Pith. sign in

REVIEW 3 major objections 5 minor 46 references

Machine learning Landau free energy potentials

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A 14-parameter Landau polynomial reproduces PbTiO3's thermodynamics and predicts a clamped-crystal rhombohedral transition near 450 K.

desk verdict A solid methods paper: the eta=0 clamped-strain prediction is genuinely out-of-sample and the model selection scheme is clever, but the in-sample validation and lack of error bars keep the predictive claim modest. read the letter →

arxiv 2507.23369 v1 pith:KRBH4ORV submitted 2025-07-31 cond-mat.mtrl-sci

Mauro Pulzone , Natalya S. Fedorova , Hugo Aramberri , Jorge Íñiguez-González This is my paper
classification cond-mat.mtrl-sci
keywords LandaufreeenergymachinelearningferroelectricPbTiO3second-principlesmodelstrain-polarizationcouplingphasetransitionselection
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper establishes that a machine-learning scheme can build a simple Landau-like free energy polynomial directly from thermodynamic data—temperature-dependent polarization and strain, with and without applied electric fields and stresses—rather than from microscopic information. Applied to the ferroelectric perovskite PbTiO3, the method selects a minimal potential, called F7, in which only two coefficients depend linearly on temperature, and this potential quantitatively reproduces the Monte Carlo thermodynamics of the underlying atomistic model. Because the training data are of a kind that could be collected in the laboratory, the same procedure offers a route to predictive Landau models for ferroelectrics and related materials. The paper argues these 'third-principles' potentials can underpin macroscopic and mesoscopic simulations.

What carries the argument

The load-bearing object is the polynomial free energy F(P, eta; E, $\sigma$, T), written as a sum of symmetry-invariant monomials in polarization and strain with coefficients that may depend on temperature. Because the potential is linear in these coefficients, the parameters of any candidate model are obtained exactly and cheaply from the normal equations M phi = y applied to the equilibrium conditions grad_P F = 0 and grad_eta F = 0. The selection step is carried by a second, predictive error e: for every training state the fitted potential is minimized from the recorded state under the recorded external fields, and e measures the largest difference between predicted and recorded polarization and strain. This validation step rejects overcomplex models with spurious minima and flat regions that a purely fitting error would accept.

What would settle it

Train the same algorithm on experimental thermodynamic data for bulk PbTiO3, or on an independent atomistic model, and compare the resulting potential's prediction for the clamped eta = 0 crystal against direct simulation or experiment; if the predicted transition is not continuous, does not occur near 450 K, or has the wrong polar symmetry, the central extrapolation claim is refuted.

Watch

Extended reading notes

Core claim

The central claim is that a Landau model with a small number of physically motivated, symmetry-invariant monomials can be learned automatically and is predictive outside its training range. For PbTiO3, the optimal strain-dependent model F7 has 14 parameters, with temperature entering only through the quadratic coefficient A_{2i} and the elastic coefficient C_1; it reproduces the second-principles Monte Carlo polarization and strain almost perfectly. The model also extrapolates to the clamped-crystal condition eta = 0, where it predicts a continuous transition near 450 K to a rhombohedral ferroelectric phase, in agreement with direct simulations, even though no training state came close to eta = 0. This is taken as evidence that the polarization–strain couplings controlling the symmetry and order of PbTiO3's transition are encoded in the simple polynomial.

Load-bearing premise

The entire training set comes from one second-principles atomistic model of PbTiO3, so the claimed predictive power for the real material stands or falls with that model being a faithful representation of PbTiO3; the paper's only out-of-sample check is generated by the same model.

