Abstract:Quasicrystals represent a class of ordered materials that have diffraction symmetry forbidden in periodic crystals. Since the first discovery of quasicrystals in 1984, approximately 100 thermodynamically stable quasicrystals have been synthesized. The discovery of new quasicrystals has led to the observation of novel physical phenomena, such as robust quantum criticality, fractal superconductivity, and peculiar long-range magnetic ordering. However, the pace of discovery of new quasicrystals has significantly slowed down, which is attributed to the lack of design principles for exploring new quasicrystals. Here, we demonstrate that machine learning can greatly accelerate the process of material discovery. Our model can predict stable quasicrystalline phases with high accuracy. With this model, we discovered three stable decagonal quasicrystals through an exhaustive screening of more than 1000 ternary aluminum alloy systems. https://doi.org/10.1103/PhysRevMaterials.7.093805 Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article's title, journal citation, and DOI. Published by the American Physical Society

What problem does this paper attempt to address?

This paper mainly discusses how to use machine learning to accelerate the discovery of new quasicrystalline materials. Quasicrystals are materials that have ordered but non-periodic diffraction symmetry in a crystalline structure. Since the first discovery of quasicrystals in 1984, although approximately 100 thermodynamically stable quasicrystals have been synthesized, the rate of discovering new quasicrystals has significantly slowed down. The research team has developed a machine learning model that can accurately predict the chemical composition of stable quasicrystalline phases. Through large-scale screening in over 1000 ternary aluminum alloy systems, they successfully discovered three new decagonal quasicrystals. The paper mentions that the traditional method based on the electron concentration rule (e/a rule) proposed by D. Shechtman and others has limitations in finding quasicrystals, while the machine learning algorithm can autonomously learn this rule and predict more types of quasicrystals. In the aluminum-transition metal alloy system, the predicted quasicrystal formation regions by the model mostly coincide with the e/a=1.8 rule. In addition, the researchers extracted five rules from the machine learning model, which are related to the physical and chemical properties of the elements and contribute to understanding the mechanism of quasicrystal formation. Through virtual screening, the research team selected three systems out of 185 predicted systems that may contain quasicrystals, and ultimately successfully synthesized three decagonal quasicrystals. These newly discovered quasicrystals are outside the composition range of previously discovered quasicrystals, indicating the capability of the machine learning model to predict new materials. Overall, this paper demonstrates the potential application of machine learning in materials science, especially in accelerating quasicrystal exploration, providing new avenues for the discovery of more quasicrystals with unique properties in the future.

Quasicrystals predicted and discovered by machine learning

Deep Learning Enables Rapid Identification of a New Quasicrystal from Multiphase Powder Diffraction Patterns

Quasicrystals as an intermediate form of matter between crystalline and amorphous solids

Quasicrystalline Nanocrystal Superlattice with Partial Matching Rules

The Coming of Quasicrystal: Discovery, Principle and Application

Quasicrystal structure prediction: A review

Quasicrystal-related Mosaics with Periodic Lattices Interlaid with Aperiodic Tiles

VQCrystal: Leveraging Vector Quantization for Discovery of Stable Crystal Structures

On the Robustness of Machine Learning Models in Predicting Thermodynamic Properties: a Case of Searching for New Quasicrystal Approximants

Quasicrystal synthesis by shock compression

An atomic scale study of two-dimensional quasicrystal nucleation controlled by multiple length scale interactions

Simulated structure and thermodynamics of decagonal Al-Co-Cu quasicrystals

High-Temperature Atomic Diffusion and Specific Heat in Quasicrystals

Atomic diffusion due to hyperatomic fluctuation for quasicrystals

Spontaneous formation of thermodynamically stable Al--Cu--Fe icosahedral quasicrystal from realistic atomistic simulations

Emergent Moiré fringes in direct-grown quasicrystal

Inverse Design of Crystals and Quasicrystals in a Non-Additive Binary Mixture of Hard Disks

Quasicrystal bulk and surface energies from density functional theory

Superconducting Quasicrystals

Aperiodic approximants bridging quasicrystals and modulated structures

Photonic quasicrystal of spin angular momentum