Predicting Molecular Properties via Computational Chemistry

Predicting Molecular Properties via Computational Chemistry

Computational Chemistry is a growing field intersecting with nearly all disciplines of chemistry, as its ability to accurately model and predict chemical reactivity, molecular structures and properties is crucial to modern molecular design and synthesis.

J-STAR has computational chemists, and the necessary infrastructure dedicated to computational chemistry. Our computational chemistry is integrated in our scientifically driven chemical process development. Integrated in route scouting, optimization of reactions, better understanding of structures and properties of working catalysts, computational chemistry aids in the first principal driven development of our processes. Using relatively routine methods, our skilled computational chemists can provide predictions that are close to experimental data with proper considerations of methods and basis sets of various important chemical properties, strongly complementing classical experimental results.

Properties that can be predicted from computational models include:

  • 3-D molecular geometries, torsions potential energy scan, conformers population distribution in solutions
  • Stereoisomers population distribution in solvent systems
  • Absolute and Relative Free Energies of chemical reactions in solvent systems
  • Electronic Properties (charge distribution, molecular orbitals, reactivity indices, etc.)
  • Spectroscopic Properties (IR, NMR, UV-vis, ECD, VCD, Raman)
  • Bond Dissociation Energies (BDE)
  • Free energy of atom abstraction scan in a solvent system (e.g. hydrogen/proton, Li)
  • Acid Dissociation Constants (pKa

In addition to molecular properties, entire or partial reaction pathways can be examined at the molecular scale predicting values like activation energy and site reactivity, giving insights into reaction kinetics, transition states, reaction orders, product distributions and more.

The reasonable predictions produced through computational investigation can help guide reaction development by understanding non-trivial molecular interactions. Using these predictions, J-STAR Research scientists can gain deeper mechanistic insight, rationalize experimental observations, and identify improved strategies and routes towards their targeted molecules.

Using up to 24-cores/48-threads and 192GB of RAM, the computational resources at J-STAR Research provides a wide range of capabilities. A few of our most frequently employed software packages include:

  • Gaussian 16
  • xtb
  • CREST

And a few of the most commonly used methods include:

  • Dispersion-corrected Density Functional Theory (DFT-D)
  • Molecular Mechanics and Dynamics (MM & MD)
  • Semi-empirical quantum chemistry methods (GFNn-xTB)
  • ONIOM (QM/MM methods)
Scroll to Top