journal article·2026-04

Modular hybrid machine learning and physics-based potentials for scalable modeling of Van der Waals heterostructures

Hekai Bu, Wenwu Jiang, Penghua Ying, Ting Liang, Zheyong Fan, Wengen Ouyang

Publisher page

Abstract

Accurately modeling the structural reconstruction and thermodynamic behavior of van der Waals (vdW) heterostructures remains a significant challenge due to the limitations of conventional force fields in capturing their complex mechanical, thermal, electronic, and tribological properties. To address these limitations, we develop a hybrid framework that combines single-layer machine-learned potential ($s$MLP) with physics-based anisotropic interlayer potential (ILP), effectively decoupling intralayer and interlayer interactions. This $s$MLP+ILP approach modularizes the modeling of vdW heterostructures like assembling LEGOs, reducing the required training configurations by at least an order of magnitude compared to the pure MLP approach, while retaining predictive accuracy and computational efficiency. We validate our framework by accurately reproducing the mechanical and thermal transport properties of graphite and bulk hexagonal boron nitride ($h$-BN), and by resolving intricate Moiré patterns in graphene/$h$-BN bilayer and graphene/graphene/$h$-BN trilayer heterostructures, achieving excellent agreement with experimental observations. Leveraging the developed $s$MLP+ILP approach, we reveal the stacking order-dependent formation of Moir{é} superlattice in trilayer graphene/$h$-BN/MoS$_2$ heterostructures, demonstrating its ability to accurately model large-scale vdW systems comprising hundreds of thousands of atoms with near $ab$ $initio$ precision. These findings demonstrate that the hybrid $s$MLP+ILP framework remarkably outperforms existing pure machine-learned or empirical potentials, offering a scalable and transferable solution for accurately and extensively modeling complex vdW materials across diverse applications, including sliding ferroelectricity, thermal management, resistive switching, and superlubric nanodevices.
Source: openaire · Quality: candidate unverified · Source record ↗

Research context

OpenAlex topics: Machine Learning in Materials Science · Advanced Electron Microscopy Techniques and Applications · Quantum many-body systems

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

References (100+)

