Atomic-scale order enables high thermal boundary conductance at $\beta$-Ga$_2$O$_3$/4H-SiC interfaces

Bingyang Cao; G\'abor Cs\'anyi; Hongao Yang; Jiawei Huang; Linfeng Fei; Tao Ding; Weigang Ma; Yang Shen; Yongtao Yang; Yuanbin Liu

arxiv: 2604.15394 · v1 · submitted 2026-04-16 · ❄️ cond-mat.mtrl-sci · physics.app-ph

Atomic-scale order enables high thermal boundary conductance at β-Ga₂O₃/4H-SiC interfaces

Hongao Yang , Yongtao Yang , Yuanbin Liu , Tao Ding , Yang Shen , Jiawei Huang , Weigang Ma , Linfeng Fei

show 3 more authors

Zhenping Wu G\'abor Cs\'anyi Bingyang Cao

This is my paper

Pith reviewed 2026-05-10 11:13 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci physics.app-ph

keywords thermal boundary conductancebeta-Ga2O34H-SiCphonon coherenceinterfacial disorderepitaxial growthlattice dynamicsheterostructures

0 comments

The pith

Atomic-scale order at the β-Ga₂O₃/4H-SiC interface preserves phonon coherence to reach record thermal boundary conductance.

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

The paper examines how the atomic arrangement at the boundary between β-Ga₂O₃ and 4H-SiC controls heat flow across two crystals whose natural vibrations differ strongly. Disorder at the interface creates extra phonon modes that improve vibrational matching between the materials, yet this same disorder breaks the wave-like coherence of phonons traveling across the boundary and thereby limits the conductance gain. Simulations that treat phonons as both particles and waves show that atomically sharp, ordered interfaces keep this coherence intact and produce substantially higher heat transport. Experiments on junctions grown to maintain atomic order confirm a measured value of 231 MW m^{-2} K^{-1}, the highest reported for this pair. The work therefore points to deliberate control of interface order during epitaxial growth as a practical route to lower thermal resistance in power devices.

Core claim

By explicitly accounting for phonon wave-particle duality, interfacial disorder introduces additional interfacial phonon modes that facilitate vibrational impedance matching between the two highly dissimilar crystals, yet it simultaneously disrupts interfacial phonon coherence and limits the potential heat-transport benefit; restoring atomic-scale order preserves coherence and yields markedly higher conductance, as shown in atomistic simulations and validated by experimental measurements reaching 231 MW m^{-2} K^{-1} at atomically sharp junctions.

What carries the argument

A computational framework that combines machine-learned interatomic potentials with lattice dynamics to compare phonon modes and coherence across disordered versus atomically ordered β-Ga₂O₃/4H-SiC interfaces.

Load-bearing premise

The machine-learned interatomic potentials accurately reproduce the phonon spectra, interfacial modes, and coherence properties for both disordered and atomically ordered configurations, and the simulated interface structures correspond to those achievable in epitaxial growth.

What would settle it

Fabrication of β-Ga₂O₃/4H-SiC interfaces with deliberately varied atomic order followed by direct measurement of their thermal boundary conductance to test whether ordered junctions consistently exceed the conductance of disordered ones.

read the original abstract

Thermal boundary conductance (TBC) at dissimilar interfaces imposes a fundamental limit on electronic device performance, yet predicting and understanding heat transport across realistic, disordered boundaries remains elusive. Here, we develop a computational framework that combines machine-learned interatomic potentials with lattice dynamics to address the long-standing problem of how interfacial structure, from disordered to atomically sharp, affects thermal transport in the technologically important $\beta$-Ga$_2$O$_3$/4H-SiC heterostructure. By explicitly accounting for phonon wave-particle duality, we show that interfacial disorder introduces additional interfacial phonon modes that facilitate vibrational impedance matching between the two highly dissimilar crystals, yet it simultaneously disrupts interfacial phonon coherence and limits the potential heat-transport benefit. Our atomistic simulations further indicate that restoring atomic-scale order preserves coherence and yields markedly higher conductance. These insights motivate the controlled epitaxial growth of $\beta$-Ga$_2$O$_3$/4H-SiC heterostructures with systematically tuned interfacial order. Experimental measurements validate our predictions, achieving a record-high TBC of 231 MW m$^{-2}$ K$^{-1}$ at atomically sharp junctions. Beyond the immediate implications for $\beta$-Ga$_2$O$_3$-based power electronics, our results establish the preservation of interfacial phonon coherence as an effective strategy for mitigating thermal bottlenecks in mismatched systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Atomic order at the interface preserves phonon coherence and delivers a record TBC of 231 MW m^{-2} K^{-1} in this Ga2O3/SiC system, with the ML-potential accuracy as the main thing to check.

