arxiv: 2604.06539 · v2 · submitted 2026-04-08 · ❄️ cond-mat.mtrl-sci · physics.chem-ph

Recognition: no theorem link

The effects of dispersion damping and three-body interactions for accurate layered-material exfoliation energies

Adrian F. Rumson , Kyle R. Bryenton , Erin R. Johnson

Authors on Pith no claims yet

Pith reviewed 2026-05-12 01:50 UTC · model grok-4.3

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

keywords layered materialsexfoliation energiesdispersion correctionsXDM modelthree-body interactionsdensity functional theoryLondon dispersionAxilrod-Teller-Muto term

0 comments

The pith

Adding three-body Axilrod-Teller-Muto terms to XDM dispersion corrections produces the most accurate exfoliation energies yet for layered materials using semi-local functionals.

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

The paper establishes that London dispersion must be modeled with proper short-range damping and explicit three-body contributions to obtain reliable exfoliation energies in materials such as graphite, boron nitride, and transition-metal dichalcogenides. It compares the established Becke-Johnson damping with a newer Z-damping function inside the XDM framework and demonstrates that both benefit from the addition of the Axilrod-Teller-Muto three-body term. These improvements bring semi-local density-functional results into closer agreement with random-phase-approximation reference values on the LM26 benchmark than any prior combination tested. The central object is the damped pairwise dispersion energy plus the ATM three-body correction, which together control the balance between interlayer binding and intralayer lattice constants.

Core claim

The central claim is that inclusion of three-body interactions via the Axilrod-Teller-Muto term further improves the computed exfoliation energies for both XDM(BJ) and XDM(Z), yielding the best performance achieved on LM26 using semi-local functionals to date, relative to reference data from the random-phase approximation.

What carries the argument

The XDM dispersion model with either Becke-Johnson or Z damping, augmented by the Axilrod-Teller-Muto three-body term, which supplies the non-additive correction needed for accurate interlayer energies.

If this is right

Lattice constants and exfoliation energies for graphite, hBN, PbO, and transition-metal dichalcogenides can be predicted more reliably with semi-local functionals.
Both BJ and Z damping functions become viable when paired with the ATM term, expanding the range of usable dispersion models.
Three-body contributions are required to reach the accuracy level previously associated only with higher-level methods.
The approach remains computationally inexpensive enough for routine screening of new layered compounds.

Where Pith is reading between the lines

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

The same corrections could improve binding energies in molecular crystals or organic solids where three-body dispersion is also significant.
If Z-damping proves more transferable than BJ damping across chemical environments, it may become the default choice in future XDM implementations.
The benchmark results suggest that explicit three-body terms should be tested in other dispersion-corrected DFT workflows for 2D heterostructures.

Load-bearing premise

The LM26 benchmark and its RPA reference values are accurate and representative enough that the observed error reductions reflect genuine physical improvement rather than benchmark-specific fitting.

What would settle it

New, independent experimental measurements of exfoliation energies for several LM26 materials that fall outside the current RPA reference range while the improved XDM+ATM calculations remain unchanged.

read the original abstract

Accurate predictions of exfoliation energies and lattice constants of layered materials hinge on a correct description of London dispersion physics. Modern a posteriori dispersion corrections in density-functional theory (DFT), such as the exchange-hole dipole moment (XDM) model, capture the proper asymptotic behaviour at long range while making use of damping functions to prevent unphysical divergence at short range. In the united-atom limit, the dispersion energy is damped to a finite, non-zero value by both the canonical Becke--Johnson (BJ) damping function and the new Z-damping function. XDM(BJ) has previously demonstrated exceptional accuracy for modelling layered materials, such as in the LM26 benchmark, which includes graphite, hexagonal boron nitride, lead(II) oxide, and transition-metal dichalcogenides. This work presents the first assessment of XDM(Z) on the same benchmark. We also show that inclusion of three-body interactions via the Axilrod--Teller--Muto (ATM) term further improves the computed exfoliation energies for both XDM(BJ) and XDM(Z), yielding the best performance achieved on LM26 using semi-local functionals to date, relative to reference data from the random-phase approximation.

