How AI is Revolutionizing the Discovery of Materials

Materials Discovery Precursor to Progress in Society

Materials Changed Societies and Enabled new Technology:

How are Materials Discovered?

Stainless steel, vulcanized rubber (car tires), Teflon, Play-doh, Saccharin, Super Glue,…

Edisonian (Trial and Error) Approach:

He tested over 6,000 plant materials to discover the final light bulb filament

wikimedia.org,

Materials-Discovery over Time

Of all (solid state) materials that we know of today, how many were discovered in the last 10 years?

pollev.com/peterschindler

Materials-Discovery over Time

Materials in ICSD

Top 500 HPC

Doubles every ~22 yrs

1st : Empirical Science

~33%

2nd: Modelbased Science

Doubles every ~1.3 yrs

3rd: Computational Science

4th: Data-driven Science

The 3rd Paradigm: Computational Discovery

Bulk modulus

Stress tensor

Surface

Work function

Surface/cleavage energy

Adsorption energy

Magnetic

Magnetic ordering

Magnetic moment

Dielectric constant

Absorption spectra

Density of states

Band structure

Thermodynamic

Vibrational entropy

Phase stability (Hull diagram)

Time Required for Experiment vs. Computation

Experiments (Synthesis) ~ weeks to months (to a Ph.D.)

First principles calculations ~ hours to days (to weeks)

Still too long to screen >100,000 candidates

Discovery Cluster

Over 20,000 CPU Cores and Over 200 GPUs

Hosted at MGHPCC (90,000-square-foot facility)

The 4th Paradigm: Data-Driven Discovery

Derived or measured properties

Physical & Chemical insight/intuition

* = fingerprint = feature (vector)

High-Quality Data?

Types of Materials Data

Types of Materials Data

XRD, TEM, etc.

NLP and LLMs

Types of Materials Data

Atomistic simulations

Materials properties

High-throughput computations of properties

Micrographs

XRD, TEM, etc.

Scientific Literature

NLP and LLMs

ML Paradigms in Materials Science

Model-centric AI

How change the model/architecture to improve performance?

Shallow ML + feature engineering

Experimental Computational Small Data

Big Data Spectra, Images

Active Learning

Data-centric AI

How systematically change data (x/y) to improve performance?

Big Data “Good Data”

Deep Learning

Experimental Computational Small Data

Big Data Spectra, Images

Experimental Computational Small Data

Big Data Spectra, Images

Transfer Learning

Experimental Computational Small Data

Big Data Spectra, Images

Materials Descriptors

Examples of Data-Driven Discovery

Industrial Perspective

Crystal Structures vs. Molecules

Crystal

Coordinates

Atom types

Lattice vectors

Molecule

Coordinates

Atom types

Periodic

Space group symmetry

Non-periodic

Point group symmetry

E(3) invariant

Requirements for an Ideal Materials Descriptor

i. Meaningful and Universal (and fixed in number)

ii. Compact and Cheap(er) to Compute

iii. Invariant Under Crystal Symmetries (and atom permutations)

iv. Continuous (small change in atomic structure = small change in descriptor)

v. Reversible

vi. Unique

vii. Additive

viii. Uncorrelated

Hierarchy of Materials Descriptors

CGCNN

Crystal Graph Convolutional NN

ALIGNN

Atomistic Line Graph NN

Others: M3GNet, SchNet, PointNet, PAINN, DimeNet++, …

Invariant to E(3)

E(3) Equivariant GNNs

Requires data augmentation

(inefficient & not physical)

No additional data required

Improved transferability and data efficiency

E(3) Equivariant GNNs

Interested in the math/CS details?

→ Prof. Robin Walters at Northeastern (Khoury College)

“Affordable Accuracy”

HIV Capsid with 44 Million Atoms

Majority of Energy Goes to Waste(-Heat)

Thermionic Energy Converter (TEC)

Vacuum gap

Heat Input

• No moving parts

• Power output scales with area

TEC Efficiency

High-Throughput DFT Workflow

Work Function Database

Model Performance with Physics-motivated Descriptors

Elemental features

200 features

15 features

~105 faster than DFT

Structural features

Promising New Low Work Function Surfaces

After ionic relaxation: Discovery of metallic surfaces with WF < 1.5 eV:

CsScCl3, Hexagonal, (100) Surface

BaX [X=Si, Sn, Ge], Orthorhombic, (110) Surface

Discovery of New High-Brightness Photocathodes for XFEL

Electron emission from Photocathodes Depends on Work Function

Work Function

Photocathode Brightness ∝ 1 / spread in transverse momentum of electrons

Intrinsic Emittance

Physical Review B, 101 (2020).

E.R.Antoniuk, Y. Yue, Y. Zhou, P. Schindler, W.A. Schroeder, B. Dunham, P. Pianetta, T. Vecchione, & E.J. Reed. Generalizable DFT-based photoemission model for the accelerated development of photocathodes and other photoemissive devices

Ab-initio Photoemission Model

Novel Ultra-bright and Air-Stable Photocathodes

Discovered through ML/DFT Driven Screening

11 materials with intrinsic emittance < 0.3 µm/mm

+ 3 air stable low intrinsic emittance materials M2O (M = Na, K, Rb)

Advanced Materials, 33, 44 (2021)

E. R. Antoniuk, P. Schindler, W. A. Schroeder, B. Dunham, P. Pianetta, T. Vecchione, and E. J. Reed

Materials Descriptors

Examples of Data-Driven Discovery

Industrial Perspective

“Materials Informatics” in Industry?

Industry Perspective: Aionics Inc.

Structure generator leveraging AI potentials to construct crystals, surfaces, molecules, etc.

Database of 10B+ candidates, searchable by physical properties, safety, supply chain, price, etc.

Synthesize New Formulations

AI platform to incorporate latest data, train new model, and guide next selection

Cloud-based DFT to compute properties of candidate

Source at Production Scale

Co-innovation partnerships with electrochemical materials manufacturers:

Aionics tools used internally to lead client’s in-house R&D

Physics-informed ML & ML-informed Physics