Pith. sign in

REVIEW 1 major objections 5 minor 3 references

Large grown-in offcuts make slow (100) Ga2O3 grow as fast as the fastest face, with low doping and usable diodes.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-13 05:42 UTC pith:HFCNV4F5

load-bearing objection Solid materials result: large grown-in EFG offcuts restore (100) MBE rates to the (010) benchmark with clean morphology and low UID; device numbers are secondary. the 1 major comments →

arxiv 2607.08929 v1 pith:HFCNV4F5 submitted 2026-07-09 cond-mat.mtrl-sci

Fast Homoepitaxy on (100) b{eta}-Ga2O3 Substrates with Large Grown-In Offcut

M. Brooks Tellekamp , Drew Haven , David Joyce , John Mangum , Henry Garland , Robert Lavelle , Kevin Schulte , Anna Sacchi
show 2 more authors
Matt Young Andriy Zakutayev
This is my paper
classification cond-mat.mtrl-sci
keywords beta-Ga2O3homoepitaxymolecular beam epitaxyoffcutedge-defined film-fed growthSchottky barrier diodeunintentional dopingstep-flow growth
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Monoclinic beta-Ga2O3 is anisotropic, so the crystal face you choose controls growth speed, defects, and breakdown field. The (100) face is theoretically attractive for power devices and can be made in large scalable ribbons, yet conventional epitaxy on it is 10–30 times slower than on (010). This paper shows that intentionally growing the ribbon with a large offcut (up to 13.4° toward −c) by rotating the seed in edge-defined film-fed growth recovers the full growth rate—matching or exceeding the rate measured on (010) under the same flux—while still producing step-flow morphology. The same films show unintentional doping as low as 2×10^15 cm−3 and simple planar Schottky diodes that break down at an average field of 1.56 MV/cm. The practical claim is that crystal face and offcut angle must be co-designed, and that scalable (100) wafers are now viable for fast MBE and device work.

Core claim

Molecular-beam epitaxy of beta-Ga2O3 on (100) substrates that already contain large grown-in offcuts (3.4°–13.4° toward −c) reaches growth rates of 4–5.1 nm/min—equal to the rate obtained on the fast (010) face under identical conditions—while preserving (100) terraces, step-flow growth, unintentional doping of 2–7×10^15 cm−3, and planar Schottky diodes with ~10^5 rectification and 1.56 MV/cm average breakdown.

What carries the argument

Grown-in offcut ribbons produced by edge-defined film-fed growth with a rotated seed. The large offcut supplies a high density of (2−01) step edges that enable step-flow growth and suppress the desorption-limited rate penalty that normally makes (100) slow.

Load-bearing premise

That the 1.56 MV/cm average breakdown extracted from simple, unterminated planar Schottky diodes on a thin UID layer fairly represents the epilayer rather than being limited by punch-through into the heavily doped substrate or by edge effects.

What would settle it

Fabricate identically processed planar diodes on thicker UID epilayers (or with edge termination) grown on the same high-offcut (100) material and on (001)/(010) controls; if the average breakdown field no longer reaches or exceeds literature values for the other faces, the claim of comparable device quality collapses.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Scalable (100) EFG ribbons become a practical substrate choice for MBE drift layers instead of being discarded for slow growth.
  • Growth rate on (100) continues to rise with offcut at least to 13.4°, removing the previous 6° ceiling set by boule-cutting waste.
  • Unintentional doping in the low-10^15 cm−3 range is now routinely available by MBE, enabling thicker lightly doped layers.
  • Device designers can exploit the higher theoretical [100] breakdown field without paying a growth-rate penalty.
  • CMP and pre-growth etch/anneal recipes developed for these extreme offcuts can be transferred to other low-symmetry oxides.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The near-coincidence of the 13.4° offcut with the monoclinic β-angle may be more than accidental; it may systematically expose the most favorable step-edge facets and could be generalized to other monoclinic crystals.
  • If the same large-offcut strategy works in HVPE or MOCVD, the (100) orientation could become competitive for commercial drift-layer growth, not only research MBE.
  • The low UID achieved partly by higher growth rate suggests that further rate increases (suboxide or indium-catalyzed MBE on these offcuts) may drive background doping even lower.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

1 major / 5 minor

Summary. The manuscript demonstrates that large grown-in offcuts (3.4°–13.4° toward [001̅]) on (100) β-Ga2O3 substrates, produced by EFG with rotated seeds, restore plasma-assisted MBE growth rates to 4–5.1 nm/min—matching or exceeding the authors’ (010) benchmark under identical Ga flux—while preserving (100) terraces, step-flow morphology, and low unintentional doping (2–7×10^15 cm−3). RHEED, AFM, and STEM confirm terrace structure and a clean homoepitaxial interface; planar unterminated Pt SBDs on ~550 nm UID layers show ~10^5 rectification and an average breakdown field of 1.56 MV/cm. The work argues that co-design of face and offcut makes scalable (100) wafers practical for high-rate epitaxy and devices.