Editorial extensions

If this is right

  • The same fitting recipe can be run on experimental measurements of polarization and strain as functions of temperature and applied fields, since nothing in the method requires information about the microscopic Hamiltonian.
  • The selected F7 model, or close variants F8 and F9, gives quantitative predictions for PbTiO3's response to electric fields and stresses within the training range, and captures the weakly first-order character of the bulk transition.
  • Under the clamped eta = 0 condition, the model predicts a continuous cubic-to-rhombohedral transition near 450 K, matching dedicated simulations; this is an out-of-sample prediction.
  • The demonstrated success provides a route to 'third-principles' Ginzburg-Landau and dynamical mesoscale simulations, with gradient terms and inertial or damping constants supplied by atomistic models or experiment.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If this selection strategy generalizes, the usual practice of hand-fitting Landau coefficients could be replaced by an automatic search over invariant monomials, with the predictive error e acting as a built-in guard against overfitting.
  • A natural test is to apply the same scheme to BaTiO3 or BiFeO3: because those materials have more complex transition sequences, a successful simple polynomial there would substantially strengthen the claim that such potentials are generally predictive.
  • The method's reliance on equilibrium data alone means it could potentially turn published experimental polarization–temperature–field tables into mesoscale models without any atomistic input; the clamped-crystal extrapolation suggests the trained potential encodes physical couplings, not just fitting curves.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper introduces a machine-learning scheme to construct Landau free energy potentials for ferroelectric materials from thermodynamic training data. For PbTiO3, the training set consists of equilibrium polarizations and strains obtained from Monte Carlo simulations of a second-principles atomistic potential, sampled over temperature, electric field, and stress. Because the Landau potential is linear in its parameters, fitting reduces to linear least squares, and the authors search over tens of thousands of candidate polynomial models. Model selection is based on a new "predictive error" e, defined as the maximum deviation between the training states and the minima of the fitted potential when the potential is minimized from those states. The central result is that a 14-parameter model F7, with only two temperature-dependent coefficients, reproduces the Monte Carlo thermodynamics of PbTiO3 and, when evaluated under the clamped condition eta=0, predicts a continuous rhombohedral transition near 450 K, in agreement with the generating atomistic model. The authors argue this demonstrates the predictive power of simple, physics-informed Landau models and propose such "third-principles" potentials as a basis for mesoscale and macroscopic simulations.

Significance. If the claims are upheld, the paper offers a useful and computationally light method for deriving transparent Landau potentials from thermodynamic data, with obvious extensions to other ferroelectrics and to experimental data. The strengths of the work include the linear-in-parameters formulation, which makes the fit exact and fast; the automatic exploration of a huge model space; and the physically meaningful interpretation of selected couplings, e.g., the sign of A22 controlling tetragonal versus rhombohedral preference and the sign of A4i controlling the order of the transition. The eta=0 prediction is a genuine extrapolation in strain space and a nontrivial check of the fitted model. However, the validation methodology is not out-of-sample, and the benchmark is the same second-principles model that produced the training data, so the quantitative predictions are for the atomistic model, not directly for real PbTiO3. The paper is a solid methodological contribution, but its central "remarkable predictive power" claim requires a stronger validation protocol than the one presented.

major comments (3)
  1. [Section V B and Eq. (19)] The predictive error e is computed on the same training set used to fit the parameters; the paper explicitly states that no held-out dataset is used. Moreover, the minimization that defines e is initialized at the training state, so e tests only whether that state remains a local minimum of the fitted potential, not whether the model reproduces the global energy landscape or avoids spurious minima elsewhere. Because model selection and the central claim of predictive power rest on e, this is a load-bearing limitation. I recommend a concrete held-out test: for example, remove a block of temperatures or field values from the training set, fit the candidate models on the remainder, and compute e on the excluded states. Additionally, running minimizations from random initial polarizations and strains would test for spurious minima that the current initialization scheme cannot detect.
  2. [Section V A, Fig. 7, and Section III A] The eta=0 clamped-crystal benchmark is generated by the same second-principles model (Ref. 28) that produced the training data. The agreement shown in Fig. 7 therefore demonstrates that F7 is a good surrogate for the atomistic model, not that it predicts real PbTiO3; indeed, the paper notes that the atomistic model gives TC about 510 K versus 760 K experimentally. The claim of "remarkable predictive power" should be re-scaled to predictive power for the generating atomistic model. In addition, the extrapolation is less severe than implied: 450 K lies inside the training temperature range (100-700 K), and F7 already contains the C1(1)T thermal expansion term, so the extrapolation is only in strain, from states with |eta| at least 0.5% to eta=0. I would like the authors to state explicitly this limitation and to discuss what experimental or higher-fidelity test would be needed to support predictions for the real material.
  3. [Section IV B and Eq. (13)] No uncertainties are reported for the Monte Carlo thermal averages of polarization and strain, nor for the fitted Landau coefficients. Without error bars on the training data, the relative weighting of electric and elastic residuals in Eq. (13) rests on the dimensionless rescaling of Section III C, which is justified only qualitatively. The e<0.1 threshold used to select F7 therefore cannot be assigned a statistical meaning, and the physical conclusions that depend on the signs of coefficients such as A22, A4i, and C1(1) are presented without confidence intervals. I recommend adding bootstrap or error-propagation estimates for the fitted parameters and reporting the spread of e across the training states rather than only its maximum.
minor comments (5)
  1. [Section I] The phrase "organized as follow" should be "organized as follows."
  2. [Section II B 2] The paper states that conventional cross-validation schemes "yield models that seem unnecessarily complex and offer poor predictive power," but no systematic comparison is provided. A short benchmark of leave-p-out or a held-out validation set on the same candidate models would make this methodological claim persuasive rather than anecdotal.
  3. [Eqs. (17) and (18)] Equation (18) appears to contain a typo: the right-hand side uses P instead of the strain functional for eta; it should read eta(E, sigma, T).
  4. [Section IV B and Fig. 6] The abstract claims quantitative reproduction of Monte Carlo thermodynamics, but Fig. 6 shows a noticeable deviation for the strain component eta_zz even for model F7. The text should either explain this discrepancy in the abstract or qualify the claim of quantitative agreement.
  5. [Section III C] The statement that the chosen dimensionless rescaling makes further weighting unnecessary is supported only by "we checked these choices." It would be helpful to report what was checked, e.g., comparing fits with different relative weights of electric and elastic residuals.

