Machine learning for autonomous crystal structure identification

Reinhart, Wesley F.; Long, Andrew; Howard, Michael P.; Ferguson, Andrew L.; Panagiotopoulos, Athanassios Z.

doi:10.1039/c7sm00957g

Cited by 101 publications

(101 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Recently, several models were developed to predict the properties of various material classes such as perovskites [27,28], oxides [29], elpasolites [30,31], thermoelectrics [32][33][34], and metallic glasses [35]. Additionally, generalized approaches have been devised for inorganic materials [36][37][38][39][40][41][42][43] and for systematically identifying efficient physical models of materials properties [44].…”

Section: Introductionmentioning

confidence: 99%

AFLOW-ML: A RESTful API for machine-learning predictions of materials properties

Gossett

Toher

Oses

et al. 2018

Computational Materials Science

View full text Add to dashboard Cite

Machine learning approaches, enabled by the emergence of comprehensive databases of materials properties, are becoming a fruitful direction for materials analysis. As a result, a plethora of models have been constructed and trained on existing data to predict properties of new systems. These powerful methods allow researchers to target studies only at interesting materialsneglecting the non-synthesizable systems and those without the desired properties -thus reducing the amount of resources spent on expensive computations and/or time-consuming experimental synthesis. However, using these predictive models is not always straightforward. Often, they require a panoply of technical expertise, creating barriers for general users. AFLOW-ML (AFLOW Machine Learning) overcomes the problem by streamlining the use of the machine learning methods developed within the AFLOW consortium. The framework provides an open RESTful API to directly access the continuously updated algorithms, which can be transparently integrated into any workflow to retrieve predictions of electronic, thermal and mechanical properties. These types of interconnected cloud-based applications are envisioned to be capable of further accelerating the adoption of machine learning methods into materials development.

show abstract

Section: Introductionmentioning

confidence: 99%

AFLOW-ML: A RESTful API for machine-learning predictions of materials properties

Gossett

Toher

Oses

et al. 2018

Computational Materials Science

View full text Add to dashboard Cite

show abstract

“…The PCA calculations for the hard-disk model can be trivially extended to detect the ordering transition in the three-dimensional hard-sphere model, for which there are various detection schemes in the literature. 40,[48][49][50][51][52] However, key practical differences regarding crystallization processes in these two systems are worth considering. First, once hard spheres crystallize in a simulation, there is an almost complete arrest of appreciable collective particle rearrangements.…”

Section: A Freezingmentioning

confidence: 99%

“…Such domains can be identified via established structure identification methods. 40,[49][50][51][52] Annealing out any resultant mosaic pattern is practically impossible due to the aforementioned kinetic and thermodynamic issues. Robust PCA detection of the ordering transitions should reflect these practical differences.…”

Section: A Freezingmentioning

confidence: 99%

Unsupervised machine learning for detection of phase transitions in off-lattice systems. I. Foundations

Jadrich

Lindquist

Truskett

2018

The Journal of Chemical Physics

View full text Add to dashboard Cite

We demonstrate the utility of an unsupervised machine learning tool for the detection of phase transitions in off-lattice systems. We focus on the application of principal component analysis (PCA) to detect the freezing transitions of two-dimensional hard-disk and three-dimensional hard-sphere systems as well as liquid-gas phase separation in a patchy colloid model. As we demonstrate, PCA autonomously discovers order-parameter-like quantities that report on phase transitions, mitigating the need for a priori construction or identification of a suitable order parameterthus streamlining the routine analysis of phase behavior. In a companion paper, we further develop the method established here to explore the detection of phase transitions in various model systems controlled by compositional demixing, liquid crystalline ordering, and non-equilibrium active forces.

show abstract

“…[30] Novel method has been implemented to identify the crystalline structures during the evaporation. [31] The evaporation can also take place on a curved surface, as the solution takes a cylindrical or spherical geometry. Su et al have developed a method to produce nanowires composed of polymers or nanoparticles.…”

Section: Wwwadvmatde Wwwadvancedsciencenewscommentioning

confidence: 99%

Structure Formation in Soft‐Matter Solutions Induced by Solvent Evaporation

et al. 2017

View full text Add to dashboard Cite

etc., and the interplay between different processes determines the final structure of the dried materials.Here, we report recent advances of selfassembled structures induced by uniform evaporation, examples including drying in a thin-film geometry or evaporating a spherical droplet on a superhydrophobic surface. [4] In these situations, the structure formation takes place in the direction of the evaporation. This excludes the case where lateral flow is induced by nonuniform drying. The lateral flow is important in explaining the coffee-ring effect, where excellent reviews exist in literature. [5,6] After introducing the general procedure in experiments, we explain the physical mechanisms responsible for the structure formation in solvent evaporation. We then discuss recent advances based on single-and binarycomponent solutions, with a special emphasis on the theoretical approaches. Drying of Single-Component SolutionsA general experiment starts with spin-coating a thin-film solution on a substrate, then the sample is left to dry. In the first step, it is important that the spin-cast film is uniform. [7,8] In the second step, the evaporation flux, i.e., the mass removed from the liquid phase to the vapor phase in unit time from unit surface area needs to be controlled precisely. Even for the pure solvent evaporation, the process is complex and involves the diffusion of solvent molecules in both the liquid and vapor phases, and external flow in the vapor phase to prevent saturation. The evaporation flux can be computed using the Hertz-Knudsen equation that is based on the thermodynamics variables [9][10][11] or solving the diffusion equation in the vapor phase using suitable boundary conditions. [12][13][14] The solvent evaporation drives the free surface to recede toward the substrate, resulting in an accumulation of the solutes at the surface. Two competing processes determine the spatial distribution of the solutes. One is the Brownian motion, which is characterized by the diffusion constant D of the solutes. The other is the evaporation, characterized by the rate v ev at which the free surface recedes. Two timescales are related to these two processes: the diffusion time τ D =  2 /D, where  is a characteristic length and the evaporation time τ ev = /v ev . To quantify the relative importance between the diffusion and the evaporation, one can define a dimensionless number called film formation Peclet number: [15]

show abstract

Machine learning for autonomous crystal structure identification

Cited by 101 publications

References 52 publications

AFLOW-ML: A RESTful API for machine-learning predictions of materials properties

AFLOW-ML: A RESTful API for machine-learning predictions of materials properties

Unsupervised machine learning for detection of phase transitions in off-lattice systems. I. Foundations

Structure Formation in Soft‐Matter Solutions Induced by Solvent Evaporation

Contact Info

Product

Resources

About