistors (CAVETs) and FinFETs.Wong et al.134 fabricated a current aperture vertical transistor(CAVET) with a Mg-implanted current blocking layer (CBL).They recently showed that the N-implanted CBL helps to improvethe device characteristics as N atoms are more stable in β-Ga2O3,whereas Mg diffuses during the postimplantation-annealing.89It must be noted that in this device structure, holes that are gener-ated during the reverse bias will be collected in the CBL. If the CBLis not grounded, the accumulation of holes in this region willlead to a large positive voltage in the CBL and result in a shift inthreshold voltage and huge instability in the device performance.Unfortunately, both Mg and N are very deep acceptors (1 eV) inβ-Ga2O3, and although the N/Mg-implanted layers can be used toblock the current, N/Mg-doping of β-Ga2O3 does not lead to a con-ductive p-type β-Ga2O3 film. Therefore, it is not possible to makethe ohmic contact with these current blocking layers. Movingforward, one needs to consider this serious issue while designingβ-Ga2O3-based CAVETs.Hu et al.,70 on the other hand, employed a deep-etch processto fabricate ∼1 μm thick vertical structures on a low-doped Ga2O3substrate. Very recently, normally-off Ga2O3 vertical FinFETs witha threshold voltage of 4 V, breakdown voltage of 1.6 kV, and adrain current density of 600 A/cm 2 were demonstrated.113 Thehigh threshold voltage was achieved by Fin-shaped channels withsubmicrometer widths. The main issue in these devices is the lowelectron mobility in the channel (10–30 cm2/V s) due to the surfaceroughness and charged surface states caused by dry etching.The low electron mobility in the channel leads to higher Ron.Additionally, the surface states due to surface damage caused bydry etching may lead to a hysteresis and a shift in the thresholdvoltage and device degradation. Therefore, to enhance the perfor-mance of β-Ga2O3 FinFETs and improve their reliability, varioustechniques including wet etching, regrowth, and annealing must beinvestigated to improve the surface quality of the sidewalls, and con-sequently, the electron mobility of the channel in these structures.As mentioned earlier, vertical devices are more sensitive todefects and the dislocation densities in the substrate. Therefore, inan immature material system or in a material system in which high-quality substrates are not available, typically increasing the devicearea (necessary to increase the current for high power applications)leads to reduction in the breakdown voltage and an increase in thereverse-biased leakage current. This is due to the increase in thenumber of defects or dislocations in the device (which are typicallyelectrically active and form leakage paths) when the device areaincreases. This phenomenon has also been reported in β-Ga2O3Schottky diodes. However, very recently, Novel Crystal Technologyannounced that in the epistructures consisting of 5 μm of lightly-doped (2–9 × 1016 cm−3) Ga2O3 films grown on n+-doped β-Ga2O3(001) substrates by HVPE, the film quality is such that the break-down voltage in the reverse direction shall not decrease even if theelectrode size of the SBD is enlarged on all over the entire 2-in. epiwafer. This rapid progress in bulk and the epitaxial growth ofβ-Ga2O3 indicate its great potential, particularly, for high-powervertical devices.V. CONCLUSION AND FUTURE PERSPECTIVETo summarize, both α- and β-Ga2O3 are promising semicon-ductors for future power electronics applications. For Ga2O3 toachieve its full potential, significant sustained funding is necessaryfor scientific discovery and technology development. This will becatalyzed by “a killer application” in which the advantages ofGa2O3 are irrefutable. Such an application is needed to be identifiedas was done in the case of GaN for next-generation RF transistors.As discussed earlier, high quality single crystal β-Ga2O3 canbe grown cost-effectively via conventional melt growth techniquesand is already commercially available. The availability of the bulksubstrate is the main advantage of β-Ga2O3 over α-Ga2O3 andother UWB semiconductors and makes the β- phase particularlyinteresting and provides more advantages for devices with verticaltopologies. This is because the performances of vertical devices,such as gate leakage and reliability, are more sensitive to the dislo-cation density. Additionally, since the electron mobility of Ga2O3 ismainly limited by high phonon scattering and high electroneffective mass, modulation-doped heterostructures do not lead toany significant improvement in the electron mobility. Therefore, incontrast to GaN power transistors that are dominantly based onAlGaN-GaN high electron mobility transistors (HEMTs), the fabri-cation of lateral β-Ga2O3-based MODFETs or MOSFETs may notprovide substantial advantages for high power application. Insteadsuch devices with lateral topologies will increase the device foot-print that will lead to an increase in cost. Therefore, design andfabrication of β-Ga2O3-based lateral power conversion devices dis-regards the main advantage of this semiconductor which distin-guishes it from all other wide bandgap semiconductors and that isthe availability of the cost-effective high-quality bulk substrate.Lateral devices still remain viable for RF applications.Nonetheless, as mentioned earlier, low thermal conductivity ofβ-Ga2O3 is one of the major obstacles of this material system thatmust be overcome before it can be considered for practical applica-tions. It has been shown135 that a β-Ga2O3 homoepitaxial devicesuffers from an unacceptable junction temperature rise of ∼1500 °Cunder a targeted power density of 10W/mm with a wafer thicknessof 500 μm. The effectiveness of various active and passive