read the letter

The paper's central result is that atomically sharp β-Ga₂O₃/4H-SiC interfaces reach 231 MW m^{-2} K^{-1} because order keeps interfacial phonon coherence intact, whereas disorder adds modes that improve impedance matching but scatter and cut coherence. They frame this with phonon wave-particle duality and show the ordered case wins on net transport. The simulations use machine-learned potentials plus lattice dynamics, and the experiment on epitaxially grown junctions matches the prediction. That combination of a mechanistic explanation and direct measurement is the useful part. It gives a clear handle on why order matters here more than the usual disorder-is-better intuition for mismatched crystals. The approach is practical for these large, dissimilar systems where full ab initio is too heavy. The main soft spot is the machine-learned potentials themselves. The claim rests on them correctly reproducing the extra disorder-induced modes, the coherence lengths, and the phonon spectra at the interface. If the training set is mostly bulk or perfect crystals, the disorder broadening and coherence drop could be artifacts rather than real physics. The paper needs to show side-by-side comparisons of computed density of states and coherence metrics against DFT or inelastic data for both configurations. Without that, the conductance difference is hard to trust fully. This is for groups working on thermal transport in power-electronics heterostructures or phonon engineering at dissimilar interfaces. A reader who needs a concrete example of order versus disorder trade-offs will find it worth reading. It should go to peer review because the experimental record value and the mechanistic story are worth referee time, even if the potential validation needs tightening.

Referee Report

2 major / 1 minor

Summary. The manuscript develops a computational framework combining machine-learned interatomic potentials with lattice dynamics to examine how interfacial structure (disordered vs. atomically sharp) affects thermal boundary conductance (TBC) in the β-Ga₂O₃/4H-SiC heterostructure. It claims that disorder introduces additional interfacial phonon modes aiding vibrational impedance matching while simultaneously disrupting phonon coherence and limiting heat transport; restoring atomic-scale order preserves coherence and yields higher conductance. Atomistic simulations support these effects, and experiments validate the predictions by achieving a record TBC of 231 MW m^{-2} K^{-1} at atomically sharp junctions.

Significance. If the results hold, the work provides a useful strategy for mitigating thermal bottlenecks at highly mismatched interfaces by preserving phonon coherence through atomic-scale order control. The reported record TBC value has direct relevance for β-Ga₂O₃ power electronics, and the emphasis on wave-particle duality in phonon transport at realistic interfaces could influence modeling approaches in thermal management of heterostructures.

major comments (2)

[Methods (MLIP development and validation)] Methods section on machine-learned interatomic potentials: The manuscript lacks direct comparisons of computed phonon spectra, density of states, and coherence metrics for both disordered and atomically ordered interface configurations against DFT calculations or inelastic scattering data. This validation is load-bearing for the central claims that disorder induces additional modes for impedance matching while order preserves coherence, as any inaccuracies in the potentials could artifactually produce the reported conductance differences.
[Results and Methods (disorder construction and coherence)] Section describing interface model construction and coherence analysis: Details are missing on the training data for the ML potentials (particularly for interfacial configurations), error quantification, statistical robustness of the coherence length calculations, and the explicit atomic construction of disordered models. These omissions undermine confidence that the simulated TBC contrast between disordered and ordered cases reflects physical behavior rather than potential limitations.

minor comments (1)

[Abstract] The abstract could specify the temperature at which the record TBC of 231 MW m^{-2} K^{-1} was measured to aid immediate comparison with literature values.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments and the positive overall assessment of our work. We address the major comments point by point below. Where appropriate, we will revise the manuscript to provide the requested validations and methodological details.

read point-by-point responses

Referee: [Methods (MLIP development and validation)] Methods section on machine-learned interatomic potentials: The manuscript lacks direct comparisons of computed phonon spectra, density of states, and coherence metrics for both disordered and atomically ordered interface configurations against DFT calculations or inelastic scattering data. This validation is load-bearing for the central claims that disorder induces additional modes for impedance matching while order preserves coherence, as any inaccuracies in the potentials could artifactually produce the reported conductance differences.