Significance. If the growth-rate recovery and materials quality hold, the result removes a long-standing practical barrier to (100) β-Ga2O3, which is attractive for its predicted higher critical field, dislocation geometry, and wafer scalability. The grown-in-offcut EFG approach avoids the material waste of post-growth grinding, and the reported UID is among the lowest for MBE Ga2O3. The combination of rate-vs-offcut data, surface/interface microscopy, and basic diode metrics constitutes a solid materials-platform contribution that is of clear interest to the UWBG power-electronics community.

major comments (1)
  1. The average breakdown field of 1.56 MV/cm (Fig. 7b, ~83 V on a 550 nm UID layer) is presented as comparable to literature planar SBDs, but C–V (Fig. S10) and reverse I–V show punch-through into the Sn-doped substrate beyond ~−20 V and no edge termination. The manuscript should more carefully state that this is an average field under non-ideal geometry rather than a critical-field figure of merit for the [100] direction, and should note the limitation relative to the theoretical anisotropy claims in the introduction.
minor comments (5)
  1. Abstract and main text alternately use “10–30× slower” and “equal to the fast growth direction”; a single consistent comparison (e.g., to the authors’ own (010) rate under identical flux) would improve clarity.
  2. Table 1 lists several “n.d.” and “est.” entries; a short note on which values are measured vs. estimated would help readers assess the co-design argument.
  3. The dielectric constant used for ECV (ϵr ≈ 11.5 for 13.4° offcut) is stated without derivation; a brief justification or reference would strengthen the doping extraction.
  4. Typographical inconsistencies appear throughout (e.g., “oOcut,” “diOraction,” “eOectively,” “film”); these should be cleaned in production.
  5. Figure 3a would benefit from error bars or a statement of thickness-fringe uncertainty so that the continuous rise with offcut can be judged quantitatively.

Circularity Check

0 steps flagged

No circularity: all central claims are direct experimental measurements against external standards, not derived quantities forced by inputs or self-citation.

full rationale

The paper reports measured MBE growth rates (via XRD Laue fringes), surface morphology (AFM, RHEED, STEM), UID densities (ECV + SIMS), and planar SBD I–V characteristics on (100) β-Ga2O3 substrates with grown-in offcuts up to 13.4°. Growth-rate recovery to the (010) benchmark is obtained under identical Ga flux (1×10^{-6} torr BEP) and is not a fitted or predicted quantity. Offcut angles are measured by HRXRD; terrace structure is observed, not assumed. Low UID (~2×10^{15} cm^{-3}) is attributed to process choices (8N Ga, reduced plasma power, gate-valved Si cell, higher rate) and is independently profiled. Device metrics (on/off ~10^5, average field 1.56 MV/cm) are raw electrical data on unterminated diodes and are presented only as comparable to literature planar devices, not as a first-principles critical-field derivation. Self-citations appear only for methods or prior context and are not load-bearing for the rate, morphology, or doping results. No free parameters are fitted and then re-labeled as predictions; no uniqueness theorems or ansatzes are imported to force the outcome. The derivation chain is therefore experimental and self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

2 free parameters · 3 axioms · 0 invented entities

The work is experimental materials science. It rests on standard crystallographic and growth-domain assumptions plus a few literature values (dielectric constant, theoretical breakdown anisotropy). No new physical entities or free parameters are introduced to force the central claim; growth rates and doping are measured, not fitted.