coolingsolutions was tested to achieve a goal of reducing the device operat-ing temperature below 200 °C at a power density of 10W/mm. Thisstudy showed that 100 μm of β-Ga2O3, even when integrated with adiamond substrate, is still too thick to achieve the target cooling.In contrast, when the Ga2O3 substrate was thinned down to 10 μm,integration with a diamond substrate reduced the junction topackage thermal resistance by a factor of six. Flip-chip heterointe-gration, as shown in Fig. 6, has also been suggested as an effectivetechnique for thermal management. The metal bumps shown inthis figure help in removing the heat from the device active regionthrough the contacts, however the dominant thermal resistance isstill the contribution from the Ga2O3 as the heat flows laterallythrough the semiconductor to the location of the device-side bondJournal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-12Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/jappads. Incorporating high thermal conductivity into nano-crystaldiamond passivation further helps with heat removal from the deviceactive region. Figure 7(a) shows the comparison of the device junc-tion temperature rise at 10W/mm upon varying the carrier wafermaterial (AlN vs diamond), carrier thickness (100 μm vs 250 μm),and electrical bump material (In vs Au). Figure 7(b) shows the effectof the thermal conductivity of the epoxy under-fill material on thedevice temperature rise. Therefore, although the low thermal conduc-tivity of Ga2O3 is the main drawback of this material system, thereare feasible pathways to address this issue.One of the advantages of α-Ga2O3 is its corundum crystalstructure which allows for an all-oxide p-n junction through theepitaxial growth of heterostructures with corundum-structuredp-type oxides such as Rh2O3 or Ir2O3. Moreover, there are severaltransition metal oxides (α-M2O3; M = Fe, Cr, V, Ti, Rh, and Ir)which form in the corundum crystal structure (Fig. 8) withferroelectric and ferromagnetic properties, enabling integrationwith multifunctional devices.136,137 Additionally, α-Ga2O3 has asimilar crystal structure as that of sapphire (Al2O3) and α-In2O3.Therefore, the epitaxial growth of α-(In,Ga,Al)2O3 alloys allowsbandgap engineering from 3.8 eV to 8.8 eV. The possibility of epitax-ial growth of α-(AlxGa1-x)2O3 films with a larger Al content onα-Ga2O3 enables larger conduction band discontinuity in these het-erostructures which will, consequently, allow for higher densities oftwo-dimensional electron gas in these heterostructures beyond whatis achievable in β-(Al,Ga)2O3-Ga2O3 heterostructures. Nonetheless,α-Ga2O3 bulk substrates are not currently available and α-Ga2O3films must be grown heteroepitaxially on sapphire substrates, and thethreading dislocation density is several orders of magnitude largerthan that in β-Ga2O3 substrates. As mentioned earlier, devices withvertical topologies are more sensitive to threading dislocation density.Therefore, at least until high quality free-standing α-Ga2O3 substratesFIG. 6. Schematic of Flip-chip hetero-integration of the β-Ga2O3 FET forthermal management. Reproduced withpermission from Chatterjee et al.,“Device-Level thermal management ofgallium oxide field-effect transistors,”IEEE Trans. Compon. Packag. Manuf.Technol. (published online), Copyright2019 IEEE.Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-13Published under license by AIP Publishing. 29 January 2024 02:24:51https://aip.scitation.org/journal/japwith low threading dislocations are available, horizontal structuresare more suitable for α-Ga2O3-based electronic devices. Additionally,since α-Ga2O3 can be lifted off sapphire relatively easily, the lowthermal conductivity of this material can be addressed simply bybonding it to other substrates (SiC, AlN, diamond, etc.) withhigher thermal conductivity. Therefore, α-(In,Ga,Al)2O3 hetero-structures are very promising for high power switching and RFapplications. Nevertheless, as mentioned earlier, there have beenonly a few reports on α-Ga2O3 devices, which is mainly due tothe unavailability of substrates. Therefore, to discover the fullpotential of the α-(Al,Ga,In)2O3 material system, it is crucial torealize high-quality α-Ga2O3 freestanding substrates or templatesubstrates. HVPE will play a key role in the development of suchα-Ga2O3 substrates, just like it did in GaN technology. The keytechnologies, such as rapid growth, conductivity control, defectcontrol, have already been demonstrated as described in Sec. III B 2.Therefore, demonstration of high-quality α-Ga2O3 substrates in thenear future is expected and will require further development of thekey technologies.ACKNOWLEDGMENTSProfessor Umesh Mishra is acknowledged for many fruitfuldiscussions on device perspectives.REFERENCES1Green Power Electronics, Council on Competitiveness-Nippon 2009 Report (inJapanese).2M. Rosina, Yole Development Report Presented in SEMICON Europa (2018).3Y. M. Tairov and V. F. Tsvetkov, “Investigation of growth processes of ingots ofsilicon carbide single crystals,” J. Cryst. Growth 43, 209–212 (1978).4Y. Oshima, T. Eri, M. Shibata, H. Sunakawa, K. Kobayashi, T. Ichihashi, andA. Usui, “Preparation of freestanding GaN wafers by hydride vapor phaseepitaxy with void-assisted separation,” Jpn. J. Appl. Phys. 42, L1–3 (2003).5K. Motoki, T. Okahisa, N. Matsumoto, M. Matsushima, H. Kimura, H. Kasai,K. Takemoto, K. Uematsu, T. Hirano, M. Nakayama, S. Nakahata, M. Ueno,D. Hara, Y. Kumagai, A. Koukitu, and H. Seki, “Preparation of large freestandingGaN substrates by hydride vapor phase epitaxy using GaAs as a starting sub-strate,” Jpn. J. Appl. Phys. 40, L140–3 (2001).6E. G. Víllora, K. Shimamura, Y. Yoshikawa, K. Aoki, and N. Ichinose,“Large-size β-Ga2O3 single