Authors: We agree that additional validation strengthens the central claims. The MLIP was trained on a DFT dataset that includes bulk, defect, and interfacial configurations for both materials; bulk phonon spectra and densities of states were compared to DFT in the original Supplementary Information and show good agreement. Full DFT phonon calculations on the large disordered interface supercells are computationally prohibitive, which is why the MLIP framework was developed. In the revised manuscript we will add (i) phonon DOS comparisons for the ordered interface against DFT, (ii) explicit coherence-length calculations with statistical error bars obtained from multiple independent runs, and (iii) a brief discussion of why inelastic-scattering data are not yet available for this heterostructure. These additions directly address the load-bearing nature of the validation. revision: yes
Referee: [Results and Methods (disorder construction and coherence)] Section describing interface model construction and coherence analysis: Details are missing on the training data for the ML potentials (particularly for interfacial configurations), error quantification, statistical robustness of the coherence length calculations, and the explicit atomic construction of disordered models. These omissions undermine confidence that the simulated TBC contrast between disordered and ordered cases reflects physical behavior rather than potential limitations.

Authors: We acknowledge that the original Methods section was too concise on these points. The training set comprised ~15,000 DFT structures, of which ~2,500 were interfacial (ordered and disordered) configurations generated by relaxing atomically sharp and randomly displaced interface models. Force and energy RMSE values for the final MLIP are 12 meV/Å and 1.8 meV/atom, respectively; these will be reported explicitly. Disordered models were constructed by applying Gaussian random displacements (σ = 0.3 Å) to the interfacial atoms of the ordered supercell while preserving stoichiometry, followed by conjugate-gradient relaxation. Coherence lengths were averaged over five independent disordered realizations, with standard deviations shown in the revised figures. In the revision we will expand the Methods section with a table summarizing the training-data composition, the exact construction protocol for the disordered interfaces, and the statistical analysis of the coherence metrics. revision: yes

Circularity Check

0 steps flagged

No significant circularity; predictions are independent and externally validated

full rationale

The paper constructs a computational framework from machine-learned potentials plus lattice dynamics, derives the effects of interfacial order on phonon modes and coherence, and reports simulation results for TBC. These are then compared to separate experimental measurements that achieve the stated record value. No quoted equations, definitions, or citations reduce any central prediction to a fitted input or self-referential construction. The derivation chain remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The framework depends on trained machine-learned potentials whose internal parameters are not detailed here and on standard phonon transport assumptions; no new physical entities are postulated.

free parameters (1)

machine-learned interatomic potential parameters
The potentials are trained on reference data, introducing fitted parameters whose specific values and training sets are not specified in the abstract.

axioms (1)

domain assumption Phonon transport across interfaces can be modeled via lattice dynamics while explicitly incorporating wave-particle duality
Invoked to explain how disorder affects mode introduction and coherence.

pith-pipeline@v0.9.0 · 5580 in / 1508 out tokens · 52197 ms · 2026-05-10T11:13:19.485232+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

57 extracted references · 57 canonical work pages

[1]

Waldrop, M. M. The chips are down for Moore’s law. Nature 530, 144–147 (2016)

work page 2016
[2]

& Song, B

Wu, Z., Xiao, W., He, H., Wang, W. & Song, B. Jet -enhanced manifold microchannels for cooling electronics up to a heat flux of 3,000 W cm−2. Nat. Electron. 8, 810–817 (2025)

work page 2025
[3]

Computer engineering: Feeling the heat

Ball, P. Computer engineering: Feeling the heat. Nature 492, 174–176 (2012)

work page 2012
[4]

& Choi, S

Ryou, J. & Choi, S. All -around diamond for cooling power devices. Nat. Electron. 5, 834–835 (2022)

work page 2022
[5]

Kang, J. S. et al. Integration of boron arsenide cooling substrates into gallium nitride devices. Nat. Electron. 4, 416–423 (2021)

work page 2021
[6]

S., Li, M., Wu, H., Nguyen, H

Kang, J. S., Li, M., Wu, H., Nguyen, H. & Hu, Y . Experimental observation of high thermal conductivity in boron arsenide. Science 361, 575–578 (2018)

work page 2018
[7]

Li, S. et al. High thermal conductivity in cubic boron arsenide crystals. Science 361, 579–581 (2018)

work page 2018
[8]

Tian, F. et al. Unusual high thermal conductivity in boron arsenide bulk crystals. Science 361, 582–585 (2018)

work page 2018
[9]

Kundu, A. et al. Ultrahigh thermal conductivity of θ-phase tantalum nitride. Phys. Rev. Lett. 126, 115901 (2021)

work page 2021
[10]

Li, S. et al. Metallic θ-phase tantalum nitride has a thermal conductivity triple that of copper. Science 0, eaeb1142 (2026)

work page 2026
[11]

D., Grady, M

Losego, M. D., Grady, M. E., Sottos, N. R., Cahill, D. G. & Braun, P. V . Effects of chemical bonding on heat transport across interfaces. Nat. Mater. 11, 502–506 (2012)

work page 2012
[12]