Show source-provided references
1. Thermal fluctuations and effective bending stiffness of elastic thin sheets and graphene: a nonlinear analysis10.1016/j.jmps.2017.07.011 ↗
2. Gaussian approximation potentials: the accuracy of quantum mechanics, without the electrons10.1103/PhysRevLett.104.136403 ↗
3. MACE: higher order equivariant message passing neural networks for fast and accurate force fields10.52202/068431-0830 ↗
4. E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials10.1038/s41467-022-29939-5 ↗
5. Generalized neural-network representation of high-dimensional potential-energy surfaces10.1103/PhysRevLett.98.146401 ↗
6. Elasticity and constitution of the Earth’s interior10.1029/JZ057i002p00227 ↗
7. Elastic constants of compression-annealed pyrolytic graphite10.1063/1.1659428 ↗
8. A second-generation reactive Empirical bond order (REBO) potential energy expression for hydrocarbons10.1088/0953-8984/14/4/312 ↗
9. ReaxFF reactive force field for molecular dynamics simulations of hydrocarbon oxidation10.1021/jp709896w ↗
10. Elastic plate basis for the deformation and electron diffraction of twisted bilayer graphene on a substrate10.1103/PhysRevB.110.024116 ↗
11. Few-layer graphene under high pressure: Raman and X-ray diffraction studies10.1016/j.ssc.2012.10.002 ↗
12. Tuning the through-plane lattice thermal conductivity in van der Waals structures through rotational (dis)ordering10.1021/acsnano.3c09717 ↗
13. A discrete-to-continuum model of weakly interacting incommensurate two-dimensional lattices: the hexagonal case10.1016/j.jmps.2023.105229 ↗
14. Efficient molecular dynamics simulations with many-body potentials on graphics processing units10.1016/j.cpc.2017.05.003 ↗
15. Homogeneous nonequilibrium molecular dynamics method for heat transport and spectral decomposition with many-body potentials10.1103/PhysRevB.99.064308 ↗
16. GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations10.1063/5.0106617 ↗
17. Neuroevolution machine learning potentials: combining high accuracy and low cost in atomistic simulations and application to heat transport10.1103/PhysRevB.104.104309 ↗
18. Anisotropic interfacial force field for interfaces of water with hexagonal boron nitride10.1021/acs.langmuir.3c01612 ↗
19. Registry-dependent potential for interfaces of water with graphene10.1021/acs.jpcc.2c08464 ↗
20. Machine-learned potentials for next-generation matter simulations10.1038/s41563-020-0777-6 ↗
21. LMFF: Efficient and scalable layered materials force field on heterogeneous many-core processors
22. Anisotropic interlayer force fields for van der waals interfaces: development and applications10.1063/5.0273079 ↗
23. Spontaneous curvature in two-dimensional van der Waals heterostructures10.1038/s41467-025-56055-x ↗
24. A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu10.1063/1.3382344 ↗
25. Sub-μm c-axis structural domain size of graphene paper uncovered by low-momentum phonon scattering10.1016/j.carbon.2017.10.070 ↗
26. Graphite under pressure: equation of state and first-order Raman modes10.1103/PhysRevB.39.12598 ↗
27. Density functional model for van der Waals interactions: unifying many-body atomic approaches with nonlocal functionals10.1103/PhysRevLett.124.146401 ↗
28. Structural superlubricity and ultralow friction across the length scales10.1038/s41586-018-0704-z ↗
29. Probing fracture mechanics of graphene through heterocrack propagation in a moiré superlattice10.1016/j.jmps.2025.106151 ↗
30. The origin of Moiré-level stick-slip behavior on graphene/h-BN heterostructures10.1002/adfm.202204209 ↗
31. Imaging dual-Moiré lattices in twisted bilayer graphene aligned on hexagonal boron nitride using microwave impedance microscopy10.1021/acs.nanolett.1c00601 ↗
32. Thickness-dependent cross-plane thermal conductivity measurements of exfoliated hexagonal boron nitride10.1021/acsami.2c21306 ↗
33. Anisotropic thermal transport in bulk hexagonal boron nitride
34. Moiré-driven interfacial thermal transport in twisted transition metal dichalcogenides
35. Twist-dependent anisotropic thermal conductivity in homogeneous MoS2 stacks10.1016/j.ijheatmasstransfer.2023.124662 ↗
36. Anisotropic interlayer force field for interfaces of graphene and h-BN with transition metal dichalcogenides10.1021/acs.jpcc.4c05924 ↗
37. Anisotropic interlayer force field for group-VI transition metal dichalcogenides10.1021/acs.jpca.3c04540 ↗
38. Nucleation mechanism for the direct graphite-to-diamond phase transition10.1038/nmat3078 ↗
39. Extremely anisotropic van der Waals thermal conductors10.1038/s41586-021-03867-8 ↗
40. Registry-dependent interlayer potential for graphitic systems10.1103/PhysRevB.71.235415 ↗
41. Coherent commensurate electronic states at the interface between misoriented graphene layers10.1038/nnano.2016.85 ↗
42. Interlayer interaction and relative vibrations of bilayer graphene10.1039/c0cp02614j ↗
43. Measurement of the elastic properties and intrinsic strength of monolayer graphene10.1126/science.1157996 ↗
44. Inter-layer potential for hexagonal boron nitride10.1063/1.4867272 ↗
45. Multiwalled nanotube faceting unravelled10.1038/nnano.2016.151 ↗
46. Interlayer potential for graphene/h-BN heterostructures10.1021/acs.jctc.6b00147 ↗
47. Anisotropic interlayer force field for two-dimensional hydrogenated carbon materials and their heterostructures10.1021/acs.jpcc.3c03275 ↗
48. Probing the ideal limit of interfacial thermal conductance in two-dimensional van der Waals heterostructures10.1038/s41524-025-01885-y ↗
49. Optimized Tersoff and Brenner empirical potential parameters for lattice dynamics and phonon thermal transport in carbon nanotubes and graphene10.1103/PhysRevB.81.205441 ↗
50. Homogenization of two-dimensional materials integrating monolayer bending and surface layer effects10.1016/j.jmps.2024.105911 ↗