crystals and wafers,” J. Cryst. Growth 270, 420–426(2004).7H. Aida, K. Nishiguchi, H. Takeda, N. Aota, K. Sunakawa, and Y. Yaguchi,“Growth of β-Ga2O3 single crystals by the edge-defined, film fed growthmethod,” Jpn. J. Appl. Phys. 47, 8506–8509 (2008).8A. Kuramata, K. Koshi, S. Watanabe, Y. Yamaoka, T. Masui, and S. Yamakoshi,“High-quality β-Ga2O3 single crystals grown by edge-defined film-fed growth,”Jpn. J. Appl. Phys. 55, 1202A2 (2016).9M. Kasu, K. Hanada, T. Moribayashi, A. Hashiguchi, T. Oshima, T. Oishi,K. Koshi, K. Sasaki, A. Kuramata, and O. Ueda, “Relationship between crystaldefects and leakage current in β-Ga2O3 Schottky barrier diodes,” Jpn. J. Appl.Phys. 55, 1202BB (2016).10K. Nakai, T. Nagai, K. Noami, and T. Futagi, “Characterization of defects inβ-Ga2O3 single crystals,” Jpn. J. Appl. Phys. 54, 051103 (2015).11Y. Tomm, P. Reiche, D. Klimm, and T. Fukuda, “Czochralski grown Ga2O3crystals,” J. Cryst. Growth 220, 510–514 (2000).12Z. Galazka, K. Irmscher, R. Uecker, R. Bertram, M. Pietsch, A. Kwasniewski,M. Naumann, T. Schulz, R. Schewski, D. Klimm, and M. Bickermann, “On thebulk β-Ga2O3 single crystals grown by the Czochralski method,” J. Cryst. Growth404, 184–191 (2014).FIG. 7. (a). Device junction temperature rise at 10 W/mm as a function ofthe carrier wafer material (AlN vs diamond), carrier thickness (100 μm vs250 μm), and electrical bump material (In vs Au). (b) The effect of thermalconductivity of the epoxy under-fill material on device temperature rise.Chatterjee et al., “Device-Level thermal management of gallium oxide field-effect transistors,” IEEE Trans. Compon. Packag. Manuf. Technol. ( publishedonline). Copyright 2019 IEEE.FIG. 8. Relationship between the bandgap and lattice constant along a axisof 9 types of corundum-structured oxides. Reproduced with permission fromKaneko et al., Jpn. J. Appl. Phys. 57 02CB18 (2018). Copyright 2018 TheJapan Society of Applied Physics.Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-14Published under license by AIP Publishing. 29 January 2024 02:24:51https://doi.org/10.1016/0022-0248(78)90169-0https://doi.org/10.1143/JJAP.42.L1https://doi.org/10.1143/JJAP.40.L140https://doi.org/10.1016/j.jcrysgro.2004.06.027https://doi.org/10.1143/JJAP.47.8506https://doi.org/10.7567/JJAP.55.1202A2https://doi.org/10.7567/JJAP.55.1202BBhttps://doi.org/10.7567/JJAP.55.1202BBhttps://doi.org/10.7567/JJAP.54.051103https://doi.org/10.1016/S0022-0248(00)00851-4https://doi.org/10.1016/j.jcrysgro.2014.07.021https://aip.scitation.org/journal/jap13J. Blevins, K. Chabak, G. Jessen, D. Thomson, K. Stevens, G. Foundos,J. Leach, J. Rumsey, and A. Green, “13.5 growth of 50 mm beta-gallium oxide(β-Ga2O3) substrates,” in CSMantech, 2018.14Z. Galazka, R. Uecker, D. Klimm, K. Irmscher, M. Naumann, M. Pietsch,A. Kwasniewski, R. Bertram, S. Ganschow, and M. Bickermann, “Scaling-Up ofbulk β-Ga2O3 single crystals by the Czochralski method,” ECS J. Solid State Sci.Technol. 6, Q3007–11 (2017).15E. G. Víllora, K. Shimamura, Y. Yoshikawa, T. Ujiie, and K. Aoki, “Electricalconductivity and carrier concentration control in β-Ga2O3 by Si doping,” Appl.Phys. Lett. 92, 202120 (2008).16N. Ueda, H. Hosono, R. Waseda, and H. Kawazoe, “Synthesis and control ofconductivity of ultraviolet transmitting β-Ga2O3 single crystals,” Appl. Phys.Lett. 70, 3561–3563 (1997).17K. Hoshikawa, E. Ohba, T. Kobayashi, J. Yanagisawa, C. Miyagawa, andY. Nakamura, “Growth of β-Ga2O3 single crystals using vertical Bridgmanmethod in ambient air,” J. Cryst. Growth 447, 36–41 (2016).18T. Kawaharamura, G. T. Dang, and M. Furuta, Jpn. J. Appl. Phys. 51, 040207(2012).19Y. Oshima, E.G. Víllora, and K. Shimamura, Appl. Phys. Express 8, 055501 (2015).20D. Machon, P. F. McMillan, B. Xu, and J. Dong, “High-pressure study of the β-to- α transition in Ga2O3,” Phys. Rev. B 73, 094125 (2006).21J. P. Remeika and M. Marezio, “Growth of α-Ga2O3 single crystals at 44kbars,” Appl. Phys. Lett. 8, 87–88 (1966).22R. Schewski, M. Baldini, K. Irmscher, A. Fiedler, T. Markurt,B. Neuschulz, T. Remmele, T. Schulz, G. Wagner, Z. Galazka, and M. Albrecht,“Evolution of planar defects during homoepitaxial growth of β-Ga2O3 layerson (100) substrates—A quantitative model,” J. Appl. Phys. 120, 225308(2016).23K. Sasaki, A. Kuramata, T. Masui, E. G. Víllora, K. Shimamura, andS. Yamakoshi, “Device-quality b-Ga2O3 epitaxial films fabricated by ozonemolecular beam epitaxy,” Appl. Phys. Express 5, 035502 (2012).24P. Vogt and O. Bierwagen, “Reaction kinetics and growth window forplasma-assisted molecular beam epitaxy of Ga2O3: Incorporation of Ga vs. Ga2Odesorption,” Appl. Phys. Lett. 108, 072101 (2016).25Y. Oshima, E. Ahmadi, S. Kaun, and F. Wu, “Growth and etching characteris-tics of (001) β-Ga2O3 by plasma-assisted molecular beam epitaxy,” Mater. Chem.Phys. 133, 1–6 (2012).26E. Ahmadi, O. S. Koksaldi, X. Zheng, T. Mates, Y. Oshima, U. K. Mishra, andJ. S. Speck, “Demonstration of β-(AlxGa1−x)2O3/β-Ga2O3 modulation dopedfield-effect transistors with Ge as dopant grown via plasma-assisted molecularbeam epitaxy,” Appl. Phys. Express 10, 071101 (2017).27V. G. Hill, R. Roy, and E. F. Osborn, “The system alumina-gallia-water,”J. Am. Ceram. Soc. 35, 135–142 (1952).28S. W. Kaun, F. Wu, and J. S. Speck, “β-(AlxGa1−x)2O3/Ga2O3 (010) hetero-structures grown on β-Ga2O3 (010) substrates by plasma-assisted molecularbeam epitaxy,” J. Vac. Sci. Technol. A 33, 041508 (2015).29E. Ahmadi, Y. Oshima, F. Wu, and J. S. Speck, “Schottky barrier height of Nito β-(AlxGa1−x)2O3 with different compositions grown by plasma-assisted molec-ular beam epitaxy,” Semicond. Sci. Technol. 