Welcome to the
Sutton Lab

Welcome to the
Sutton Lab 

Machine
Learning Methods
Machine learning to predict quantum mechanical properties of atomistic systems (e.g., energy, bandgap, density, etc) 
Computational Materials Discovery
High-throughput methods and statistical learning methods for the computational screening of new materials
Electronic Structure Calculations
Calculated ground and excited state properties of materials and molecules to the gain a quantitative understanding of atomic-scale phenomena
Machine Learning
Machine learning to predict quantum mechanical  properties of atomistic systems (e.g., energy, bandgap, density, etc) 
Computational Materials Discovery
High-throughput methods and statistical learning methods for the computational screening of new materials
Electronic Structure Calculations
Calculated ground and excited state properties of materials and molecules to the gain a quantitative understanding of atomic-scale phenomena
Latest News from the Sutton Lab

Descriptors of materials properties to uncover functionality 

We used a novel data analytics approach called SISSO (sure independence screening and sparsifying operator), to derive a new tolerance factor to predict the stability of perovskites.
Read here
New Tolerance Factor to Predict the Stability of Perovskite Oxides and Halides

New Tolerance Factor to Predict the Stability of Perovskite Oxides and Halides

Button

Descriptors of materials properties to uncover functionality 

We used a novel data analytics approach called SISSO (sure independence screening and sparsifying operator), to derive a new tolerance factor to predict the stability of perovskites.
Read more
New Tolerance Factor to Predict the Stability of Perovskite Oxides and Halides

New Tolerance Factor to Predict the Stability of Perovskite Oxides and Halides

Button

Descriptors of materials properties to uncover functionality 

We used a novel data analytics approach called SISSO (sure independence screening and sparsifying operator), to derive a new tolerance factor to predict the stability of perovskites 
Read here
New Tolerance Factor to Predict the Stability of Perovskite Oxides and Halides

New Tolerance Factor to Predict the Stability of Perovskite Oxides and Halides

Button

Noncovalent intermolecular interactions in organic electronic materials: implications for the molecular packing vs electronic properties of acenes

Noncovalent intermolecular interactions can be tuned through the toolbox of synthetic chemistry and determine not only the molecular packing but also the resulting electronic, optical, and mechanical properties of materials derived from π-conjugated molecules, oligomers, and polymers.
Read here
View more

 Noncovalent intermolecular interactions in organic electronic materials: implications for the molecular packing vs electronic properties of acenes

Noncovalent intermolecular interactions can be tuned through the toolbox of synthetic chemistry and determine not only the molecular packing but also the resulting electronic, optical, and mechanical properties of materials derived from π-conjugated molecules, oligomers, and polymers.
Read more

Noncovalent intermolecular interactions in organic electronic materials: implications for the molecular packing vs electronic properties of acenes

Noncovalent intermolecular interactions can be tuned through the toolbox of synthetic chemistry and determine not only the molecular packing but also the resulting electronic, optical, and mechanical properties of materials derived from 
π-conjugated molecules, oligomers, and polymers.
Read here
View more

Identifying Domains of Applicability of MachineLearning Models of Materials 

A critical obstacle for this effort is that the complex choices involved in designing an ML model are currently made based on the overly simplistic metric of the average model error. This motivated the development of a new procedure that allows for opening the black box of ML models by identifying where and why ML models perform accurately, which we label as the domain of applicability (DA) of ML models.
Read here

Identifying Domains of Applicability of Machine Learning Models for Materials Science

A critical obstacle for this effort is that the complex choices involved in designing an ML model are currently made based on the overly simplistic metric of the average model error. This motivated the development of a new procedure that allows for opening the black box of ML models by identifying where and why ML models perform accurately, which we label as the domain of applicability (DA) of ML models.
Read more

Machine Learning for Heterogeneous Catalyst Design and Discovery

We recently highlighted several examples where machine learning is making an impact on hetero- geneous catalysis research, such as: accelerating the determination of catalyst active sites and catalyst screening; finding descriptors and patterns in catalysis data; determining interatomic potentials for catalyst simulation; and discovering and analyzing catalytic mechanisms.
Read here
Machine learning for heterogeneous catalyst design and discovery

Machine learning for heterogeneous catalyst design and discovery

Button