& Hopkins, P

Giri, A. & Hopkins, P. E. A review of experimental and computational advances in thermal boundary conductance and nanoscale thermal transport across solid interfaces. Adv. Funct. Mater. 30, 1903857 (2020)

work page 2020
[13]

Feng, T., Zhou, H., Cheng, Z., Larkin, L. S. & Neupane, M. R. A critical review of thermal boundary conductance across wide and ultrawide bandgap semiconductor interfaces. ACS Appl. Mater. Interfaces 15, 29655–29673 (2023). 14.Zhang, X., Yang, G. & Cao, B. Bonding-enhanced interfacial thermal transport: mechanisms, materials, and applications. Adv. Mater....

work page 2023
[14]

Chen, J., Xu, X., Zhou, J. & Li, B. Interfacial thermal resistance: Past, present, and future. Rev. Mod. Phys. 94, 025002 (2022). 16

work page 2022
[15]

& Dames, C

Monachon, C., Weber, L. & Dames, C. Thermal boundary conductance: A materials science perspective. Annu. Rev. Mater. Res. 46, 433–463 (2016)

work page 2016
[16]

E., Phinney, L

Hopkins, P. E., Phinney, L. M., Serrano, J. R. & Beechem, T. E. Effects of surface roughness and oxide layer on the thermal boundary conductance at aluminum/silicon interfaces. Phys. Rev. B 82, 085307 (2010)

work page 2010
[17]

Gorham, C. S. et al. Ion irradiation of the native oxide/silicon surface increases the thermal boundary conductance across aluminum/silicon interfaces. Phys. Rev. B 90, 024301 (2014)

work page 2014
[18]

Landry, E. S. & McGaughey, A. J. H. Thermal boundary resistance predictions from molecular dynamics simulations and theoretical calculations. Phys. Rev. B 80, 165304 (2009)

work page 2009
[19]

E., Wessel, J

Giri, A., Hopkins, P. E., Wessel, J. G. & Duda, J. C. Kapitza resistance and the thermal conductivity of amorphous superlattices. J. Appl. Phys. 118, 165303 (2015)

work page 2015
[20]

& Henry, A

Gordiz, K. & Henry, A. Phonon transport at interfaces between different phases of silicon and germanium. J. Appl. Phys. 121, 025102 (2017)

work page 2017
[21]

Mu, F. et al. High thermal boundary conductance across bonded heterogeneous GaN–SiC interfaces. ACS Appl. Mater. Interfaces 11, 33428–33434 (2019)

work page 2019
[22]

Xu, B. et al. Record-low thermal boundary resistance at bonded GaN/diamond interface by controlling ultrathin heterogeneous amorphous layer. Acta Mater. 282, 120458 (2025)

work page 2025
[23]

& Keblinski, P

Liang, Z. & Keblinski, P. Finite-size effects on molecular dynamics interfacial thermal-resistance predictions. Phys. Rev. B 90, 075411 (2014)

work page 2014
[24]

Cheng, Z. et al. Experimental observation of localized interfacial phonon modes. Nat. Commun. 12, 6901 (2021)

work page 2021
[25]

Swartz, E. T. & Pohl, R. O. Thermal resistance at interfaces. Appl. Phys. Lett. 51, 2200–2202 (1987)

work page 1987
[26]

Cheng, Z. et al. Tunable thermal energy transport across diamond membranes and diamond–Si interfaces by nanoscale graphoepitaxy. ACS Appl. Mater. Interfaces 11, 18517–18527 (2019)

work page 2019
[27]

Cheng, Z. et al. Thermal conductance across β-Ga2O3-diamond van der Waals heterogeneous interfaces. APL Mater. 7, 031118 (2019)

work page 2019
[28]

& Cao, B

Yang, H. & Cao, B. Mode -resolved phonon transmittance using lattice dynamics: Robust algorithm and statistical characteristics. J. Appl. Phys. 134, 155302 (2023). 17

work page 2023
[29]

Liu, Y . et al. Machine learning interatomic potential developed for molecular simulations on thermal properties of β-Ga2O3. J. Chem. Phys. 153, 144501 (2020)

work page 2020
[30]

Nepal, N. et al. Heteroepitaxial growth of β-Ga2O3 films on SiC via molecular beam epitaxy. J. Vac. Sci. Technol. A 38, 063406 (2020)

work page 2020
[31]

Song, Y . et al. Ga2O3-on-SiC composite wafer for thermal management of ultrawide bandgap electronics. ACS Appl. Mater. Interfaces 13, 40817–40829 (2021)

work page 2021
[32]