32, 035004 (2017).30E. Ahmadi, O. S. Koksaldi, S. W. Kaun, Y. Oshima, D. B. Short, U. K. Mishra,and J. S. Speck, “Ge doping of β-Ga2O3 films grown by plasma-assisted molecu-lar beam epitaxy,” Appl. Phys. Express 10, 041102 (2017).31T. Oshima, Y. Kato, N. Kawano, A. Kuramata, S. Yamakoshi, S. Fujita,T. Oishi, and M. Kasu, “Carrier confinement observed at modulation-dopedβ-(AlxGa1-x)2O3/Ga2O3 heterojunction interface,” Appl. Phys. Express 10,035701 (2017).32Y. Zhang, Z. Xia, J. Mcglone, W. Sun, C. Joishi, A. R. Arehart,S. A. Ringel, and S. Rajan, “Evaluation of Low-temperature saturation velocity in-(AlxGa1–x)2O3/Ga2O3 modulation-doped field-effect transistors,” IEEE Trans.Electron Devices 66, 1574–1578 (2019).33K. Nomura, K. Goto, R. Togashi, H. Murakami, Y. Kumagai, A. Kuramata,S. Yamakoshi, and A. Koukitu, “Thermodynamic study of β-Ga2O3 growth byhalide vapor phase epitaxy,” J. Cryst. Growth 405, 19–22 (2014).34Y. Oshima, E. G. Vίllora, and K. Shimamura, “Quasi-heteroepitaxial growth ofβ-Ga2O3 on off-angled sapphire (0 0 0 1) substrates by halide vapor phaseepitaxy,” J. Cryst. Growth 410, 53–58 (2015).35K. Konishi, K. Goto, R. Togashi, H. Murakami, M. Higashiwaki, A. Kuramata,S. Yamakoshi, B. Monemar, and Y. Kumagai, “Comparison of O2 and H2O asoxygen source for homoepitaxial growth of β-Ga2O3 layers by halide vaporphase epitaxy,” J. Cryst. Growth 492, 39–44 (2018).36H. Murakami, K. Nomura, K. Goto, K. Sasaki, K. Kawara, Q. T. Thieu,R. Togashi, Y. Kumagai, M. Higashiwaki, A. Kuramata, S. Yamakoshi,B. Monemar, and A. Koukitu, “Homoepitaxial growth of β-Ga2O3 layers byhalide vapor phase epitaxy,” Appl. Phys. Express 8, 015503 (2015).37J. H. Leach, K. Udwary, J. Rumsey, G. Dodson, H. Splawn, and K. R. Evans,“Halide vapor phase epitaxial growth of β-Ga2O3 and α-Ga2O3 films,” APLMater. 7, 022504 (2019).38J. B. Varley, J. R. Weber, A. Janotti, and C. G. Van de Walle, “Oxygen vacan-cies and donor impurities in β-Ga2O3,” Appl. Phys. Lett. 97, 142106 (2010).39M. Higashiwaki, K. Sasaki, H. Murakami, Y. Kumagai, A. Koukitu,A. Kuramata, T. Masui, and S. Yamakoshi, “Recent progress in Ga2O3 powerdevices,” Semicond. Sci. Technol. 31, 034001 (2016).40F. Alema, B. Hertog, A. Osinsky, P. Mukhopadhyay, M. Toporkov, andW. V. Schoenfeld, “Fast growth rate of epitaxial β–Ga2O3 by close coupled show-erhead MOCVD,” J. Cryst. Growth 475, 77–82 (2017).41J. L. Lyons, D. Steiauf, A. Janotti, and C. G. Van de Walle, “Carbon as ashallow donor in transparent conducting oxides,” Phys. Rev. Appl. 2, 064005(2014).42Z. Feng, A. F. M. Anhar Uddin Bhuiyan, M. R. Karim, and H. Zhao,“MOCVD homoepitaxy of Si-doped (010) β-Ga2O3 thin films with superiortransport properties,” Appl. Phys. Lett. 114, 250601 (2019).43R. Miller, F. Alema, and A. Osinsky, “Epitaxial β-Ga2O3 and β-(AlxGa1−x)2O3--Ga2O3 heterostructures growth for power electronics,” IEEE Trans. Semicond.Manuf. 31, 467–474 (2018).44Y. Zhang, F. Alema, A. Mauze, O. S. Koksaldi, R. Miller, A. Osinsky, andJ. S. Speck, “MOCVD grown epitaxial β-Ga2O3 thin film with an electron mobil-ity of 176 cm 2 /V s at room temperature,” APL Mater. 7, 022506 (2019).45F. Alema, B. Hertog, P. Mukhopadhyay, Y. Zhang, A. Mauze, A. Osinsky,W. V. Schoenfeld, J. S. Speck, and T. Vogt, “Solar blind Schottky photodiodebased on an MOCVD-grown homoepitaxial β-Ga2O3 thin film,” APL Mater. 7,022527 (2019).46D. Gogova, G. Wagner, M. Baldini, M. Schmidbauer, K. Irmscher,R. Schewski, Z. Galazka, M. Albrecht, and R. Fornari, “Structural properties ofSi-doped β-Ga2O3 layers grown by MOVPE,” J. Cryst. Growth 401, 665–669(2014).47D. Shinohara and S. Fujita, “Heteroepitaxy of corundum-structured α-Ga2O3thin films on α-Al2O3 substrates by ultrasonic mist chemical vapor deposition,”Jpn. J. Appl. Phys. 47, 7311–7313 (2008).48R. Jinno, T. Uchida, K. Kaneko, and S. Fujita, “Reduction in edge dislocationdensity in corundum-structured α-Ga2O3 layers on sapphire substrates withquasi-graded α-(Al,Ga)2O3 buffer layers,” Appl. Phys. Express 9, 071101 (2016).49S. Fujita and K. Kaneko, “Epitaxial growth of corundum-structured wide bandgap III-oxide semiconductor thin films,” J. Cryst. Growth 401, 588–592 (2014).50K. Kaneko, K. Suzuki, Y. Ito, and S. Fujita, “Growth characteristics ofcorundum-structured α-(AlxGa1−x)2O3/Ga2O3 heterostructures on sapphire sub-strates,” J. Cryst. Growth 436, 150–154 (2016).51K. Akaiwa and S. Fujita, “Electrical conductive corundum-structured α-Ga2O3thin films on sapphire with Tin-doping grown by spray-assisted mist chemicalvapor deposition,” Jpn. J. Appl. Phys. 51, 070203 (2012).52M. Oda, R. Tokuda, H. Kambara, T. Tanikawa, T. Sasaki, and T. Hitora,“Schottky barrier diodes of corundum-structured gallium oxide showingon-resistance of 0.1 mΩ cm 2 grown by MIST EPITAXY®,” Appl. Phys. Express9, 021101 (2016).53K. Akaiwa, K. Kaneko, K. Ichino, and S. Fujita, “Conductivity control ofSn-doped α-Ga2O3 thin films grown on sapphire substrates,” Jpn. J. Appl. Phys.55, 1202BA (2016).Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-15Published under license by AIP Publishing. 