Circularity Check

0 steps flagged · score 1.0 of 10

No circular derivation: F7 is fit to equilibrium equations of state, and the load-bearing eta=0 extrapolation is a genuinely out-of-sample test within the disclosed second-principles surrogate model.

full rationale

The derivation chain is not circular. F7 is obtained by least-squares fitting of the linear equations of state ∇_P F = E and ∇_η F = σ (Eqs. 6–7, 10–11, 13–14) to the 273-entry training set D; no target quantity such as the η = 0 transition temperature is used as a fitting constraint. The validation error e (Eq. 19) is computed on the same training states, but it is not the fitting objective, and the paper explicitly discloses that no held-out dataset is used (Section V B: “we do not employ a different dataset to evaluate the models”). The central predictive claim rests on the η = 0 test of Fig. 7, where F7, F8, and F9 are minimized under the elastic constraint η = 0 and compared with second-principles Monte Carlo data. The paper states that all configurations in D have strains of at least 0.5%, so the η = 0 condition is not in the training data; the test is therefore out-of-sample for the Landau model. Using the same second-principles model (Ref. 28) to generate D and to benchmark the η = 0 prediction is a validity caveat for statements about real PbTiO3, especially since that model gives T_C ≈ 510 K versus 760 K experimentally (Section III B), but it is not circularity: the Landau model's output is not an input to its own fit. Self-citations to the second-principles model are used as a tool source, not as an external uniqueness theorem or ansatz-forcing citation, and the paper itself flags the relevant limitations. The mild in-sample character of e lowers confidence in the claimed quantitative accuracy, but it does not reduce the main η = 0 prediction to the fitting data by construction.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central claim rests on fitted Landau coefficients, the assumption that a low-order polynomial captures the relevant thermodynamics, and the faithfulness of the second-principles model used to generate both training and validation data. No new physical entities are introduced; the term 'third-principles' is a naming convention, not a new object.