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

XDM(Z) plus the ATM three-body term edges out prior XDM(BJ) on LM26 exfoliation energies versus RPA, but the gains look incremental and rest on the benchmark's representativeness.

read the letter

The paper's core result is that switching to the Z-damping function in XDM and then adding the Axilrod-Teller-Muto three-body term produces the lowest mean errors on the LM26 exfoliation energies among the semi-local DFT options tested. It is the first reported run of XDM(Z) on this set, and the three-body addition helps both damping variants without introducing any new fitted parameters in this study. That is useful incremental information for people who already use XDM in layered-material work. The comparison stays external to the RPA reference values, so there is no obvious circularity in the reported ordering. The authors also keep the focus narrow on exfoliation energies and lattice constants, which matches the practical need for these quantities in stability predictions. The methods appear to follow the established XDM protocol, and the stress-test note finds no internal contradictions with the asymptotic requirements of the damping functions. On the soft side, LM26 is still a modest collection of systems, and any improvement could partly reflect error cancellation that is particular to this set rather than a general physical fix. RPA references themselves carry their own uncertainties for some of these materials, and without the full error tables or per-system breakdowns it is hard to judge how uniform the gains are. The abstract mentions lattice constants but the headline claim centers on energies, so the paper's strength is narrower than the title might suggest. This is the kind of targeted benchmark extension that computational materials groups will want to see, even if it does not rewrite the dispersion-correction landscape. It is solid enough and grounded enough to go to referees rather than a desk reject; a serious review would mainly press on the statistical robustness of the error reductions and whether the LM26 set needs expansion or cross-validation against other references.

Referee Report

0 major / 3 minor

Summary. The manuscript evaluates the XDM dispersion correction using both the standard Becke-Johnson (BJ) damping and a newly proposed Z-damping function on the LM26 benchmark of layered materials (graphite, hBN, PbO, TMDs). It reports that adding the Axilrod-Teller-Muto (ATM) three-body term improves exfoliation energies for both damping schemes, with XDM(Z)+ATM delivering the lowest mean errors relative to RPA reference values among semi-local DFT functionals tested to date.

Significance. If the reported error reductions hold, the work supplies a zero-parameter route to better dispersion physics for layered materials by combining an improved damping function with explicit three-body interactions. The direct comparison to prior XDM(BJ) results on the identical LM26 set and the use of an external RPA benchmark provide independent support for the performance ordering. This is relevant for computational materials science where accurate exfoliation energies are needed for van der Waals heterostructures.

minor comments (3)

The abstract states that XDM(Z)+ATM yields the best performance on LM26, but the main text should include an explicit table (or expanded Table X) listing mean absolute errors, maximum errors, and standard deviations for all four combinations (XDM(BJ), XDM(BJ)+ATM, XDM(Z), XDM(Z)+ATM) together with the previous XDM(BJ) reference values for direct comparison.
Section 3 (or Methods): the functional form and short-range behavior of the new Z-damping function are only sketched; a short derivation or explicit equation showing how it differs from BJ damping in the united-atom limit would aid reproducibility.
Results section: while the ATM term is shown to reduce errors, the manuscript should report whether the improvement is uniform across all 26 materials or driven by a subset (e.g., the TMDs), to address possible benchmark-specific error cancellation.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of our manuscript, accurate summary of the key results, and recommendation for minor revision. We are pleased that the significance of combining the Z-damping function with the Axilrod-Teller-Muto three-body term for improved exfoliation energies on the LM26 benchmark was recognized.

Circularity Check

0 steps flagged

Minor self-citation of prior XDM work but benchmark assessment remains independent of fitted inputs

full rationale

The paper conducts an empirical benchmark of existing XDM dispersion models (with BJ and Z damping) plus the standard ATM three-body term against external RPA reference exfoliation energies on the LM26 set. No derivation chain reduces the reported error reductions to quantities defined or fitted within the present study; the central claim is a comparative performance ordering on held-out reference data. Prior XDM papers by the same group are cited for methodological context, but this constitutes normal self-citation of an independently validated model rather than a load-bearing justification that collapses the current results. The approach is zero-parameter with respect to the LM26 benchmark itself.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Only the abstract is available, so the ledger is necessarily incomplete. The central claim rests on the assumption that RPA provides reliable reference exfoliation energies and that the LM26 set is representative.

axioms (1)

domain assumption Random-phase approximation calculations supply sufficiently accurate reference exfoliation energies for the LM26 materials.
The abstract uses RPA data as the benchmark against which DFT results are judged.

pith-pipeline@v0.9.0 · 5523 in / 1166 out tokens · 51829 ms · 2026-05-12T01:50:45.937359+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Reference graph

Works this paper leans on

11 extracted references · 11 canonical work pages · 1 internal anchor

[1]

Chhowalla, D

1 M. Chhowalla, D. Jena and H. Zhang, Nat. Rev.Mater., 2016, 1, 16052. 2 W. Choi, N. Choudhary , G. H. Han, J. Park, D. Akinwande and Y. H. Lee, Mater. Today, 2017,20, 116–130. 3 S. Manzeli, D. Ovchinnikov, D. Pasquier, O. V. Yazyev and A. Kis, Nat. Rev.Mater., 2017,2, 1–15. 4 X.-F. Wang, H. Tian, Y. Liu, S. Shen, Z. Yan, N. Deng, Y. Yang and T.-L. Ren, A...

work page 2016
[2]