29 January 2024 02:24:51https://doi.org/10.1149/2.0021702jsshttps://doi.org/10.1149/2.0021702jsshttps://doi.org/10.1063/1.2919728https://doi.org/10.1063/1.2919728https://doi.org/10.1063/1.119233https://doi.org/10.1063/1.119233https://doi.org/10.1016/j.jcrysgro.2016.04.022https://doi.org/10.1143/JJAP.51.040207https://doi.org/10.7567/APEX.8.055501https://doi.org/10.1103/PhysRevB.73.094125https://doi.org/10.1063/1.1754500https://doi.org/10.1063/1.4971957https://doi.org/10.1143/APEX.5.035502https://doi.org/10.1063/1.4942002https://doi.org/10.1016/j.matchemphys.2011.12.070https://doi.org/10.1016/j.matchemphys.2011.12.070https://doi.org/10.7567/APEX.10.071101https://doi.org/10.1111/j.1151-2916.1952.tb13087.xhttps://doi.org/10.1116/1.4922340https://doi.org/10.1088/1361-6641/aa53a7https://doi.org/10.7567/APEX.10.041102https://doi.org/10.7567/APEX.10.035701https://doi.org/10.1109/TED.2018.2889573https://doi.org/10.1109/TED.2018.2889573https://doi.org/10.1016/j.jcrysgro.2014.06.051https://doi.org/10.1016/j.jcrysgro.2014.10.038https://doi.org/10.1016/j.jcrysgro.2018.04.009https://doi.org/10.7567/APEX.8.015503https://doi.org/10.1063/1.5055680https://doi.org/10.1063/1.5055680https://doi.org/10.1063/1.3499306https://doi.org/10.1088/0268-1242/31/3/034001https://doi.org/10.1016/j.jcrysgro.2017.06.001https://doi.org/10.1103/PhysRevApplied.2.064005https://doi.org/10.1063/1.5109678https://doi.org/10.1109/TSM.2018.2873488https://doi.org/10.1109/TSM.2018.2873488https://doi.org/10.1063/1.5058059https://doi.org/10.1063/1.5064471https://doi.org/10.1016/j.jcrysgro.2013.11.056https://doi.org/10.1143/JJAP.47.7311https://doi.org/10.7567/APEX.9.071101https://doi.org/10.1016/j.jcrysgro.2014.02.032https://doi.org/10.1016/j.jcrysgro.2015.12.013https://doi.org/10.1143/JJAP.51.070203https://doi.org/10.7567/APEX.9.021101https://doi.org/10.7567/JJAP.55.1202BAhttps://aip.scitation.org/journal/jap54Y. Oshima, K. Kawara, T. Shinohe, T. Hitora, M. Kasu, and S. Fujita,“Epitaxial lateral overgrowth of α-Ga2O3 by halide vapor phase epitaxy,” APLMater. 7, 022503 (2019).55Y. Oshima, “Heteroepitaxial growth of α and ε-Ga2O3,” in Gallium Oxide:Crystal Growth, Materials Properties, and Devices (Springer, 2019).56H. Sun, K.-H. Li, C. G. T. Castanedo, S. Okur, G. S. Tompa, T. Salagaj,S. Lopatin, A. Genovese, and X. Li, “HCl flow-induced phase change of α-, β-, andε-Ga2O3 films grown by MOCVD,” Cryst. Growth Des. 18, 2370–2376 (2018).57Z. Cheng, M. Hanke, P. Vogt, O. Bierwagen, and A. Trampert, “Phase formationand strain relaxation of Ga2O3 on c-plane and a-plane sapphire substrates as studiedby synchrotron-based x-ray diffraction,” Appl. Phys. Lett. 111, 162104 (2017).58R. Jinno, Y. Cho, K. Lee, V. Protasenko, H. G. Xing, and D. Jena, “Stabilizationof alpha-Ga2O3 on m-plane alpha-Al2O3 by Plasma-Assisted MBE,” in 66thJSAP Spring Meeting, Tokyo, Japan (2019).59G. T. Dang, T. Kawaharamura, M. Furuta, and M. W. Allen, “Mist-CVDgrown Sn-doped alpha-Ga2O3 MESFETs,” IEEE Trans. Electron Devices 62,3640–3644 (2015).60Z. Hu, K. Nomoto, W. Li, R. Jinno, T. Nakamura, D. Jena, and H. Xing,“1.6 kV vertical Ga2O3 FinFETs with source-connected field plates andnormally-off operation,” in 2019 31st International Symposium on PowerSemiconductor Devices and ICs (ISPSD), (IEEE, 2019), pp. 483–486.612018 Ground breaking work on gallium oxide (Ga2O3) normally-off transistorFlosfia.62M. Higashiwaki, K. Sasaki, T. Kamimura, M. Hoi Wong, D. Krishnamurthy,A. Kuramata, T. Masui, and S. Yamakoshi, “Depletion-mode Ga2O3metal-oxide-semiconductor field-effect transistors on β-Ga2O3 (010) substratesand temperature dependence of their device characteristics,” Appl. Phys. Lett.103, 1–5 (2013).63M. H. Wong, K. Sasaki, A. Kuramata, S. Yamakoshi, and M. Higashiwaki,“Electron channel mobility in silicon-doped Ga2O3 MOSFETs with a resistivebuffer layer,” Jpn. J. Appl. Phys. 55, 1202–1211 (2016).64A. J. Green, K. D. Chabak, E. R. Heller, R. C. Fitch, M. Baldini, A. Fiedler,K. Irmscher, G. Wagner, Z. Galazka, S. E. Tetlak, A. Crespo, K. Leedy, andG. H. Jessen, “3.8-MV/cm breakdown strength of MOVPE-grown Sn-dopedβ-Ga2O3 MOSFETs,” IEEE Electron Device Lett. 37, 902–905 (2016).65Y. Zhang, A. Neal, Z. Xia, C. Joishi, J. M. Johnson, Y. Zheng, S. Bajaj,M. Brenner, D. Dorsey, K. Chabak, G. Jessen, J. Hwang, S. Mou, J. P. Heremans,and S. Rajan, “Demonstration of high mobility and quantum transport inmodulation-doped β-(AlxGa1-x)2O3/Ga2O3 heterostructures,” Appl. Phys. Lett.112, 173502 (2018).66C. Joishi, Y. Zhang, Z. Xia, W. Sun, A. R. Arehart, S. Ringel, S. Lodha, andS. Rajan, “Breakdown characteristics of β-(Al0.22Ga0.78)2O3/Ga2O3 field-platedmodulation doped field effect transistors with SiNx passivation,” IEEE ElectronDevice Lett. 40, 1241 (2019).67Y. Zhang, C. Joishi, Z. Xia, M. Brenner, S. Lodha, and S. Rajan,“Demonstration of β-(AlxGa1-x)2O3/Ga2O3 double heterostructure field effecttransistors,” Appl. Phys. Lett. 112, 233503 (2018).68Y. Zhang, M. Sun, D. Piedra, J. Hu, Z. Liu, Y. Lin, X. Gao, K. Shepard, andT. Palacios, “1200 V GaN vertical fin power field-effect transistors,” in 2017 IEEEInternational Electron Devices Meeting (IEDM), (IEEE, 2017), pp. 9.2.1–9.2.4.69K. D. Chabak, N. Moser, A. J. Green, D. E. Walker, S. E. Tetlak, E. Heller,A. Crespo, R. Fitch, J. P. McCandless, K. Leedy, M. Baldini, G. Wagner,Z. Galazka, X. Li, and G. Jessen, “Enhancement-mode Ga2O3 wrap-gate finfield-effect transistors on native (100) β-Ga2O3 substrate with high breakdownvoltage,” Appl. Phys. Lett. 109, 213501 (2016).70Z. Hu, K. Nomoto, W. Li, L. Zhang, J.-H. Shin, N. Tanen, T. Nakamura,D. Jena, and H. G. Xing, “Vertical fin Ga2O3 power field-effect transistors withon/off ratio >109,” in 2017 75th Annual Device Research Conference (DRC)(IEEE, South Bend, IN, 2017).71Z. Hu, K. Nomoto, W. Li, Z. Zhang, N. Tanen, Q. T. Thieu, K. Sasaki,A. Kuramata, T. Nakamura, D. Jena, and H. G. Xing, “Breakdown mechanism in1 kA/cm 2 and 960 V E-mode β-Ga2O3 vertical transistors,” Appl. Phys. Lett.113, 122103 (2018).72B. Song, A. K. Verma, K. Nomoto, M. Zhu, D. Jena, and H. G. Xing, “VerticalGa2O3 Schottky barrier diodes on single-crystal β-Ga2O3 (-201) substrates,” in2016 74th Annual Device Research Conference (IEEE, 2016), Vol. 3, pp. 1–2.73W. Li, K. Nomoto, Z. Hu, N. Tanen, K. Sasaki, A. Kuramata, D. Jena, andH. G. Xing, “1.5 kV Vertical Ga2O3 Trench-MIS Schottky barrier diodes,” in2018 76th Device Research Conference (IEEE, 2018), Vol. 110, pp. 1–2.74W. Li, Z. Hu, K. Nomoto, Z. Zhang, J-Y. Hsu, Q. T. Thieu, K. Sasaki,A. Kuramata, D. Jena, and H. G. Xing, “1230 V β-Ga2O3 trench Schottky barrierdiodes with an ultra-low leakage current of <1 μA/cm2,” Appl. Phys. Lett., 113202101 (2018).75E. Ahmadi, F. Wu, H. Li, S. W. Kaun, M. Tahhan, K. Hestroffer,S. Keller, J. S. Speck, and U. K. Mishra, “N-face GaN/AlN/GaN/InAlN andGaN/AlN/AlGaN/GaN/InAlN high-electron-mobility transistor structures grownby plasma-assisted molecular beam epitaxy on vicinal substrates,” Semicond. Sci.Technol. 