Many-body benchmarking of DFT local-registry energetics in bilayer InSe

summary

The gist

Density functional theory (DFT) is widely used to model twisted bilayers, but the accuracy of the local stacking energetics underlying such models remains uncertain Here…

In short

The study used diffusion quantum Monte Carlo (DMC) to benchmark Density Functional Theory (DFT) predictions for local stacking energetics in bilayer InSe. DMC revealed that DFT incorrectly predicts near-degeneracy among AB, AAr, and ABr stackings. DMC shows much larger energy separations, indicating that the true stacking stability depends on the full many-body electronic response of the registry, not just atomic motifs.

Key concepts

Local Registry Landscape
This refers to how different atomic arrangements (like AB vs. AAr) at a specific interface affect the energy of a material. DFT often predicts this landscape is shallow and functional-dependent, meaning its predictions for stacking stability can be unreliable.
Diffusion Quantum Monte Carlo (DMC)
DMC is a high-accuracy computational method used to calculate the true ground state energies of materials by simulating quantum mechanical particles. It was used here to test and correct the less accurate DFT predictions regarding stacking energies.
Many-Body Hierarchy
This describes the actual, detailed energy ordering among different atomic stackings, which DMC reveals is much more complex than DFT suggests. The many-body hierarchy shows that certain arrangements (like AAr) are significantly more stable or less stable than DFT predicts.
Charge Redistribution
This analysis looks at how electrons move between the layers when the bilayer forms. The study found that AAr stabilizes by moving charge away from the interlayer gap, while ABr shows an opposite trend, providing a microscopic reason for its different energy stability.

Terminology used across episodes

This episode discusses

The paper

Many-body benchmarking of DFT local-registry energetics in bilayer InSe · Read on arXiv

Jeonghwan Ahn, Abdulgani Annaberdiyev, Jovan Nelson, Nathaniel P. Stern, Hyeondeok Shin

Anthony J Leggett Institute for Condensed Matter Theory, University of Illinois at Urbana-Champaign · Center for Nanophase Materials Sciences, Oak Ridge National Laboratory · Applied Physics Program, Northwestern University · Department of Physics and Astronomy, Northwestern University · Computational Science Division, Argonne National Laboratory

Density functional theory (DFT) is widely used to model twisted bilayers, but the accuracy of the local stacking energetics underlying such models remains uncertain. Here, we benchmark the local-registry landscape of bilayer InSe using diffusion quantum Monte Carlo (DMC). DFT predicts that AB, AAr, and ABr stackings, which share the same interfacial Se registry, are nearly degenerate within 1.5 meV/f.u. and exhibit nearly indistinguishable DFT charge-density responses. DMC instead separates these stackings by 8(5) and 41(4) meV/f.u., while the energy difference between the most stable and least stable registries reaches 60(7) meV/f.u.. These large energy separations show that the stacking energetics are not determined by the interfacial atomic motif alone but depend on the full registry and its associated many-body electronic response. More broadly, these results show that DFT-based moiré models can substantially underestimate local stacking-energy corrugation, with direct consequences for predicted structural relaxation, domain formation, and electronic reconstruction in twisted layered materials.

DOI: 10.1103/1mnm-51c5

Transcript

Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.

Kai: We're looking at the paper: "Many-body benchmarking of DFT local-registry energetics in bilayer InSe".

Mira: Here's the gist of the paper. Density functional theory (DFT) is widely used to model twisted bilayers, but the accuracy of the local stacking energetics underlying such models remains uncertain Here, we benchmark the localregistry landscape of bilayer InSe using diffusion quantum Monte Carlo (DMC).

Kai: Benchmarking DFT Predictions The study benchmarks the local-registry landscape of bilayer InSe using DMC to test DFT predictions.

Mira: DFT predicts that AB, AAr, and ABr stackings, which share the same interfacial Se registry, are nearly degenerate within 1.5 meV/f.u.

Kai: and exhibit nearly indistinguishable DFT charge-density responses.

Paper summary: Kai: So, to wrap up the "Many-body benchmarking of DFT local-registry energetics in bilayer InSe" paper by Ahn and colleagues. They used DMC to show that DFT underestimates the energy scale driving moiré relaxation in bilayer InSe >

Mira: The authors highlight that DFT predicts AB, AAr, and ABr stackings are nearly degenerate within one point five meV per formula unit, but DMC separates them significantly more by eight point five and forty one point four meV per formula unit <ref:2607.01682#pg1>. Lev What does this mean for the actual physics we’re trying to model in moiré systems? It means that the weak twist dependence DFT predicted shouldn't be interpreted as intrinsically weak registry dependence, but rather a result of a flattened local-registry landscape >

Kai: Exactly. The paper shows that DFT can artificially compress energy differences among locally similar registries <ref:2607.01682#pg3>. Mira So the big point is that reliable moiré modeling requires validating the local-registry hierarchy and its full energetic scale, not merely relying on relaxed geometry or what DFT identifies as the minimum >

Lev: From an engineering standpoint, if we're building systems that rely on moiré patterns, we need to incorporate this many-body physics because it dictates the actual energy landscape and how things will relax in a real device >

Conclusion: Kai: So, this paper, "Many-body benchmarking of DFT local-registry energetics in bilayer InSe," it’s basically using diffusion quantum Monte Carlo to test how well density functional theory predicts the stacking energies in these twisted bilayers >

Mira: Right. The authors are showing that DFT predicts AB, AAr, and ABr stackings are nearly the same energy within one point five meV per formula unit, but DMC shows they actually separate by eight point five and forty one point four meV per formula unit >

Lev: That separation is huge for a quantum error correction system because it means the true energy scale is much bigger than what standard DFT gives you >

Kai: So what does this mean for the whole moiré problem? It suggests that we can't just trust the geometry or which stacking DFT picks as its minimum anymore >

Mira: Exactly. The local registry, like AB versus AAr, isn't just about the atoms touching; it’s about how charge actually moves between the layers and stays in the structure >

Lev: From an engineering standpoint, if we design a moiré device based on DFT predictions alone, we’re missing a big part of the physics that governs how stable those patterns actually are >

Kai: It means that when DFT says there's this weak twist dependence, it’s probably not because the registry is inherently weak; it’s because the local-registry landscape itself is flattened >

Mira: That’s right. The authors argue that DFT can artificially compress those energy differences among similar-looking registries, which messes up our moiré models >

Lev: So what this paper really tells us is that for any real moiré simulation, you need to validate the full energetic scale of the local registry, not just the relaxed geometry or whatever minimum DFT finds >

More episodes

← Home