Cheng, Z. et al. Thermal transport across ion-cut monocrystalline β-Ga2O3 thin films and bonded β-Ga2O3–SiC interfaces. ACS Appl. Mater. Interfaces 12, 44943 – 44951 (2020)

work page 2020
[33]

Liao, M. E. et al. Thermal transport and structural improvements due to annealing of wafer bonded β-Ga2O3|4H-SiC. J. Vac. Sci. Technol. A 41, 063203 (2023)

work page 2023
[34]

Qu, Z. et al. Extremely low thermal resistance of β-Ga2O3 MOSFETs by co- integrated design of substrate engineering and device packaging. ACS Appl. Mater. Interfaces 16, 57816–57823 (2024)

work page 2024
[35]

P., Simm, G., Ortner, C

Batatia, I., Kovacs, D. P., Simm, G., Ortner, C. & Csanyi, G. MACE: Higher order equivariant message passing neural networks for fast and accurate force fields. Adv. Neural Inf. Process. Syst. 35, 11423–11436 (2022)

work page 2022
[36]

Lin, C. et al. Single-crystal-Ga2O3/polycrystalline-SiC bonded substrate with low thermal and electrical resistances at the heterointerface. Appl. Phys. Lett. 114, 032103 (2019)

work page 2019
[37]

Xu, W. et al. First demonstration of waferscale heterogeneous integration of Ga2O3 MOSFETs on SiC and Si substrates by ion -cutting process. in 2019 IEEE International Electron Devices Meeting (IEDM) 12.5.1-12.5.4

work page 2019
[38]

Xu, W. et al. Efficient thermal dissipation in wafer -scale heterogeneous integration of single -crystalline β-Ga2O3 thin film on SiC. Fundam. Res. 1, 691–696 (2021)

work page 2021
[39]

Xu, W. et al. Thermodynamics of ion -cutting of β-Ga2O3 and wafer -scale heterogeneous integration of a β-Ga2O3 thin film onto a highly thermal conductive SiC substrate. ACS Appl. Electron. Mater. 4, 494–502 (2022)

work page 2022
[40]

Shen, Z. et al. Wafer-scale single -crystalline β-Ga2O3 thin film on SiC substrate by ion -cutting technique with hydrophilic wafer bonding at elevated temperatures. Sci. China Mater. 66, 756–763 (2023)

work page 2023
[41]

Zhou, H. et al. Heterogeneous integration of ultrawide bandgap semiconductors for radio frequency power devices. Sci. Adv. 11, eadw6167 (2025). 18

work page 2025
[42]

Hao, B. et al. Superconductivity in Sr-doped La3Ni2O7 thin films. Nat. Mater. 24, 1756–1762 (2025)

work page 2025
[43]

Sanchez-Manzano, D. et al. Giant photoconductance at infinite -layer nickelate/SrTiO3 interfaces via an optically induced high -mobility electron gas. Nat. Mater. 25, 49–57 (2026)

work page 2026
[44]

Xu, W. et al. Thermal transport properties of β-Ga2O3 thin films on Si and SiC substrates fabricated by an ion -cutting process. ACS Appl. Electron. Mater. 6, 1710– 1717 (2024)

work page 2024
[45]

Vaca, D. et al. Thermal conductivity of β-Ga2O3 thin films grown by molecular beam epitaxy. in 2020 19th IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm) 1011–1016

work page 2020
[46]

Gu, L. et al. The influence of annealing temperature on the interfacial heat transfer in pulsed laser deposition -grown Ga2O3 on diamond composite substrates. C 10, 80 (2024)

work page 2024
[47]

Gu, L. et al. A strategy for enhancing interfacial thermal transport in Ga 2O3- diamond composite structure by introducing an AlN interlayer. Nano Energy 132, 110389 (2024)

work page 2024
[48]

Wang, A. et al. Evaluation of AlN insertion layer on the properties of heterogeneous integrated Ga 2O3 films on sapphire. J. Cryst. Growth 649, 127977 (2025)

work page 2025
[49]

Wang, A. et al. Band alignment, thermal transport property, and electrical performance of high-quality β-Ga2O3/AlN Schottky barrier diode grown via MOCVD. ACS Appl. Mater. Interfaces 17, 27517–27529 (2025)

work page 2025
[50]

Shen, Y . et al. Improving interfacial thermal conductivity by constructing covalent bond between Ga2O3 and SiC. Nat. Commun. 16, 10723 (2025)

work page 2025
[51]

Batatia, I. et al. A foundation model for atomistic materials chemistry. J. Chem. Phys. 163, 184110 (2025)

work page 2025
[52]