51. In situ twistronics: a new platform based on superlubricity
52. Intrinsic long-range bond-order potential for carbon: performance in Monte Carlo simulations of graphitization10.1103/PhysRevB.68.024107 ↗
53. Improved long-range reactive bond-order potential for carbon. I. construction
54. Effect of high pressure on the lattice parameters of diamond, graphite, and hexagonal boron nitride10.1063/1.1726442 ↗
55. Catalytic growth of ultralong graphene nanoribbons on insulating substrates10.1002/adma.202200956 ↗
56. Graphene nanoribbons grown in hBN stacks for high-performance electronics10.1038/s41586-024-07243-0 ↗
57. The princess and the nanoscale pea: long-range penetration of surface distortions into layered materials stacks10.1021/acsnano.9b00645 ↗
58. Raman evidence for pressure-induced formation of diamondene10.1038/s41467-017-00149-8 ↗
59. The compressibility of media under extreme pressures10.1073/pnas.30.9.244 ↗
60. Stronger interlayer interactions contribute to faster hot carrier cooling of bilayer graphene under pressure10.1103/PhysRevLett.126.027402 ↗
61. Ferroelectricity with concomitant Coulomb screening in van der Waals heterostructures10.1038/s41565-024-01846-4 ↗
62. Mechanical and tribological properties of layered materials under high pressure: assessing the importance of many-body dispersion effects10.1021/acs.jctc.9b00908 ↗
63. Parity-dependent moiré superlattices in graphene/h-BN heterostructures: a route to mechanomutable metamaterials10.1103/PhysRevLett.126.216101 ↗
64. Registry-dependent peeling of layered material interfaces: the case of graphene nanoribbons on hexagonal boron nitride10.1021/acsami.1c09529 ↗
65. Nanoserpents: graphene nanoribbon motion on two-dimensional hexagonal materials10.1021/acs.nanolett.8b02848 ↗
66. Controllable thermal conductivity in twisted homogeneous interfaces of graphene and hexagonal boron nitride10.1021/acs.nanolett.0c02983 ↗
67. Anisotropic interlayer force field for transition metal dichalcogenides: the case of molybdenum disulfide10.1021/acs.jctc.1c00782 ↗
68. Unconventional domain tessellations in moiré-of-moiré lattices10.1038/s41586-025-08932-0 ↗
69. Generalized gradient approximation made simple10.1103/PhysRevLett.77.3865 ↗
70. High-pressure Raman spectroscopy of graphene10.1103/PhysRevB.80.073408 ↗
71. A comprehensive assessment of empirical potentials for carbon materials10.1063/5.0052870 ↗
72. Potential energy surface for graphene on graphene: Ab initio derivation, analytical description, and microscopic interpretation
73. Interlayer shear behaviors of bilayer graphene with a moiré pattern10.1016/j.jmps.2025.106123 ↗
74. Development of a machine learning potential for graphene10.1103/PhysRevB.97.054303 ↗
75. Energy transport by radiation in hyperbolic material comparable to conduction10.1002/adfm.201905830 ↗
76. High dimensions and heavy tails for natural evolution strategies
77. Tuning the direct and indirect excitonic transitions of h -BN by hydrostatic pressure10.1021/acs.jpcc.1c02082 ↗
78. General-purpose machine-learned potential for 16 elemental metals and their alloys10.1038/s41467-024-54554-x ↗
79. Computer simulation of local order in condensed phases of silicon10.1103/PhysRevB.31.5262 ↗
80. A reactive potential for hydrocarbons with intermolecular interactions10.1063/1.481208 ↗
81. Torsional periodic lattice distortions and diffraction of twisted 2D materials10.1038/s41467-022-35477-x ↗
82. Tan, C. W., Descoteaux, M. L., Kotak, M., de Miranda, N. G., Kavanagh, S. R., Zichi, L., Wang, M., Saluja, A., Hu, Y. R., Smidt, T., Johansson, A., Witt, W. C., Kozinsky, B., Musaelian, A., 2025. High-performance training and inference for deep equivariant interatomic potentials. 2504.16068. 10.48550/arXiv.2504.16068.
83. Empirical interatomic potential for carbon, with applications to amorphous carbon10.1103/PhysRevLett.61.2879 ↗
84. LAMMPS-a flexible simulation tool for particle-based materials modeling at the atomic, meso, and continuum scales10.1016/j.cpc.2021.108171 ↗
85. Accurate molecular van der Waals interactions from ground-state electron density and free-atom reference data10.1103/PhysRevLett.102.073005 ↗
86. Temperature effects on the universal equation of state of solids10.1103/PhysRevB.35.1945 ↗
87. Interfacial ferroelectricity by van der Waals sliding10.1126/science.abe8177 ↗
88. Deducing the internal interfaces of twisted multilayer graphene via Moiré-regulated surface conductivity10.1093/nsr/nwad175 ↗
89. Kinetic friction of structurally superlubric 2d material interfaces10.1016/j.jmps.2023.105396 ↗
90. Hybrid neural network potential for multilayer graphene10.1103/PhysRevB.100.195419 ↗
91. The phonon dispersion of graphite revisited10.1016/j.ssc.2004.04.042 ↗
92. Facet-governed frictional behavior in graphene/h-BN heteronanotubes10.1016/j.eml.2024.102144 ↗
93. GPUMD 4.0: A high-performance molecular dynamics package for versatile materials simulations with machine-learned potentials
94. Scanning tunnelling microscopy and spectroscopy of ultra-flat graphene on hexagonal boron nitride10.1038/nmat2968 ↗
95. Peeling and sliding of graphene nanoribbons with periodic van der waals interactions10.1016/j.jmps.2021.104698 ↗
96. Shape-dependent friction scaling laws in twisted layered material interfaces10.1016/j.jmps.2024.105555 ↗
97. Moiré superlattice effects on interfacial mechanical behavior: a concise review
98. Origin of frictional scaling law in circular twist layered interfaces: simulations and theory10.1016/j.jmps.2022.105114 ↗
99. Ultrahigh thermal conductance across superlubric interfaces in twisted graphite10.1103/PhysRevLett.134.146302 ↗
100. Pressure-induced commensurate stacking of graphene on boron nitride10.1038/ncomms13168 ↗