Embedded Files
I received my Ph.D. in physical chemistry at Rice University.
I am the leading developer of SchrodingerWave software.
Please feel free to check my ResearchGate and Linked-In profiles:
A) Studying Nano-Bio Interface by Molecular Modeling and Spectroscopy
A) Studying Nano-Bio Interface by Molecular Modeling and Spectroscopy
DOI: 10.1021/acs.jpclett.3c02100
Single-walled carbon nanotubes (SWCNTs) can be covalently functionalized, altering their spectroscopic and photophysical behavior. We present a structure-resolved Raman metric to quantify functionalization: covalent sp³ defects induce intermediate-frequency modes (IFMs, ~350–650 cm⁻¹) whose frequencies vary with nanotube diameter and excitation wavelength. As IFMs grow, radial-breathing-mode (RBM) intensities decline; the IFM/RBM intensity ratio thus provides a sensitive, chirality-specific measure of functionalization.
DOI: 10.1002/cphc.201601000
Amyloid-β (Aβ) fibrillation is a pathological hallmark of Alzheimer’s disease. Because bio–nano interfacial interactions modulate Aβ aggregation, clarifying peptide–surface interactions at the molecular level is essential for assessing nanoparticle neurotoxicity. Here, atomistic molecular dynamics simulations of the fibril-forming fragment Aβ₍₂₅–₃₅₎ on Au and Ag (111) and (100) facets show that adsorption stabilizes β-sheet–rich conformations, increasing aggregation propensity and suggesting a mechanism for nanoparticle-accelerated fibrillation. To quantify single-peptide β-sheet content, we introduce a Ramachandran-based metric.
DOI: 10.1021/acs.jpcb.5c02824
In guanine functionalization, ssDNA-wrapped single-walled carbon nanotubes (SWCNTs) form covalent bonds at guanine bases; the resulting electronic and optical shifts depend on inter-defect spacing. Using replica-exchange with solute tempering molecular dynamics, we modeled (GT)₁₀ ssDNA conformations on SWCNTs and the pre-reaction distribution of guanine positions. Simulations varied interstrand interactions, nanotube end effects, ionic strength, DNA/SWCNT mass ratio, and tube diameter. From these, we quantified axial spacing distributions of guanine nucleobases. Irregular spacings are predicted to induce excitonic energy disorder and contribute to spectral broadening in guanine-functionalized SWCNTs.
B) Deep Learning to Predict Physical Properties of Chemical Systems
B) Deep Learning to Predict Physical Properties of Chemical Systems
DOI: 10.1016/j.fluid.2013.07.004
We estimate polymer specific volumes using a hybrid approach that combines an artificial neural network (ANN) with a simple group-contribution method (GCM). The model was trained, validated, and tested on 2,865 data points across multiple temperatures and pressures for 25 polymer systems, yielding average relative deviations of 4.03% (training), 4.39% (validation), and 4.82% (test). Compared with prior methods, the ANN–GCM offers a simple, accurate procedure that remains consistent with experiment even under high-temperature, high-pressure conditions.
We predict hydrocarbon densities using a hybrid model that couples an artificial neural network (ANN) with a simple group-contribution method (GCM). Trained, validated, and tested on 2,891 measurements across a wide temperature–pressure range for 40 hydrocarbons—including short- and long-chain alkanes (CH₄ to n-C₄₀H₈₂), cycloalkanes, highly branched alkanes, and aromatics—the ANN–GCM provides accurate, easy-to-use estimates. Compared with prior correlations, it maintains strong agreement with experiment, including under high-temperature, high-pressure (HTHP) conditions.
Page updated
Google Sites
Report abuse