30, 055012 (2015).76A. T. Neal, S. Mou, S. Rafique, H. Zhao, E. Ahmadi J. S., D. B. Speck,N. Thomson, K. D. Moser, and G. H. J. Chabak, “Donors and acceptors inβ-Ga2O3,” Appl. Phys. Lett. 113, 062101 (2018).77N. Moser, J. McCandless, A. Crespo, K. Leedy, A. Green, A. Neal, S. Mou,E. Ahmadi, J. Speck, K. Chabak, N. Peixoto, and G. Jessen, “Ge-Doped β-Ga2O3MOSFETs,” IEEE Electron Device Lett. 38, 775–778 (2017).78M. Higashiwaki, Current Status of Gallium Oxide-Based Power DeviceTechnology (IEEE, 2015), pp. 1–4.79M. H. Wong, K. Sasaki, A. Kuramata, S. Yamakoshi, and M. Higashiwaki,“Field-Plated Ga2O3 MOSFETs With a breakdown voltage of over 750 V,”IEEE Electron Device Lett. 37, 212–215 (2016).80Y. Yao, R. F. Davis, and L. M. Porter, “Investigation of different metals as ohmiccontacts to β-Ga2O3: Comparison and analysis of electrical behavior, morphology,and other physical properties,” J. Electron. Mater. 46, 2053–2060 (2017).81M.-H. Lee and R. L. Peterson, “Interfacial reactions of titanium/gold ohmiccontacts with Sn-doped β-Ga2O3,” APL Mater. 7, 022524 (2019).82M. Higashiwaki, K. Sasaki, K. Goto, K. Nomura, Q. T. Thieu, R. Togashi,H. Murakami, Y. Kumagai, B. Monemar, A. Koukitu, A. Kuramata, and S. Yamakoshi,“Ga2O3 Schottky barrier diodes with n–Ga2O3 drift layers grown by HVPE,” inDevice Research Conference-Conference Digest (DRC, 2015), Vol. 2015, pp. 29–30.83K. Sasaki, M. Higashiwaki, A. Kuramata, T. Masui, and S. Yamakoshi,“Ga2o3 Schottky barrier diodes fabricated by using single-crystal β-Ga2O3 (010)substrates,” IEEE Electron Device Lett. 34, 493–495 (2013).84D. Khan, D. Gajula, S. Okur, G. S. Tompa, and G. Koley, “β-Ga2O3 thin filmbased lateral and vertical Schottky barrier diode,” ECS J. Solid State Sci. Technol.8, Q106–10 (2019).85Y. Yao, R. Gangireddy, J. Kim, K. K. Das, R. F. Davis, and L. M. Porter,“Electrical behavior of β-Ga2O3 Schottky diodes with different Schottky metals,”J. Vac. Sci. Technol. B 35, 03D113 (2017).86D. Splith, S. Müller, F. Schmidt, H. von Wenckstern, J. J. van Rensburg,W. E. Meyer, and M. Grundmann, “Determination of the mean and the homo-geneous barrier height of Cu Schottky contacts on heteroepitaxial β-Ga2O3 thinfilms grown by pulsed laser deposition,” Phys. Status Solidi 211, 40–47 (2014).87Z. Hu, K. Nomoto, W. Li, N. Tanen, K. Sasaki, A. Kuramata, T. Nakamura,D. Jena, and H. G. Xing, “Enhancement-Mode Ga2O3 vertical transistors Withbreakdown voltage >1 kV,” IEEE Electron Device Lett. 39, 869–872 (2018).88M. H. Wong, Y. Nakata, A. Kuramata, S. Yamakoshi, and M. Higashiwaki,“Enhancement-mode Ga2O3 MOSFETs with Si-ion-implanted source anddrain,” Appl. Phys. Express 10, 3–6 (2017).89M. H. Wong, K. Goto, H. Murakami, Y. Kumagai, and M. Higashiwaki,“Current aperture vertical β-Ga2O3 MOSFETs fabricated by N- and Si-Ionimplantation doping,” IEEE Electron Device Lett. 40, 431–434 (2019).90K. Sasaki, M. Higashiwaki, A. Kuramata, T. Masui, and S. Yamakoshi, “Si-Ionimplantation doping in β-Ga2O3 and its application to fabrication of low-resistance ohmic contacts,” Appl. Phys. Express 6, 086502 (2013).91M. H. Wong, C.-H. Lin, A. Kuramata, S. Yamakoshi, H. Murakami,Y. Kumagai, and M. Higashiwaki, “Acceptor doping of β-Ga2O3 by Mg and Nion implantations,” Appl. Phys. Lett. 113, 102103 (2018).Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-16Published under license by AIP Publishing. 29 January 2024 02:24:51https://doi.org/10.1063/1.5051058https://doi.org/10.1063/1.5051058https://doi.org/10.1021/acs.cgd.7b01791https://doi.org/10.1063/1.4998804https://doi.org/10.1109/TED.2015.2477438https://doi.org/10.1063/1.4821858https://doi.org/10.7567/JJAP.55.1202B9https://doi.org/10.1109/LED.2016.2568139https://doi.org/10.1063/1.5025704https://doi.org/10.1063/1.5037095https://doi.org/10.1063/1.4967931https://doi.org/10.1109/DRC.2017.7999512https://doi.org/10.1063/1.5038105https://doi.org/10.1063/1.5052368https://doi.org/10.1088/0268-1242/30/5/055012https://doi.org/10.1088/0268-1242/30/5/055012https://doi.org/10.1063/1.5034474https://doi.org/10.1109/LED.2017.2697359https://doi.org/10.1109/CSICS.2015.7314495https://doi.org/10.1109/CSICS.2015.7314495https://doi.org/10.1109/CSICS.2015.7314495https://doi.org/10.1109/LED.2015.2512279https://doi.org/10.1007/s11664-016-5121-1https://doi.org/10.1063/1.5054624https://doi.org/10.1109/LED.2013.2244057https://doi.org/10.1149/2.0211906jsshttps://doi.org/10.1116/1.4980042https://doi.org/10.1002/pssa.201330088https://doi.org/10.1109/LED.2018.2830184https://doi.org/10.7567/APEX.10.041101https://doi.org/10.1109/LED.2018.2884542https://doi.org/10.7567/APEX.6.086502https://doi.org/10.1063/1.5050040https://aip.scitation.org/journal/jap92M. H. Wong, K. Goto, Y. Morikawa, A. Kuramata, S. Yamakoshi, H. Murakami,Y. Kumagai, and M. Higashiwaki, Appl. Phys. Express 11, 064102 (2018).93A. Y. Polyakov, I.-H. Lee, N. B. Smirnov, E. B. Yakimov, I. V. Shchemerov,A. V. Chernykh, A. I. Kochkova, A. A. Vasilev, P. H. Carey, F. Ren, D. J. Smith,and S. J. Pearton, “Defects at the surface of β-Ga2O3 produced by Ar plasmaexposure,” APL Mater. 7, 061102 (2019).94J. Yang, S. Ahn, F. Ren, S. Pearton, R. Khanna, K. Bevlin, D. Geerpuram, andA. Kuramata, “Inductively coupled plasma etching of bulk, single-crystalGa2O3,” J. Vac. Sci. Technol. B 35, 031205 (2017).95S. Jang, S. Jung, K. Beers, J. Yang, F. Ren, A. Kuramata, S. J. Pearton, andK. H. Baik, “A comparative study of wet etching and contacts on (2¯01) and(010) oriented β-Ga2O3,” J. Alloys Compd. 