free parameters (4)
  • A2i(0) and A2i(1): quadratic polarization coefficient = A2i(0) = -3.81307e8 C^-2 m^2 N; A2i(1) = 0.83591e6 C^-2 m^2 N K^-1
    Main temperature-dependent polarization stiffness, fitted to MC equilibrium data via normal equations.
  • A4i(0), A22(0), A6i(0): higher-order polarization couplings = A4i(0)=0.90945e8, A22(0)=-1.67506e8, A6i(0)=2.01344e8 (SI)
    Higher-order invariants fitted to data; signs determine transition order and polarization anisotropy.
  • B12(0), B1'11(0): polarization-strain couplings = B12(0) = -69.23618e8 C^-2 m^2 N, B1'11(0) = 0.01283e8 C^-2 m^2 N
    Couplings between polarization and strain fitted to stress-dependent training states.
  • C1(0), C1(1), C2(0), C11(0), C2'(0): elastic and thermal expansion coefficients = C1(0)=-0.00068e11 m^-2 N, C1(1)=-2.50047e6 m^-2 N K^-1, C2(0)=0.81154e11 m^-2 N, C11(0)=0.57923e11 m^-2 N…
    Elastic and thermal expansion terms fitted to the training set; C1(1) captures thermal expansion.
assumptions (6)
  • domain assumption The free energy is expandable as a polynomial in homogeneous polarization and strain up to 8th order in P, quadratic in eta, with linear T dependence.
    Section II A; this restricts the model universe and is justified by Landau theory but not derived.
  • standard math Equilibrium states satisfy stationarity of the free energy with respect to P and eta (Eqs. 6-7).
    Section II B 1; standard thermodynamic equilibrium condition.
  • domain assumption The second-principles potential of Ref. 28 accurately represents PbTiO3 thermodynamics.
    Section III B; all training data and validation targets come from this model, which is the authors' own previous work.
  • domain assumption Training set states are representative equilibrium states and Monte Carlo averages are converged.
    Section III B; 10,000/40,000 sweeps and supercell sizes are claimed sufficient, but no convergence analysis is shown.
  • ad hoc to paper The dimensionless rescaling of energy, polarization, and temperature enables balanced errors without additional weights.
    Section III C; this choice affects the error functions E and e and the selected models.
  • ad hoc to paper Temperature dependence of coefficients is truncated at linear order, with no T dependence for higher-order couplings.
    Section II A; this restricts the model space and may influence the resulting simple-model outcome.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Machine learning Landau free energy potentials." pith.science (2026). https://pith.science/paper/KRBH4ORV

@misc{pith2026250723369,
  author       = {Pith},
  title        = {Pith review of: Machine learning Landau free energy potentials},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KRBH4ORV}},
  note         = {Machine review of arXiv:2507.23369}
}
abstract

We show how to construct Landau-like free energy potentials using a machine-learning approach. For concreteness, we focus on perovskite oxide PbTiO$_{3}$. We work with a training set obtained from Monte Carlo simulations based on an atomistic ''second-principles'' potential for PbTiO$_{3}$. We rely exclusively on data that would be experimentally accessible -- i.e., temperature-dependent polarization and strain, both with and without external electric fields and stresses applied --, to explore scenarios where the training set could be obtained from laboratory measurements. We introduce a scheme that allows us to identify optimal polynomial models of the temperature-dependent free energy surface, mapped as a function of the homogeneous electric polarization and homogeneous strain. Our results for PbTiO$_{3}$ show that a very simple polynomial -- where only two parameters depend linearly on temperature -- is sufficient to yield a correct description of the material's behavior. Remarkably, the obtained models also capture the subtle couplings by which elastic strain controls key features of ferroelectricity in PbTiO$_{3}$ -- i.e., the symmetry of the polar phase and the discontinuous character of the transition --, despite the fact that no effort was made to include such information in the training set. We emphasize the distinctive aspects of our methodology (which relies on an original form of validation step) by comparing it with the usual machine-learning approach for model construction. Our results illustrate how physically motivated models can have remarkable predictive power, even if they are derived from a limited amount of data. We argue that such ''third-principles'' models can be the basis for predictive macroscopic or mesoscopic simulations of ferroelectrics and other materials undergoing non-reconstructive structural transitions.

Figures

Figures reproduced from arXiv: 2507.23369 by the authors.

Figure 1
Figure 1. FIG. 1. Sketch illustrating our fitting and validation proce [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2. (a) Error function [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 4
Figure 4. FIG. 4. Equilibrium polarization as a function of temperature [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figures from the paper (2 more)
Figure 6
Figure 6. Figure 6: FIG. 6. Temperature dependence of polarization (a) and [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: FIG. 7. Temperature dependence of the polarization compo [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

46 extracted references · 18 canonical work pages

  1. [1]

    author author Albert P. \ Bartók \ and\ author Gábor \ Csányi ,\ title title Gaussian approximation potentials: A brief tutorial introduction , \ 10.1002/qua.24927 journal journal International Journal of Quantum Chemistry \ volume 115 ,\ pages 1051--1057 ( year 2015 ) NoStop

    doi:10.1002/qua.24927

  2. [2]

    author author Han \ Wang , author Linfeng \ Zhang , author Jiequn \ Han , \ and\ author Weinan \ E ,\ title title DeePMD -kit: A deep learning package for many-body potential energy representation and molecular dynamics , \ 10.1016/j.cpc.2018.03.016 journal journal Computer Physics Communications \ volume 228 ,\ pages 178--184 ( year 2018 ) NoStop

    doi:10.1016/j.cpc.2018.03.016

  3. [3]

    author author Ryosuke \ Jinnouchi , author Ferenc \ Karsai , \ and\ author Georg \ Kresse ,\ title title On-the-fly machine learning force field generation: Application to melting points , \ 10.1103/PhysRevB.100.014105 journal journal Phys. Rev. B \ volume 100 ,\ pages 014105 ( year 2019 ) NoStop

    doi:10.1103/physrevb.100.014105

  4. [4]

    author author Zheyong \ Fan , author Zezhu \ Zeng , author Cunzhi \ Zhang , author Yanzhou \ Wang , author Keke \ Song , author Haikuan \ Dong , author Yue \ Chen , \ and\ author Tapio \ Ala-Nissila ,\ title title Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport , \...

    doi:10.1103/physrevb.104.104309

  5. [5]

    author author Albert \ Musaelian , author Simon \ Batzner , author Anders \ Johansson , author Lixin \ Sun , author Cameron J. \ Owen , author Mordechai \ Kornbluth , \ and\ author Boris \ Kozinsky ,\ title title Learning local equivariant representations for large-scale atomistic dynamics , \ 10.1038/s41467-023-36329-y journal journal Nature Communicatio...

    doi:10.1038/s41467-023-36329-y

  6. [6]

    \ Xie , author Matthias \ Rupp , \ and\ author Richard G

    author author Stephen R. \ Xie , author Matthias \ Rupp , \ and\ author Richard G. \ Hennig ,\ title title Ultra-fast interpretable machine-learning potentials , \ 10.1038/s41524-023-01092-7 journal journal npj Computational Materials \ volume 9 ,\ pages 162 ( year 2023 ) NoStop

    doi:10.1038/s41524-023-01092-7

  7. [7]

    \ Elena , author Dávid P

    author author Ilyes \ Batatia , author Philipp \ Benner , author Yuan \ Chiang , author Alin M. \ Elena , author Dávid P. \ Kovács , author Janosh \ Riebesell , author Xavier R. \ Advincula , author Mark \ Asta , author Matthew \ Avaylon , author William J. \ Baldwin , author Fabian \ Berger , author Noam \ Bernstein , author Arghya \ Bhowmik , author Sam...

    arXiv doi:10.48550/arxiv.2401.00096

  8. [8]

    author author Ryan \ Jacobs , author Dane \ Morgan , author Siamak \ Attarian , author Jun \ Meng , author Chen \ Shen , author Zhenghao \ Wu , author Clare Yijia \ Xie , author Julia H. \ Yang , author Nongnuch \ Artrith , author Ben \ Blaiszik , author Gerbrand \ Ceder , author Kamal \ Choudhary , author Gabor \ Csanyi , author Ekin Dogus \ Cubuk , auth...

    arXiv

Show all 46 references
  1. [9]

    author author Hong-Liang \ Hu \ and\ author Long-Qing \ Chen ,\ title title Three-dimensional computer simulation of ferroelectric domain formation , \ https://doi.org/10.1111/j.1151-2916.1998.tb02367.x journal journal Journal of the American Ceramic Society \ volume 81 ,\ pag...

    1998

  2. [10]

    author author Long-Qing \ Chen ,\ title title Phase-field method of phase transitions/domain structures in ferroelectric thin films: A review , \ 10.1111/j.1551-2916.2008.02413.x journal journal Journal of the American Ceramic Society \ volume 91 ,\ pages 1835--1844 ( year 200...

    2008

  3. [11]

    author author Pierre \ Toledano \ and\ author Jean-Claude \ Toledano ,\ @noop title Landau theory of phase transitions \ ( publisher World Scientific Publishing Company ,\ year 1987 ) NoStop

    1987

  4. [12]

    author author Premi \ Chandra \ and\ author Peter B. \ Littlewood ,\ title title A landau primer for ferroelectrics , \ in\ 10.1007/978-3-540-34591-6_3 booktitle Physics of Ferroelectrics: A Modern Perspective \ ( publisher Springer Berlin Heidelberg ,\ address Berlin, Heidelb...

    doi

  5. [13]

    \ Devonshire ,\ title title Theory of ferroelectrics , \ 10.1080/00018735400101173 journal journal Advances in Physics \ volume 3 ,\ pages 85--130 ( year 1954 ) NoStop

    author author A.F. \ Devonshire ,\ title title Theory of ferroelectrics , \ 10.1080/00018735400101173 journal journal Advances in Physics \ volume 3 ,\ pages 85--130 ( year 1954 ) NoStop

    1954 doi

  6. [14]

    author author M. J. \ Haun , author E. Furman , author S. J. \ Jang , author H. A. \ McKinstry , \ and\ author L. E. \ Cross ,\ title title Thermodynamic theory of PbTiO _ 3 , \ 10.1063/1.339293 journal journal Journal of Applied Physics \ volume 62 ,\ pages 3331--3338 ( year ...

    doi

  7. [15]

    Tagantsev , author L

    author author A. Tagantsev , author L. E. \ Cross , \ and\ author J. Fousek ,\ @noop title Domains in Ferroic Crystals and Thin Films \ ( publisher Springer ,\ year 2010 ) NoStop

    2010

  8. [16]

    author author Dennis \ Meier , author Jan \ Seidel , author Marty \ Gregg , \ and\ author Ramamoorthy \ Ramesh ,\ @noop title Domain Walls : From Fundamental Properties to Nanotechnology Concepts \ ( publisher Oxford University Press ,\ year 2020 ) NoStop

    2020

  9. [17]

    Ivantchev , author J

    author author Jorge \ \'I\ niguez , author S. Ivantchev , author J. M. \ Perez-Mato , \ and\ author Alberto \ Garc\' a ,\ title title Devonshire-landau free energy of batio _ 3 from first principles , \ 10.1103/PhysRevB.63.144103 journal journal Phys. Rev. B \ volume 63 ,\ pag...

    doi

  10. [18]

    author author Anil \ Kumar \ and\ author Umesh V. \ Waghmare ,\ title title First-principles free energies and ginzburg-landau theory of domains and ferroelectric phase transitions in BaTiO _ 3 , \ 10.1103/PhysRevB.82.054117 journal journal Phys. Rev. B \ volume 82 ,\ pages 05...

    doi

  11. [19]

    author author Andrea \ Schiaffino \ and\ author Massimiliano \ Stengel ,\ title title Macroscopic polarization from antiferrodistortive cycloids in ferroelastic SrTiO _ 3 , \ @noop journal journal Phys. Rev. Lett. \ volume 119 ,\ pages 137601 ( year 2017 ) NoStop

    2017

  12. [20]

    author author Konstantin \ Shapovalov \ and\ author Massimiliano \ Stengel ,\ title title Tilt-driven antiferroelectricity in PbZrO _ 3 , \ 10.1103/PhysRevMaterials.7.L071401 journal journal Physical Review Materials \ volume 7 ,\ pages L071401 ( year 2023 ) NoStop

    doi

  13. [21]

    author author L. D. \ Landau \ and\ author E. M. \ Lifshitz ,\ @noop title Statistical Physics, Part 1 ,\ edition 3rd \ ed.,\ series Course of Theoretical Physics , Vol. volume 5 \ ( publisher Pergamon Press ,\ address Oxford ,\ year 1980 ) NoStop

    1980

  14. [22]

    author author Darrell G. \ Schlom , author Long-Qing \ Chen , author Chang-Beom \ Eom , author Karin M \ Rabe , author Stephen K \ Streiffer , \ and\ author Jean-Marc \ Triscone ,\ title title Strain tuning of ferroelectric thin films , \ 10.1146/annurev.matsci.37.061206.11301...

  15. [23]

    \ Press , author Saul A

    author author William H. \ Press , author Saul A. \ Teukolsky , author William T. \ Vetterling , \ and\ author Brian P. \ Flannery ,\ @noop title Numerical Recipes 3rd Edition: The Art of Scientific Computing ,\ edition 3rd \ ed.\ ( publisher Cambridge University Press ,\ year...

    2007

  16. [24]

    author author Carlos \ Escorihuela-Sayalero , author Jacek C. \ Wojde , \ and\ author Jorge \ \'I\ niguez ,\ title title Efficient systematic scheme to construct second-principles lattice dynamical models , \ 10.1103/PhysRevB.95.094115 journal journal Phys. Rev. B \ volume 95 ...

    doi

  17. [25]

    author author Trevor \ Hastie , author Robert \ Tibshirani , \ and\ author Jerome \ Friedman ,\ @noop title The Elements of Statistical Learning: Data Mining, Inference, and Prediction ,\ edition 2nd \ ed.\ ( publisher Springer ,\ year 2009 ) NoStop

    2009

  18. [26]

    \ Rabe , editor Charles H

    editor Karin M. \ Rabe , editor Charles H. \ Ahn , \ and\ editor Jean-Marc \ Triscone ,\ eds.,\ 10.1007/978-3-540-34591-6_1 title Physics of Ferroelectrics: A Modern Perspective \ ( publisher Springer Berlin Heidelberg ,\ address Berlin, Heidelberg ,\ year 2007 ) NoStop

    doi

  19. [27]

    author author R. D. \ King-Smith \ and\ author David \ Vanderbilt ,\ title title First-principles investigation of ferroelectricity in perovskite compounds , \ @noop journal journal Physical Review B \ volume 49 ,\ pages 5828 ( year 1994 ) NoStop

    1994

  20. [28]

    author author Jacek C \ Wojde , author Patrick \ Hermet , author Mathias P \ Ljungberg , author Philippe \ Ghosez , \ and\ author Jorge \ \'I\ niguez ,\ title title First-principles model potentials for lattice-dynamical studies: general methodology and example of application ...

    2013

  21. [29]

    author author Jacek C \ Wojde \ and\ author Jorge \ \'I \ n iguez ,\ title title Ferroelectric transitions at ferroelectric domain walls found from first principles , \ @noop journal journal Physical Review Letters \ volume 112 ,\ pages 247603 ( year 2014 a ) NoStop

    2014

  22. [30]

    author author Pavlo \ Zubko , author Jacek C \ Wojde , author Marios \ Hadjimichael , author St \'e phanie \ Fernandez-Pena , author Ana \"i s \ Sen \'e , author Igor \ Luk'yanchuk , author Jean-Marc \ Triscone , \ and\ author Jorge \ \'I \ n iguez ,\ title title Negative capa...

    2016

  23. [31]

    author author M. A. \ Pereira Gon c alves , author Carlos \ Escorihuela-Sayalero , author Pablo \ Garca-Fern \'a ndez , author Javier \ Junquera , \ and\ author Jorge \ \'I \ n iguez ,\ title title Theoretical guidelines to create and tune electric skyrmion bubbles , \ @noop j...

    2019

  24. [32]

    Das , author Y

    author author S. Das , author Y. L. \ Tang , author Z. Hong , author M.A.P. \ Gonçalves , author M.R. \ McCarter , author C. Klewe , author K.X. \ Nguyen , author F. G\'omez-Ortiz , author P. Shafer , author E. Arenholz , author V.A. \ Stoica , author S.-L. \ Hsu , author B. W...

    2019

  25. [33]

    Graf \ and\ author J

    author author M. Graf \ and\ author J. \' I \ n iguez ,\ title title A unified perturbative approach to electrocaloric effects , \ @noop journal journal Communications Materials \ volume 2 ,\ pages 60 ( year 2021 ) NoStop

    2021

  26. [34]

    author author Jacek C \ Wojde \ and\ author Jorge \ \'I \ n iguez ,\ title title Testing simple predictors for the temperature of a structural phase transition , \ @noop journal journal Physical Review B \ volume 90 ,\ pages 014105 ( year 2014 b ) NoStop

    2014

  27. [35]

    Zhong , author David \ Vanderbilt , \ and\ author K

    author author W. Zhong , author David \ Vanderbilt , \ and\ author K. M. \ Rabe ,\ title title First-principles theory of ferroelectric phase transitions for perovskites: The case of BaTiO _ 3 , \ @noop journal journal Physical Review B \ volume 52 ,\ pages 6301 ( year 1995 ) NoStop

    1995

  28. [36]

    author author Xingyue \ Ma , author Hongying \ Chen , author Ri He , author Zhanbo \ Yu , author Sergei \ Prokhorenko , author Zheng \ Wen , author Zhicheng \ Zhong , author Jorge \ \'I\ niguez-Gonz\'alez , author L. Bellaiche , author Di Wu , \ and\ author Yurong \ Yang ,\ ti...

    doi

  29. [37]

    author author I\ nigo \ Robredo , author Binayak \ Mukherjee , author Hugo \ Aramberri , \ and\ author Jorge \ \'I\ niguez-Gonz\'alez ,\ @noop title Minimalist machine-learned interatomic potentials can predict complex structural behaviors ac<curately , \ note In preparation (...

    2025

  30. [38]

    author author U. V. \ Waghmare \ and\ author K. M. \ Rabe ,\ title title Ab initio statistical mechanics of the ferroelectric phase transition in PbTiO _ 3 , \ @noop journal journal Physical Review B \ volume 55 ,\ pages 6161 ( year 1997 ) NoStop

    1997

  31. [39]

    author author Tiannan \ Yang , author Bo Wang , author Jia-Mian \ Hu , \ and\ author Long-Qing \ Chen ,\ title title Domain dynamics under ultrafast electric-field pulses , \ 10.1103/PhysRevLett.124.107601 journal journal Phys. Rev. Lett. \ volume 124 ,\ pages 107601 ( year 20...

    doi

  32. [40]

    author author Taorui \ Chen , author Bo Wang , author Yujie \ Zhu , author Shihao \ Zhuang , author Long-Qing \ Chen , \ and\ author Jia-Mian \ Hu ,\ title title Analytical model and dynamical phase-field simulation of terahertz transmission across ferroelectrics , \ 10.1103/P...

    doi

  33. [41]

    author author Jianyi \ Liu , author Haohua \ Wen , author Weijin \ Chen , \ and\ author Yue \ Zheng ,\ title title Atomistic studies of temporal characteristics of polarization relaxation in ferroelectrics , \ 10.1103/physrevb.103.014308 journal journal Physical Review B \ vol...

    doi

  34. [42]

    author author Haohua \ Wen , author Jianyi \ Liu , author Jinhong \ Li , author Bowen \ Li , author Weijin \ Chen , \ and\ author Yue \ Zheng ,\ title title A review of progress in theoretical modeling of polarization dynamics in ferroelectric materials , \ 10.1088/1361-648x/a...

    doi

  35. [43]

    \ Grisolia , author Hong Jian \ Zhao , author Otto E

    author author Peng \ Chen , author Mathieu N. \ Grisolia , author Hong Jian \ Zhao , author Otto E. \ Gonz \' a lez-V \' a zquez , author L. Bellaiche , author Manuel \ Bibes , author Bang-Gui \ Liu , \ and\ author Jorge \ \' I \ n iguez ,\ title title Energetics of oxygen-oct...

    doi

  36. [44]

    author author Pablo \ Garc\' a-Fern\'andez , author Jacek C \ Wojde , author Jorge \ \'I \ n iguez , \ and\ author Javier \ Junquera ,\ title title Second-principles method for materials simulations including electron and lattice degrees of freedom , \ @noop journal journal Ph...

    2016

  37. [45]

    author author Menghao \ Wu \ and\ author Ju Li ,\ title title Sliding ferroelectricity in 2D van der Waals materials: Related physics and future opportunities , \ 10.1073/pnas.2115703118 journal journal Proceedings of the National Academy of Sciences \ volume 118 ,\ pages e211...

    doi

  38. [46]

    author author Alexey \ Lipatov , author Pradeep \ Chaudhary , author Zhao \ Guan , author Haidong \ Lu , author Gang \ Li , author Olivier \ Crégut , author Kokou Dodzi \ Dorkenoo , author Roger \ Proksch , author Salia \ Cherifi-Hertel , author Ding-Fu \ Shao , author Evgeny ...

    doi

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.