free parameters (2)
  • Ga BEP set-point (1×10−6 torr)
    Chosen to be slightly oxygen-rich relative to the (010) calibration; the absolute value is an experimental control, not a fitted constant used to claim the rate recovery.
  • Assumed dielectric constant εr ≈ 11.5 for 13.4° offcut (100)
    Taken from literature (Schubert et al.) and used only for ECV conversion; small uncertainty does not alter the order-of-magnitude UID claim.
axioms (3)
  • domain assumption Monoclinic β-Ga2O3 exhibits strong anisotropy in growth rate, defect types, and impact-ionization coefficients that depend on crystal face and offcut direction.
    Stated in Introduction and Table 1; underpins the motivation for exploring large (100) offcuts.
  • domain assumption Step-flow growth on (100) is seeded by (2̅01) facets when the offcut is toward [001̅], suppressing twin domains.
    Cited from Schewski et al. and used to interpret RHEED/AFM/STEM of the extreme-offcut surfaces.
  • domain assumption XRD thickness fringes from an (Al,Ga)2O3 marker layer accurately give epilayer thickness under the growth conditions used.
    Standard metrology assumption for the growth-rate data in Fig. 3.

pith-pipeline@v1.1.0-grok45 · 21033 in / 2762 out tokens · 26266 ms · 2026-07-13T05:42:47.698173+00:00 · methodology

0 comments
read the original abstract

The choice of crystalline orientation and offcut angle is non-trivial for low-symmetry $\beta\text{-Ga}_2\text{O}_3$, where anisotropy impacts bulk and thin film synthesis, material properties, and power device fabrication and performance. Scalable (100)-oriented $\beta\text{-Ga}_2\text{O}_3$ wafers are desirable for electronic devices but are not typically used due to 10-30x slower growth rates compared to other orientations. Here we report molecular beam epitaxy (MBE) growth rates equal to the fast growth direction by using (100) $Ga_2O_3$ wafers with large grown-in offcuts. The offcuts (up to 13.4{\deg}) are directly grown by Edge-defined Film-fed Growth (EFG) of 2D ribbons with rotated seed crystals, avoiding material loss from crystal boule offcut methods while maintaining high crystalline quality. Chemical-mechanical polishing produces epitaxy-ready substrates, and step flow growth is observed across all offcut angles. We measure an unintentional n-type doping density of $2{\times}10^{15} cm^{-3}$, one of the lowest values reported for MBE-grown films. Planar Schottky barrier diodes on these epilayers without edge termination have an on/off ratio ~10$^5$ and an average breakdown field of 1.56 MV/cm, comparable to or exceeding similar devices fabricated on other orientations. Overall, these results illustrate the importance of both crystal face and offcut angle and validate the use of the scalable (100)-oriented $\beta\text{-Ga}_2\text{O}_3$ wafers.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

3 extracted references · 2 canonical work pages

  1. [1]

    World’s first successful epitaxial deposition of gallium oxide on a 6-inch wafer using the HVPE method | Novel Crystal Technology, Inc.,

    K. Koshi, Single Crystal Manufacturing Apparatus and Method, 2023, US11725299B2. [14] P . Mazzolini, A. Falkenstein, Z. Galazka, M. Martin, O. Bierwagen, Appl. Phys. Lett. 2020, 117, 222105. [15] R. Schewski, K. Lion, A. Fiedler, C. Wouters, A. Popp, S. V . Levchenko, T. Schulz, M. Schmidbauer, S. Bin Anooz, R. Grüneberg, Z. Galazka, G. Wagner, K. Irmsche...

    2023

  2. [2]

    Cr2O3/β-Ga2O3 Heterojunction Diodes with Orientation-Dependent Breakdown Electric Field up to 12.9 MV/cm,

    K. Sasaki, A. Kuramata, T. Masui, E. G. Víllora, K. Shimamura, S. Yamakoshi, Appl. Phys. Express 2012, 5, 035502. [48] D. B. Bayraktaroglu, Assessment of Ga2O3 Technology, Air Force Research Laboratory, 2016. [49] X. Xiu, L. Zhang, Y . Li, Z. Xiong, R. Zhang, Y . Zheng, J. Semicond. 2019, 40, 011805. [50] K. Goto, K. Konishi, H. Murakami, Y . Kumagai, B. ...

    work page doi:10.48550/arxiv.2512.11264 2012

  3. [3]

    H. Aida, K. Nishiguchi, H. Takeda, N. Aota, K. Sunakawa, Y . Yaguchi, Jpn. J. Appl. Phys. 2008, 47, 8506. [72] T.-S. Chou, J. Rehm, S. Bin Anooz, C. Wouters, O. Ernst, A. Akhtar, Z. Galazka, M. Albrecht, A. Fiedler, A. Popp, Appl. Phys. Lett. 2025, 126, 022101. [73] J. P . McCandless, C. A. Gorsak, V . Protasenko, D. G. Schlom, M. O. Thompson, H. G. Xing,...

    work page doi:10.1063/5.0238259 2008