Research in the Kulik group

Research in the Kulik group leverages computational modeling to aid the discovery of new materials and mechanisms. Our group uses first-principles modeling to unearth fundamental aspects of structure-property relationships in catalysts and materials. By taking a computational approach, we carry out studies that allow us to make connections across a wide range of catalytic systems from biological enzymes to emerging heterogeneous single-atom catalysts. We develop computational software and machine learning models that accelerate the discovery of new materials and design rules. This approach enables the prediction of new materials properties in seconds, the exploration of million-compound design spaces, and the identification of design rules and exceptions that go beyond intuition. To ensure the predictive power of our approach, our group develops new methods to increase the accuracy of density functional theory especially for materials with challenging electronic structure such as transition metal complexes and solids.

Read more about our group’s work in the recently published papers below!


Discovering inorganic complexes with an ANN

Machine learning (ML) has emerged as a powerful complement to simulation for materials discovery by reducing time for evaluation of energies and properties at accuracy competitive with first-principles methods. In this work, we take our recently developed artificial neural network (ANN) that can predict spin-state ordering to within 3 kcal/mol of DFT training data and use it for chemical exploration.

How does solid state density localize? (Ed Choice)

Widely employed approximate density functional theory (DFT) suffers from delocalization errors. DFT+U and hybrid functionals are widely employed methods to correct energetic delocalization errors, but their effect on the density is less well known. Our recent work demonstrated that in transition metal complexes both methods localize density away from the metal and onto surrounding ligands, regardless of metal or ligand identity, in a similar fashion.

Scaling relations in single-site catalysis

Computational catalyst screening is limited primarily by the efficiency with which accurate predictions can be made. In bulk heterogeneous catalysis, linear free energy relationships (LFERs) accelerate screening by relating catalytic activity back to the adsorption energies of key intermediates, but their applicability to single-site catalysts remains unclear, in view of the directional, covalent metal-ligand bonds and the broader chemical space of accessible ligand scaffolds.

Recovering exact conditions in semilocal DFT

Widely employed semi-local DFT suffers from well-known errors that prevent its robust predictio, e.g. in materials and catalyst design. This failure in semi-local DFT can be traced to the violation of exchange-correlation approximations of key exact conditions. The flat-plane condition is the union of two exact constraints in electronic structure theory: (i) energetic piecewise linearity with fractional electron removal or addition and (ii) invariant energetics with change in electron spin in a half filled orbital.

Inorganic features for machine learning

Machine learning models can accelerate the discovery of new materials by allowing rapid screening of potential candidates, but depend on the way the molecule is represented. For transition metal complexes, accurate calculations are computationally costly and available training data sets are small, so finding the correct representation becomes a critical ingredient for model predictive accuracy.

Unifying sensitivity in catalysis and spin states

Open-shell single-site transition metal (TM) complexes are efficient catalysts, but challenging to screen computationally because of 1) the need to simultaneously predict spin-splitting energies, activation barriers and reaction energies, and 2) well-known challenges of approximate DFT for these predictions that manifest as strong sensitivity to the amount of exact exchange in the functional. Hence, advancing DFT-based screening requires understanding how computed energetics are sensitive to exact exchange.

Conceptual DFT for systematic modeling

Multi-scale quantum-mechanical/molecular-mechanical (QM/MM) and large-scale QM simulation provide valuable insight into enzyme mechanism and structure-property relationships. Analysis of the electron density afforded through these methods can enhance our understanding of how the enzyme environment modulates reactivity at the enzyme active site. From this perspective, tools from conceptual density functional theory to interrogate electron densities can provide added insight into enzyme function.

Predicting electronic structure with ANNs

Direct density functional theory (DFT) simulation is the critical bottleneck in high-throughput computational screening for inorganic materials and molecular transition metal complexes. These calculations are computationally costly and the calculated properties are sensitive to the exchange–correlation functional employed. Screening efforts for molecular inorganic systems are complicated by the uncertain spin-state ordering of hypothetical complexes and difficulty in obtaining good initial guess geometries.

Discovering chemistry from large databases

Using our recently developed inorganic discovery toolkit, molSimplify code, candidate molecules were obtained from the ChEMBL-19 database (> 1M molecules) to address a critical outstanding challenge in materials science. These databases of bioactive organic molecules are typically employed for discovery of therapeutic drug-like molecules; we instead demonstrated their power as a tool to discover design rules for inorganic complexes while maintaining realism (i.e., stable, synthetically accessible substituents) and providing diversity in functional groups.

Inorganic discovery & machine learning

Virtual high throughput screening has emerged as a powerful tool for the discovery of new materials. Although the computational materials science community has benefited from open source tools for the rapid structure generation, calculation, and analysis of crystalline inorganic materials, software and strategies to address the unique challenges of inorganic complex discovery have not been as widely available. We present our recent developments in the open source molSimplify code for inorganic discovery.

Pages

About Us

The Kulik group focuses on the development and application of new electronic structure methods and atomistic simulations tools in the broad area of catalysis.

Our Interests

We are interested in transition metal chemistry, with applications from biological systems (i.e. enzymes) to nonbiological applications in surface science and molecular catalysis.

Our Focus

A key focus of our group is to understand mechanistic features of complex catalysts and to facilitate and develop tools for computationally driven design.

Contact Us

Questions or comments? Let us know! Contact Dr. Kulik: