Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
144 tokens/sec
GPT-4o
8 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Discovering High-Strength Alloys via Physics-Transfer Learning (2403.07526v2)

Published 12 Mar 2024 in cond-mat.mtrl-sci, cs.LG, and physics.comp-ph

Abstract: Predicting the strength of materials requires considering various length and time scales, striking a balance between accuracy and efficiency. Peierls stress measures material strength by evaluating dislocation resistance to plastic flow, reliant on elastic lattice responses and crystal slip energy landscape. Computational challenges due to the non-local and non-equilibrium nature of dislocations prohibit Peierls stress evaluation from state-of-the-art material databases. We propose a data-driven framework that leverages neural networks trained on force field simulations to understand crystal plasticity physics, predicting Peierls stress from material parameters derived via density functional theory computations, which are otherwise computationally intensive for direct dislocation modeling. This physics transfer approach successfully screen the strength of metallic alloys from a limited number of single-point calculations with chemical accuracy. Guided by these predictions, we fabricate high-strength binary alloys previously unexplored, utilizing high-throughput ion beam deposition techniques. The framework extends to problems facing the accuracy-performance dilemma in general by harnessing the hierarchy of physics of multiscale models in materials sciences.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (49)
  1. Jain, A. et al. Commentary: The materials project: A materials genome approach to accelerating materials innovation. \JournalTitleAPL Mater. 1, 011002 (2013).
  2. Materials design and discovery with high-throughput density functional theory: The open quantum materials database (OQMD). \JournalTitleJOM 65, 1501–1509 (2013).
  3. Mesoscopic and multiscale modelling in materials. \JournalTitleNat. Mater. 20, 774–786 (2021).
  4. Merchant, A. et al. Scaling deep learning for materials discovery. \JournalTitleNature 624, 80–85 (2023).
  5. Wu, Z. et al. MoleculeNet: A benchmark for molecular machine learning. \JournalTitleChem. Sci. 9, 513–530 (2018).
  6. PI1M: A benchmark database for polymer informatics. \JournalTitleJ. Chem. Inf. Model. 60, 4684–4690 (2020).
  7. Mansouri Tehrani, A. et al. Machine learning directed search for ultraincompressible, superhard materials. \JournalTitleJ. Am. Chem. Soc. 140, 9844–9853 (2018).
  8. Atomic stiffness for bulk modulus prediction and high-throughput screening of ultraincompressible crystals. \JournalTitleNat. Commun. 14, 4258 (2023).
  9. Polycrystalline morphology and mechanical strength of nanotube fibers. \JournalTitleNPJ Comput. Mater. 8, 15 (2022).
  10. Presto: Rapid protein mechanical strength prediction with an end-to-end deep learning model. \JournalTitleExtreme Mech. Lett. 55, 101803 (2022).
  11. Mycelium-based wood composites for light weight and high strength by experiment and machine learning. \JournalTitleCell Rep. Phys. Sci. 4, 101424 (2023).
  12. Deep learning accelerated design of mechanically efficient architected materials. \JournalTitleACS Appl. Mater. Interfaces 15, 22543–22552 (2023).
  13. Crystal Plasticity Finite Element Methods (John Wiley & Sons, 2011).
  14. Courtney, T. H. Mechanical Behavior of Materials (Waveland Press, 2005).
  15. Strain hardening due to deformation twinning in α𝛼\alphaitalic_α-titanium: Constitutive relations and crystal-plasticity modeling. \JournalTitleActa Mater. 53, 3495–3502 (2005).
  16. ⟨a⟩delimited-⟨⟩a\langle\rm a\rangle⟨ roman_a ⟩ Prismatic, ⟨a⟩delimited-⟨⟩a\langle\rm a\rangle⟨ roman_a ⟩ basal, and ⟨c+a⟩delimited-⟨⟩ca\langle\rm{c+a}\rangle⟨ roman_c + roman_a ⟩ slip strengths of commercially pure ZZ\mathrm{Z}roman_Zr by micro-cantilever tests. \JournalTitleActa Mater. 96, 249–257 (2015).
  17. Predictive crystal plasticity modeling of single crystal nickel based on first-principles calculations. \JournalTitleJOM 74, 1423–1434 (2022).
  18. Molecular dynamics simulation of stress field around edge dislocations in aluminum. \JournalTitleComput. Mater. Sci. 84, 83–96 (2014).
  19. Nabarro, F. Dislocations in a simple cubic lattice. \JournalTitleProc. Phys. Soc. 59, 256 (1947).
  20. Semidiscrete variational Peierls framework for dislocation core properties. \JournalTitlePhys. Rev. Lett. 78, 4221 (1997).
  21. Nabarro, F. Fifty-year study of the Peierls-Nabarro stress. \JournalTitleMater. Sci. Eng. A 234, 67–76 (1997).
  22. Generalized-stacking-fault energy surface and dislocation properties of aluminum. \JournalTitlePhys. Rev. B 62, 3099 (2000).
  23. Ab initio modeling of dislocation core properties in metals and semiconductors. \JournalTitleActa Mater. 124, 633–659 (2017).
  24. Density functional theory calculations of generalized stacking fault energy surfaces for eight face-centered cubic transition metals. \JournalTitleJ. Appl. Phys. 126 (2019).
  25. Considerations for choosing and using force fields and interatomic potentials in materials science and engineering. \JournalTitleCurr. Opin. Solid State Mater. Sci. 17, 277–283 (2013).
  26. Evaluating variability with atomistic simulations: The effect of potential and calculation methodology on the modeling of lattice and elastic constants. \JournalTitleModel. Simul. Mat. Sci. Eng. 26, 055003 (2018).
  27. Martin, R. M. Electronic Structure: Basic Theory and Practical Methods (Cambridge university press, 2020).
  28. Large-scale atomistic simulation of dislocation core structure in face-centered cubic metal with deep potential method. \JournalTitleComput. Mater. Sci. 218, 111941 (2023).
  29. Recent advances and outstanding challenges for machine learning interatomic potentials. \JournalTitleNat. Comput. Sci. 3, 998–1000 (2023).
  30. Hedman, D. et al. Dynamics of growing carbon nanotube interfaces probed by machine learning-enabled molecular simulations. \JournalTitlearXiv preprint arXiv:2302.09542 (2023).
  31. An accurate machine learning interatomic potential for fcc and hcp nickel. \JournalTitlearXiv preprint arXiv:2312.17596 (2023).
  32. Deep potential molecular dynamics: A scalable model with the accuracy of quantum mechanics. \JournalTitlePhys. Rev. Lett. 120, 143001 (2018).
  33. Takamoto, S. et al. Towards universal neural network potential for material discovery applicable to arbitrary combination of 45 elements. \JournalTitleNat. Commun. 13, 2991 (2022).
  34. How to validate machine-learned interatomic potentials. \JournalTitleJ. Chem. Phys. 158 (2023).
  35. Deep residual learning for image recognition. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit. (CVPR), 770–778 (2016).
  36. Learning from Data, vol. 4 (AMLBook New York, 2012).
  37. Activity-weight duality in feed-forward neural networks reveals two co-determinants for generalization. \JournalTitleNat. Mach. Intell. 5, 908–918 (2023).
  38. Big data meets quantum chemistry approximations: The ΔΔ\Deltaroman_Δ-machine learning approach. \JournalTitleJ. Chem. Theory Comput. 11, 2087–2096 (2015).
  39. Multifidelity information fusion with machine learning: A case study of dopant formation energies in hafnia. \JournalTitleACS Appl. Mater. Interfaces 11, 24906–24918 (2019).
  40. Smith, J. S. et al. Approaching coupled cluster accuracy with a general-purpose neural network potential through transfer learning. \JournalTitleNature Commun. 10, 2903 (2019).
  41. Computer Simulations of Dislocations (Oxford University Press, 2006).
  42. A multi-scale model of dislocation plasticity in α𝛼\alphaitalic_α-Fe: Incorporating temperature, strain rate and non-Schmid effects. \JournalTitleInt. J. Plast. 73, 100–118 (2015).
  43. Plimpton, S. Fast parallel algorithms for short-range molecular dynamics. \JournalTitleJ. Comput. Phys. 117, 1–19 (1995).
  44. Blöchl, P. E. Projector augmented-wave method. \JournalTitlePhys. Rev. B 50, 17953 (1994).
  45. From ultrasoft pseudopotentials to the projector augmented-wave method. \JournalTitlePhys. Rev. B 59, 1758 (1999).
  46. Generalized gradient approximation made simple. \JournalTitlePhys. Rev. Lett. 77, 3865 (1996).
  47. Special points for Brillouin-zone integrations. \JournalTitlePhys. Rev. B 13, 5188 (1976).
  48. Fan, Z. et al. Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport. \JournalTitlePhys. Rev. B 104, 104309 (2021).
  49. Train faster, generalize better: Stability of stochastic gradient descent. In Int. Conf. Mach. Learn. (ICML), 1225–1234 (2016).

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now