Anichini, W

11 C. Anichini, W. Czepa, D. Pakulski, A. Aliprandi, A. Ciesielski and P. Samorì, Chem. Soc. Rev., 2018,47, 4860–4908. 12 M. Li, J. Lu, Z. Chen and K. Amine, Adv. Mater., 2018,30, 1800561. 13 L. Li, Z. Hu, S. Zhao and S.-L. Chou, Chem. Sci., 2021,12, 15206–15218. 14 R. H. Savage, J. App. Phys., 1948,19, 1–10. 15 C. Lee, Q. Li, W. Kalb, X.-Z. Liu, H. Berge...

work page arXiv 2018
[3]

Björkman, J

37 T. Björkman, J. Chem. Phys., 2014,141, 074708. 38 S. A. Tawfik, T. Gould, C. Stampfl and M. J. Ford, Phys. Rev. Mater., 2018,2, 034005. 39 J. Ning, M. Kothakonda, J. W. Furness, A. D. Kaplan, S. Ehlert, J. G. Brandenburg, J. P. Perdew and J. Sun,Phys. Rev.B, 2022, 106, 075422. 40 A. Otero-de-la Roza, L. M. LeBlanc and E. R. Johnson, J. Phys. Chem. Lett...

work page 2014
[4]

Axilrod and E

43 B. Axilrod and E. Teller, J. Chem. Phys., 1943,11, 299–300. 44 Y. Muto, Proc. Phys. Math. Soc. Jpn, 1943, pp. 629–631. 45 A. Otero-de-la Roza and E. R. Johnson, J. Chem. Phys., 2013, 138, 054103. 46 A. Otero-de-la Roza, L. M. LeBlanc and E. R. Johnson, Phys. Chem. Chem. Phys., 2020,22, 8266–8276. 47 O. Anatole von Lilienfeld and A. Tkatchenko, J. Chem....

work page 1943
[5]

Roadmap on Advancements of the FHI-aims Software Package

51 J.-D. Chai and M. Head-Gordon, Phys. Chem. Chem. Phys., 2008,10, 6615–6620. 52 P. Giannozzi, O. Andreussi, T. Brumme, O. Bunau, M. B. Nardelli, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, M. Cococcioni et al., J Phys.: Condens. Matter, 2017,29, 465901. 53 P. Giannozzi, O. Baseggio, P. Bonfà, D. Brunato, R. Car, I. Carnimeo, C. Cavazzoni, S. De Giro...

work page internal anchor Pith review Pith/arXiv arXiv 2008
[6]

Adamo and V

56 C. Adamo and V. Barone, J. Chem. Phys., 1999,110, 6158–

work page 1999
[7]

Weigend and R

57 F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005, 7, 3297–3305. 58 M. e. Frisch, G. Trucks, H. B. Schlegel, G. Scuseria, M. Robb, J. Cheeseman, G. Scalmani, V. Barone, G. Petersson, H. Nakat- suji et al., Gaussian 16,

work page 2005
[8]

59 E. R. Johnson, R. M. Dickson and A. D. Becke, J. Chem. Phys., 2007,126, 184104. 60 A. Otero-de-la-Roza, K. R. Bryenton, F. Kannemann, E. R. Johnson, R. M. Dickson, H. Schmider and A. D. Becke, postg (release: XCDM(Z)), 2015,https://github.com/ aoterodelaroza/postg. 61 V. W.-z. Yu, F. Corsetti, A. García, W. P. Huhn, M. Jacquelin, W. Jia, B. Lange, L. L...

work page 2007
[9]

66 F. O. Kannemann and A. D. Becke, J. Chem. Theory Comput., 2010,6, 1081–1088. 67 A. Otero-de-la-Roza, refdata, 2015,https://github.com/ aoterodelaroza/refdata. 68 E. van Lenthe, E.-J. Baerends and J. G. Snijders, J. Chem. Phys., 1994,101, 9783–9792. 69 A. F. Rumson, pyxdm, 2026,https://github.com/ arumson0/pyxdm/. 70 A. J. A. Price, A. Otero-de-la-Roza ...

work page 2010
[10]

Dal Corso, Comput

72 A. Dal Corso, Comput. Mater. Sci., 2014,95, 337–350. 73 J. P. Perdew and A. Zunger, Phys. Rev.B, 1981,23,

work page 2014
[11]

74 J. P. Perdew and Y. Wang, Phys. Rev.B, 1992,45, 13244. 75 J. P. Perdew, A. Ruzsinszky , G. I. Csonka, O. A. Vydrov, G. E. Scuseria, L. A. Constantin, X. Zhou and K. Burke, Phys. Rev. Lett., 2008,100, 136406. 76 H. Rydberg, M. Dion, N. Jacobson, E. Schröder, P. Hyldgaard, S. Simak, D. C. Langreth and B. I. Lundqvist, Phys. Rev.Lett., 2003,91, 126402. 77...

work page 1992