731, 118–125 (2018).96L. Zhang, A. Verma, H. G. Xing, and D. Jena, “Inductively-coupled-plasmareactive ion etching of single-crystal β-Ga2O3,” Jpn. J. Appl. Phys. 56, 030304(2017).97J. Yang, F. Ren, R. Khanna, K. Bevlin, D. Geerpuram, L.-C. Tung, J. Lin,H. Jiang, J. Lee, E. Flitsiyan, L. Chernyak, S. J. Pearton, and A. Kuramata,“Annealing of dry etch damage in metallized and bare (-201) Ga2O3,” J. Vac. Sci.Technol. B 35, 051201 (2017).98J. Yang, S. Ahn, F. Ren, R. Khanna, K. Bevlin, D. Geerpuram, S. J. Pearton,and A. Kuramata, “Inductively coupled plasma etch damage in (-201) Ga2O3Schottky diodes,” Appl. Phys. Lett. 110, 142101 (2017).99J. E. Hogan, S. W. Kaun, E. Ahmadi, Y. Oshima, and J. S. Speck,“Chlorine-based dry etching of β-Ga2O3,” Semicond. Sci. Technol. 31, 065006(2016).100Z. A. Jian, Y. Oshima, S. Wright, K. Owen, and E. Ahmadi, “Chlorine-basedinductive coupled plasma etching of α-Ga2O3,” Semicond. Sci. Technol. 34,035006 (2019).101C.-H. Lin, Y. Yuda, M. H. Wong, M. Sato, N. Takekawa, K. Konishi,T. Watahiki, M. Yamamuka, H. Murakami, Y. Kumagai, and M. Higashiwaki,“Vertical Ga2O3 Schottky barrier diodes with guard ring formed by nitrogen-Ionimplantation,” IEEE Electron Device Lett. 40, 18937409 (2019).102W. Li, Z. Hu, K. Nomoto, R. Jinno, Z. Zhang, T. Q. Tu, K. Sasaki,A. Kuramata, D. Jena, and H. G. Xing, “2.44 kV Ga2O3 vertical trench Schottkybarrier diodes with very low reverse leakage current,” in 2018 IEEE InternationalElectron Devices Meeting (IEDM), (IEEE, 2018), pp. 8.5.1–8.5.4.103J. Yang, F. Renl, M. Tadjer, S. J. Pearton, and A. Kuramata, “2.3 kVField-Plated Vertical Ga2O3 Schottky rectifiers and 1 a forward current with650 V reverse breakdown Ga2O3 field-plated Schottky barrier diodes,” in 201876th Device Research Conference (DRC), (IEEE, 2018), pp. 1–2.104W. Li, Z. Hu, K. Nomoto, Z. Zhang, J.-Y. Hsu, Q. T. Thieu, K. Sasaki,A. Kuramata, D. Jena, and H. G. Xing, “1230 V β-Ga2O3 trench Schottky barrierdiodes with an ultra-low leakage current of >1 μA/cm2,” Appl. Phys. Lett. 113,202101 (2018).105J. Yang, S. Ahn, F. Ren, S. J. Pearton, S. Jang, and A. Kuramata, “High break-down voltage (−201) β-Ga2O3 Schottky rectifiers,” IEEE Electron Device Lett.38, 906–909 (2017).106K. Sasaki, D. Wakimoto, Q. T. Thieu, Y. Koishikawa, A. Kuramata,M. Higashiwaki, and S. Yamakoshi, “First demonstration of Ga2O3 trenchMOS-type Schottky barrier diodes,” IEEE Electron Device Lett. 38, 783–785(2017).107J. Yang, S. Ahn, F. Ren, S. J. Pearton, S. Jang, J. Kim, and A. Kuramata, “Highreverse breakdown voltage Schottky rectifiers without edge termination onGa2O3,” Appl. Phys. Lett. 110, 192101 (2017).108K. Konishi, K. Goto, H. Murakami, Y. Kumagai, A. Kuramata, S. Yamakoshi,and M. Higashiwaki, “1-kV vertical Ga2O3 field-plated Schottky barrier diodes,”Appl. Phys. Lett. 110, 103506 (2017).109J. Yang, F. Ren, M. Tadjer, S. J. Pearton, and A. Kuramata, “Ga2o3 Schottkyrectifiers with 1 ampere forward current, 650 V reverse breakdown and 26.5 MWcm−2figure-of-merit,” AIP Adv. 8, 055026 (2018).110M. Higashiwaki, K. Sasaki, A. Kuramata, T. Masui, and S. Yamakoshi,“Gallium oxide (Ga2O3) metal-semiconductor field-effect transistors onsingle-crystal β-Ga2O3 (010) substrates,” Appl. Phys. Lett. 100, 1–4 (2012).111M. Higashiwaki, K. Sasaki, M. H. Wong, T. Kamimura, D. Krishnamurthy,A. Kuramata, T. Masui, and S. Yamakoshi, “Depletion-mode Ga2O3 MOSFETson β-Ga2O3 (010) substrates with Si-ion-implanted channel and contacts,” inTechnical Digest—International Electron Devices Meeting, (IEDM, 2013), Vol. 3,pp. 707–710.112M. H. Wong, K. Sasaki, A. Kuramata, and S. Yamakoshi, “Field-Plated Ga2O3MOSFETs With a breakdown voltage of over 750 V,” IEEE Electron Device Lett.37, 212–215 (2016).113M. Singh, M. A. Casbon, M. J. Uren, J. W. Pomeroy, S. Dalcanale,S. Karboyan, P. J. Tasker, M. H. Wong, K. Sasaki, A. Kuramata, S. Yamakoshi,M. Higashiwaki, and M. Kuball, “Pulsed large signal RF performance of field-plated Ga2O3 MOSFETs,” IEEE Electron Device Lett. 39, 1572–1575 (2018).114M. H. Wong, A. Takeyama, T. Makino, T. Ohshima, K. Sasaki, A. Kuramata,S. Yamakoshi, and M. Higashiwaki, “Radiation hardness of β-Ga2O3metal-oxide-semiconductor field-effect transistors against gamma-ray irradia-tion,” Appl. Phys. Lett. 112, 023503 (2018).115L. Zhang, A. Verma, H. G. Xing, and D. Jena, “Inductively-coupled-plasma reac-tive ion etching of single-crystal β-Ga2O3,” Jpn. J. Appl. Phys. 56, 030304 (2017).116Z. Jian, Y. Oshima, S. Wright, K. Owen, and E. Ahmadi, “Chlorine-basedinductive coupled plasma etching of α-Ga2O3,” Semicond. Sci. Technol. 34,035006 (2019).117N. A. Moser, J. P. McCandless, A. Crespo, K. D. Leedy, A. J. Green,E. R. Heller, K. D. Chabak, N. Peixoto, and G. H. Jessen, “High pulsed currentdensity β-Ga2O3 MOSFETs verified by an analytical model corrected for inter-face charge,” Appl. Phys. Lett. 110, 143505 (2017).118K. D. Chabak, D. E. W. Jr, A. J. Green, A. Crespo, M. Lindquist, K. Leedy,S. Tetlak, N. A. Moser, and G. Jessen, “Sub-micron gallium oxide radio fre-quency field-effect transistors,” in 2018 IEEE MTT-S International MicrowaveWorkshop Series Advanced Materials and Processes RF THz Applications (IEEE,2018), pp. 1–3.119A. J. Green, K. D. Chabak, M. Baldini, N. Moser, R. Gilbert, R. C. Fitch,G. Wagner, Z. Galazka, J. McCandless, A. Crespo, K. Leedy, and G. H. Jessen,“β-Ga2O3 MOSFETs for radio frequency operation,” IEEE Electron Device Lett.38, 790–793 (2017).120K. Zeng, A. Vaidya, and U. Singisetti, “A field-plated Ga2O3 MOSFET withnear 2-kV breakdown voltage and 520 mΩ cm2 on-resistance,” Appl. Phys.Express 12, 081003 (2019).121Y. Zhang, A. R. Arehart, and S. Ringel, Appl. Phys. Lett. 113, 123501 (2018).122K. Zeng, A. Vaidya, and U. Singisetti, “1.85 kV breakdown voltage in lateralfield-plated Ga2O3 MOSFETs,” IEEE Electron Device Lett. 39, 1385–1388 (2018).123K. Chabak, A. Green, N. Moser, S. Tetlak, J. McCandless, K. Leedy, R. Fitch,A. Crespo, and G. Jessen, “Gate-recessed, laterally-scaled β-Ga2O3 MOSFETswith high-voltage enhancement-mode operation,” in 2017 75th Annual DeviceResearch Conference (DRC), (IEEE, 2017), Vol. 38, pp. 1–2.124K. D. Chabak, J. P. McCandless, N. A. Moser, A. J. Green, K. Mahalingam,A. Crespo, N. Hendricks, B. M. Howe, S. E. Tetlak, K. Leedy, R. C. Fitch,D. Wakimoto, K. Sasaki, A. Kuramata, and G. H. Jessen, “Recessed-Gateenhancement-mode β-Ga2O3 MOSFETs,” IEEE Electron Device Lett. 39, 67–70(2018).125S. Krishnamoorthy, Z. Xia, C. Joishi, Y. Zhang, J. McGlone, J. Johnson,M. Brenner, A. R. Arehart, J. Hwang, S. Lodha, S. Rajan, “Modulation-dopedβ-(Al0.2Ga0.8)2O3/Ga2O3 field-effect transistor,” Appl. Phys. Lett. 111, 023502(2017).126Y. Zhang, Z. Xia, J. McGlone, W. Sun, C. Joishi, A. R. Arehart, S. A. Ringel,and S. Rajan, IEEE Trans. Electron Devices 66, 1574 (2019).127H. Okumura, Y. Kato, T. Oshima, and T. Palacios, “Demonstration of lateralfield-effect transistors using Sn-doped β-(AlGa)2O3 (010),” Jpn. J. Appl. Phys.58, SBBD12 (2019).128S. Krishnamoorthy, Z. Xia, S. Bajaj, M. Brenner, and S. Rajan, “Delta-dopedβ-gallium oxide field-effect transistor,” Appl. Phys. Express 10, 051102(2017).129K. Ghosh and U. Singisetti, “Ab initio velocity-field curves in monoclinicβ-Ga2O3,” J. Appl. Phys. 122, 035702 (2017).Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-17Published under license by AIP Publishing. 29 January 2024 02:24:51https://doi.org/10.7567/APEX.11.064102https://doi.org/10.1063/1.5109025https://doi.org/10.1116/1.4982714https://doi.org/10.1016/j.jallcom.2017.09.336https://doi.org/10.7567/JJAP.56.030304https://doi.org/10.1116/1.4986300https://doi.org/10.1116/1.4986300https://doi.org/10.1063/1.4979592https://doi.org/10.1088/0268-1242/31/6/065006https://doi.org/10.1088/1361-6641/aafeb2https://doi.org/10.1109/LED.2019.2927790https://doi.org/10.1063/1.5052368https://doi.org/10.1109/LED.2017.2703609https://doi.org/10.1109/LED.2017.2696986https://doi.org/10.1063/1.4983203https://doi.org/10.1063/1.4977857https://doi.org/10.1063/1.5034444https://doi.org/10.1063/1.3674287https://doi.org/10.1109/LED.2015.2512279https://doi.org/10.1109/LED.2018.2865832https://doi.org/10.1063/1.5017810https://doi.org/10.7567/JJAP.56.030304https://doi.org/10.1088/1361-6641/aafeb2https://doi.org/10.1063/1.4979789https://doi.org/10.1109/IMWS-AMP.2018.8457153https://doi.org/10.1109/IMWS-AMP.2018.8457153https://doi.org/10.1109/IMWS-AMP.2018.8457153https://doi.org/10.1109/IMWS-AMP.2018.8457153https://doi.org/10.1109/IMWS-AMP.2018.8457153https://doi.org/10.1109/IMWS-AMP.2018.8457153https://doi.org/10.1109/IMWS-AMP.2018.8457153https://doi.org/10.1109/LED.2017.2694805https://doi.org/10.7567/1882-0786/ab2e86https://doi.org/10.7567/1882-0786/ab2e86https://doi.org/10.1063/1.5039502https://doi.org/10.1109/LED.2018.2859049https://doi.org/10.1109/LED.2017.2779867https://doi.org/10.1063/1.4993569https://doi.org/10.1109/TED.2018.2889573https://doi.org/10.7567/1347-4065/ab002bhttps://doi.org/10.7567/APEX.10.051102https://doi.org/10.1063/1.4986174https://aip.scitation.org/journal/jap130K. A. Mengle, G. Shi, D. Bayerl, and E. Kioupakis, “First-principles calcula-tions of the near-edge optical properties of β-Ga2O3,” Appl. Phys. Lett. 109,212104 (2016).131N. Ma, N. Tanen, A. Verma, Z. Guo, T. Luo, H. G. Xing, and D. Jena, “Intrinsicelectron mobility limits in β-Ga2O3,” Appl. Phys. Lett. 109, 212101 (2016).132Y. Zhang, Z. Xia, C. Joishi, and S. Rajan, “Design and Demonstration of(AlxGal-x)2O3/Ga2O3 Double Heterostructure Field Effect Transistor (DHFET),”in Device Research Conference—Conference Digest, DRC, (IEEE, 2018), Vol. 2018,pp. 1–2.133H. Zhou, K. Maize, G. Qiu, A. Shakouri, and P. D. Ye, “β-Ga2O3 on insulatorfield-effect transistors with drain currents exceeding 1.5 A/mm and their self-heating effect,” Appl. Phys. Lett. 111, 092102 (2017).134M. H. Wong, K. Goto, A. Kuramata, S. Yamakoshi, H. Murakami,Y. Kumagai, and M. Higashiwaki, “First demonstration of vertical Ga2O3MOSFET: Planar structure with a current aperture,” in 2017 75th Annual DeviceResearch Conference (DRC), 25–28 June 2017 (IEEE, South Bend, IN, 2017).135B. Chatterjee, K. Zeng, C. D. Nordquist, U. Singisetti, and S. Choi,“Device-Level thermal management of gallium oxide field-effect transistors,”IEEE Trans. Compon. Packag. Manuf. Technol. (published online).136K. Kaneko, S. Fujita, and T. Hitora, “A power device material of corundum-structured α-Ga2O3 fabricated by MIST EPITAXY® technique,” Jpn. J. Appl.Phys. 57, 02CB18 (2018).137K. Akaiwa, K. Kaneko, S. Fujita, E. Chikoidze, and Y. Dumont, “Room tem-perature ferromagnetism in conducting α-(In1−xFex)2O3 alloy films,” Appl. Phys.Lett. 106, 062405 (2015).138K. Harafuji, T. Tsuchiya, and K. Kawamura, “Molecular dynamics simulationfor evaluating melting point of wurtzite-type GaN crystal,” J. Appl. Phys. 96,2501–2512 (2004).Journal ofApplied Physics PERSPECTIVE scitation.org/journal/japJ. Appl. Phys. 126, 160901 (2019); doi: 10.1063/1.5123213 126, 160901-18Published under license by AIP Publishing. 29 January 2024 02:24:51https://doi.org/10.1063/1.4968822https://doi.org/10.1063/1.4968550https://doi.org/10.1063/1.5000735https://doi.org/10.1109/DRC.2017.7999413https://doi.org/10.1109/DRC.2017.7999413https://doi.org/10.1109/TCPMT.2019.2923356https://doi.org/10.7567/JJAP.57.02CB18https://doi.org/10.7567/JJAP.57.02CB18https://doi.org/10.1063/1.4908050https://doi.org/10.1063/1.4908050https://doi.org/10.1063/1.1772878https://aip.scitation.org/journal/jap