The Kosenkov Group

Physical Chemistry at Monmouth University


Our field of research is computational and theoretical chemistry with applications to investigation of chemical dynamics of biological systems. The goal of our research program is to understand fundamental laws that control chemistry in living cells. Primary tools used in our projects include multi-scale fragmentation methods, quantum master equation, and hybrid quantum mechanics/molecular mechanics approaches.

Computational and Theoretical Chemistry

Chemical Education

Quantum Dynamics of Excitation Energy Transfer in Complex Biomolecular Systems

Many chemical processes in biological systems involve molecular electronic excited states. These chemical processes range from energy transfer in photosynthetic proteins to photochemical transformations in artificially designed photo-switchable ion channels. My group develops and applies computational methods to reveal mechanisms of excitation energy transfer in molecular fluorescent probes and light-harvesting pigment-protein complexes. The obtained results are crucial for targeting and sensing specific DNA sequences, understanding the mechanism of photosynthesis, and development of artificial analogs of light-harvesting complexes and novel types of organic solar cells. Recent advances in QM/MM, ab-initio-based fragmentation, and multi-scale models open new opportunities for an atomistic-level description of excitation energy transfer in macromolecules. Our research program aims to elucidate the role of quantum phenomena in biological systems, in particular, the impact of molecular vibrations on energy transfer, interactions of light-absorbing pigments with solvent-protein surroundings, etc. Go to top

Development of Deep Learning Algorithms for Computational Chemistry

In the context of computational chemistry, deep learning is a class of machine learning algorithms that relies on multiple layers of data abstraction (defined by the neural network topology) with connections optimized during the network training process. Deep learning algorithms do not require pre-defined fitting equations or hard-coded descriptors. Machine learning algorithms have been successfully applied for computer-aided drug design, development and prediction of material properties, in quantum chemistry to model electronic quantum transport, to predict atomic energies and isomer energies of organic molecules, and quantum excitation dynamics. Currently, my group is working on the development of deep learning algorithms to predict interactions of macromolecules in their electronic ground and excited states to model energy transfer processes. The deep learning approach significantly accelerates computations and enables high-throughput screening of biomolecular systems by providing the ability to predict mechanisms of excitation energy transfer in pigment-protein complexes and molecular fluorescent probes. The development of new deep-learning-based computational chemistry methodologies is facilitated by recent advancements in GPU hardware architectures (e.g., CUDA) and the availability of cross-platform software libraries for deep learning (e.g., TensorFlow). The developed algorithms have been implemented in our original PyFREC software. Go to top

Multi-scale modeling of Dynamics of Neuronal and Cell Signaling

Synchronization of electrical activity in the brain is an example of non-linear spatial and temporal self-organization of cellular processes. Recently developed chemically engineered voltage-gated ion channels enable direct activation of biological neurons by light. This allows for the real-time noninvasive control and probing of neurons. My group is working on integration of biological neuron models (e.g. Hindmarsh-Rose) and QM/MM methodologies to simulate synchronization of light-sensitive neurons under normal and pharmacologically modified conditions. Go to top

Speeding-up computational chemistry using graphics processing units (GPUs)

Modern quantum chemical methods require enormous computer resources for handling molecular systems containing just a few hundred atoms. In addition to utilization of computer clusters and supercomputer systems we want to take advantage of using graphics processing units (GPUs) - devices that have hundreds of computational cores running in parallel (unlike conventional desktop computers that have only a few cores) to speed up computations. Go to top

Chemical Education

Integrating Chemistry Laboratory Instrumentation into the Industrial Internet

This project is designed to improve physical chemistry and instrumental analysis laboratory courses for undergraduate students by employing as teaching tools novel technologies in electronics and data integration using the industrial Internet. In the project students build, calibrate, and program an automatic titrator using an open-source microcontroller platform and standard pH probes, and integrate the instrument into the Internet. The proposed approach is flexible and can be used to enable integration of various laboratory instruments (e.g. temperature, pressure, salinity probes etc.) into the industrial Internet. The hardware and software are open-source which makes possible further modification and development by the academic community.

For more details see:

Nicole Famularo, Yana Kholod, and Dmytro Kosenkov, "Integrating Chemistry Laboratory Instrumentation into the Industrial Internet: Building, Programming, and Experimenting with an Automatic Titrator", J. Chem. Educ., 2016, 93(1), 175-181. DOI: 10.1021/acs.jchemed.5b00494

Software Download

Circuit schematics
11/15/2016 updated_scheme.pdf  
Go to top

Transient Absorption Spectroscopy with 3D-Printed Temperature Controlled Sample Holder

The laboratory unit demonstrates a project-based approach to teaching physical chemistry laboratory where upper-division undergraduates carry out a transient-absorption experiment investigating kinetics of cis-trans isomerization of N, N-dimethyl-4, 4-azodianiline. Students participate in modification of a standard flash-photolysis spectrometer adding a temperature controlled sample holder. The sample holder is open-source and can be reproduced with 3D-printing technology. Students build and use the experimental setup. Based on observer kinetics of cis-trans isomerization students carry out data analysis estimating energy, enthalpy, entropy, and Gibbs free energy of activation.

For more details see:

Dmytro Kosenkov, James Shaw, Jennifer Zuczek, and Yana Kholod, "Transient Absorption Spectroscopy of Cis-Trans Isomerization of N, N-Dimethyl-4, 4-Azodianiline with 3D-Printed Temperature Controlled Sample Holder", J. Chem. Educ., 2016, (ASAP) DOI: 10.1021/acs.jchemed.5b00494

3D-Printable Design Download
Go to top