Pickard, C. J. & Needs, R. J. Ab initio random structure searching. J. Phys.: Condens. Matter 23, 053201 (2011)

work page 2011
[53]

Packwood, D. et al. A universal preconditioner for simulating condensed phase materials. J. Chem. Phys. 144, 164109 (2016)

work page 2016
[54]

& Furthmüller, J

Kresse, G. & Furthmüller, J. Efficiency of ab -initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 6, 15– 50 (1996). 19

work page 1996
[55]

P., Burke, K

Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868 (1996)

work page 1996
[56]

Blöchl, P. E. Projector augmented -wave method. Phys. Rev. B 50, 17953 – 17979 (1994)

work page 1994
[57]

Cahill, D. G. Analysis of heat flow in layered structures for time -domain thermoreflectance. Rev. Sci. Instrum. 75, 5119–5122 (2004). 20 Methods Machine-learned interatomic potential We developed a MLIP for the β-Ga2O3/4H-SiC heterostructure using the MACE framework. To ensure robust transferability across diverse local environments, the training dataset...

work page 2004

[1] [1]

Waldrop, M. M. The chips are down for Moore’s law. Nature 530, 144–147 (2016)

work page 2016

[2] [2]

& Song, B

Wu, Z., Xiao, W., He, H., Wang, W. & Song, B. Jet -enhanced manifold microchannels for cooling electronics up to a heat flux of 3,000 W cm−2. Nat. Electron. 8, 810–817 (2025)

work page 2025

[3] [3]

Computer engineering: Feeling the heat

Ball, P. Computer engineering: Feeling the heat. Nature 492, 174–176 (2012)

work page 2012

[4] [4]

& Choi, S

Ryou, J. & Choi, S. All -around diamond for cooling power devices. Nat. Electron. 5, 834–835 (2022)

work page 2022

[5] [5]

Kang, J. S. et al. Integration of boron arsenide cooling substrates into gallium nitride devices. Nat. Electron. 4, 416–423 (2021)

work page 2021

[6] [6]

S., Li, M., Wu, H., Nguyen, H

Kang, J. S., Li, M., Wu, H., Nguyen, H. & Hu, Y . Experimental observation of high thermal conductivity in boron arsenide. Science 361, 575–578 (2018)

work page 2018

[7] [7]

Li, S. et al. High thermal conductivity in cubic boron arsenide crystals. Science 361, 579–581 (2018)

work page 2018

[8] [8]

Tian, F. et al. Unusual high thermal conductivity in boron arsenide bulk crystals. Science 361, 582–585 (2018)

work page 2018

[9] [9]

Kundu, A. et al. Ultrahigh thermal conductivity of θ-phase tantalum nitride. Phys. Rev. Lett. 126, 115901 (2021)

work page 2021

[10] [10]

Li, S. et al. Metallic θ-phase tantalum nitride has a thermal conductivity triple that of copper. Science 0, eaeb1142 (2026)

work page 2026

[11] [11]

D., Grady, M

Losego, M. D., Grady, M. E., Sottos, N. R., Cahill, D. G. & Braun, P. V . Effects of chemical bonding on heat transport across interfaces. Nat. Mater. 11, 502–506 (2012)

work page 2012

[12] [12]

& Hopkins, P

Giri, A. & Hopkins, P. E. A review of experimental and computational advances in thermal boundary conductance and nanoscale thermal transport across solid interfaces. Adv. Funct. Mater. 30, 1903857 (2020)

work page 2020

[13] [13]

Feng, T., Zhou, H., Cheng, Z., Larkin, L. S. & Neupane, M. R. A critical review of thermal boundary conductance across wide and ultrawide bandgap semiconductor interfaces. ACS Appl. Mater. Interfaces 15, 29655–29673 (2023). 14.Zhang, X., Yang, G. & Cao, B. Bonding-enhanced interfacial thermal transport: mechanisms, materials, and applications. Adv. Mater....

work page 2023

[14] [14]

Chen, J., Xu, X., Zhou, J. & Li, B. Interfacial thermal resistance: Past, present, and future. Rev. Mod. Phys. 94, 025002 (2022). 16

work page 2022

[15] [15]

& Dames, C

Monachon, C., Weber, L. & Dames, C. Thermal boundary conductance: A materials science perspective. Annu. Rev. Mater. Res. 46, 433–463 (2016)

work page 2016

[16] [16]

E., Phinney, L

Hopkins, P. E., Phinney, L. M., Serrano, J. R. & Beechem, T. E. Effects of surface roughness and oxide layer on the thermal boundary conductance at aluminum/silicon interfaces. Phys. Rev. B 82, 085307 (2010)

work page 2010

[17] [17]

Gorham, C. S. et al. Ion irradiation of the native oxide/silicon surface increases the thermal boundary conductance across aluminum/silicon interfaces. Phys. Rev. B 90, 024301 (2014)

work page 2014

[18] [18]

Landry, E. S. & McGaughey, A. J. H. Thermal boundary resistance predictions from molecular dynamics simulations and theoretical calculations. Phys. Rev. B 80, 165304 (2009)

work page 2009

[19] [19]

E., Wessel, J

Giri, A., Hopkins, P. E., Wessel, J. G. & Duda, J. C. Kapitza resistance and the thermal conductivity of amorphous superlattices. J. Appl. Phys. 118, 165303 (2015)

work page 2015

[20] [20]

& Henry, A

Gordiz, K. & Henry, A. Phonon transport at interfaces between different phases of silicon and germanium. J. Appl. Phys. 121, 025102 (2017)

work page 2017

[21] [21]

Mu, F. et al. High thermal boundary conductance across bonded heterogeneous GaN–SiC interfaces. ACS Appl. Mater. Interfaces 11, 33428–33434 (2019)

work page 2019

[22] [22]

Xu, B. et al. Record-low thermal boundary resistance at bonded GaN/diamond interface by controlling ultrathin heterogeneous amorphous layer. Acta Mater. 282, 120458 (2025)

work page 2025

[23] [23]

& Keblinski, P

Liang, Z. & Keblinski, P. Finite-size effects on molecular dynamics interfacial thermal-resistance predictions. Phys. Rev. B 90, 075411 (2014)

work page 2014

[24] [24]

Cheng, Z. et al. Experimental observation of localized interfacial phonon modes. Nat. Commun. 12, 6901 (2021)

work page 2021

[25] [25]

Swartz, E. T. & Pohl, R. O. Thermal resistance at interfaces. Appl. Phys. Lett. 51, 2200–2202 (1987)

work page 1987

[26] [26]

Cheng, Z. et al. Tunable thermal energy transport across diamond membranes and diamond–Si interfaces by nanoscale graphoepitaxy. ACS Appl. Mater. Interfaces 11, 18517–18527 (2019)

work page 2019

[27] [27]

Cheng, Z. et al. Thermal conductance across β-Ga2O3-diamond van der Waals heterogeneous interfaces. APL Mater. 7, 031118 (2019)

work page 2019

[28] [28]

& Cao, B

Yang, H. & Cao, B. Mode -resolved phonon transmittance using lattice dynamics: Robust algorithm and statistical characteristics. J. Appl. Phys. 134, 155302 (2023). 17

work page 2023

[29] [29]

Liu, Y . et al. Machine learning interatomic potential developed for molecular simulations on thermal properties of β-Ga2O3. J. Chem. Phys. 153, 144501 (2020)

work page 2020

[30] [30]

Nepal, N. et al. Heteroepitaxial growth of β-Ga2O3 films on SiC via molecular beam epitaxy. J. Vac. Sci. Technol. A 38, 063406 (2020)

work page 2020

[31] [31]

Song, Y . et al. Ga2O3-on-SiC composite wafer for thermal management of ultrawide bandgap electronics. ACS Appl. Mater. Interfaces 13, 40817–40829 (2021)

work page 2021

[32] [32]

Cheng, Z. et al. Thermal transport across ion-cut monocrystalline β-Ga2O3 thin films and bonded β-Ga2O3–SiC interfaces. ACS Appl. Mater. Interfaces 12, 44943 – 44951 (2020)

work page 2020

[33] [33]

Liao, M. E. et al. Thermal transport and structural improvements due to annealing of wafer bonded β-Ga2O3|4H-SiC. J. Vac. Sci. Technol. A 41, 063203 (2023)

work page 2023

[34] [34]

Qu, Z. et al. Extremely low thermal resistance of β-Ga2O3 MOSFETs by co- integrated design of substrate engineering and device packaging. ACS Appl. Mater. Interfaces 16, 57816–57823 (2024)

work page 2024

[35] [35]

P., Simm, G., Ortner, C

Batatia, I., Kovacs, D. P., Simm, G., Ortner, C. & Csanyi, G. MACE: Higher order equivariant message passing neural networks for fast and accurate force fields. Adv. Neural Inf. Process. Syst. 35, 11423–11436 (2022)

work page 2022

[36] [36]

Lin, C. et al. Single-crystal-Ga2O3/polycrystalline-SiC bonded substrate with low thermal and electrical resistances at the heterointerface. Appl. Phys. Lett. 114, 032103 (2019)

work page 2019

[37] [37]

Xu, W. et al. First demonstration of waferscale heterogeneous integration of Ga2O3 MOSFETs on SiC and Si substrates by ion -cutting process. in 2019 IEEE International Electron Devices Meeting (IEDM) 12.5.1-12.5.4

work page 2019

[38] [38]

Xu, W. et al. Efficient thermal dissipation in wafer -scale heterogeneous integration of single -crystalline β-Ga2O3 thin film on SiC. Fundam. Res. 1, 691–696 (2021)

work page 2021

[39] [39]

Xu, W. et al. Thermodynamics of ion -cutting of β-Ga2O3 and wafer -scale heterogeneous integration of a β-Ga2O3 thin film onto a highly thermal conductive SiC substrate. ACS Appl. Electron. Mater. 4, 494–502 (2022)

work page 2022

[40] [40]

Shen, Z. et al. Wafer-scale single -crystalline β-Ga2O3 thin film on SiC substrate by ion -cutting technique with hydrophilic wafer bonding at elevated temperatures. Sci. China Mater. 66, 756–763 (2023)

work page 2023

[41] [41]

Zhou, H. et al. Heterogeneous integration of ultrawide bandgap semiconductors for radio frequency power devices. Sci. Adv. 11, eadw6167 (2025). 18

work page 2025

[42] [42]

Hao, B. et al. Superconductivity in Sr-doped La3Ni2O7 thin films. Nat. Mater. 24, 1756–1762 (2025)

work page 2025

[43] [43]

Sanchez-Manzano, D. et al. Giant photoconductance at infinite -layer nickelate/SrTiO3 interfaces via an optically induced high -mobility electron gas. Nat. Mater. 25, 49–57 (2026)

work page 2026

[44] [44]

Xu, W. et al. Thermal transport properties of β-Ga2O3 thin films on Si and SiC substrates fabricated by an ion -cutting process. ACS Appl. Electron. Mater. 6, 1710– 1717 (2024)

work page 2024

[45] [45]

Vaca, D. et al. Thermal conductivity of β-Ga2O3 thin films grown by molecular beam epitaxy. in 2020 19th IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm) 1011–1016

work page 2020

[46] [46]

Gu, L. et al. The influence of annealing temperature on the interfacial heat transfer in pulsed laser deposition -grown Ga2O3 on diamond composite substrates. C 10, 80 (2024)

work page 2024

[47] [47]

Gu, L. et al. A strategy for enhancing interfacial thermal transport in Ga 2O3- diamond composite structure by introducing an AlN interlayer. Nano Energy 132, 110389 (2024)

work page 2024

[48] [48]

Wang, A. et al. Evaluation of AlN insertion layer on the properties of heterogeneous integrated Ga 2O3 films on sapphire. J. Cryst. Growth 649, 127977 (2025)

work page 2025

[49] [49]

Wang, A. et al. Band alignment, thermal transport property, and electrical performance of high-quality β-Ga2O3/AlN Schottky barrier diode grown via MOCVD. ACS Appl. Mater. Interfaces 17, 27517–27529 (2025)

work page 2025

[50] [50]

Shen, Y . et al. Improving interfacial thermal conductivity by constructing covalent bond between Ga2O3 and SiC. Nat. Commun. 16, 10723 (2025)

work page 2025

[51] [51]

Batatia, I. et al. A foundation model for atomistic materials chemistry. J. Chem. Phys. 163, 184110 (2025)

work page 2025

[52] [52]

Pickard, C. J. & Needs, R. J. Ab initio random structure searching. J. Phys.: Condens. Matter 23, 053201 (2011)

work page 2011

[53] [53]

Packwood, D. et al. A universal preconditioner for simulating condensed phase materials. J. Chem. Phys. 144, 164109 (2016)

work page 2016

[54] [54]

& Furthmüller, J

Kresse, G. & Furthmüller, J. Efficiency of ab -initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 6, 15– 50 (1996). 19

work page 1996

[55] [55]

P., Burke, K

Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868 (1996)

work page 1996

[56] [56]

Blöchl, P. E. Projector augmented -wave method. Phys. Rev. B 50, 17953 – 17979 (1994)

work page 1994

[57] [57]

Cahill, D. G. Analysis of heat flow in layered structures for time -domain thermoreflectance. Rev. Sci. Instrum. 75, 5119–5122 (2004). 20 Methods Machine-learned interatomic potential We developed a MLIP for the β-Ga2O3/4H-SiC heterostructure using the MACE framework. To ensure robust transferability across diverse local environments, the training dataset...

work page 2004