Engineering Papers⌕ Search

Engineering topics

Crowley, Michael

Publications and source records attributed to Crowley, Michael.

Investigating the Effect of Water on the Mechanical Properties of Cellulose from Multiscale Molecular Dynamics Simulations

Classical molecular dynamics (MD) simulations provide insight into the structure and physicochemical properties of materials with atomic resolution. However, the length and time scales accessible to atomistic MD are orders of magnitude smaller than many relevant processes such as the response of a bulk material to experimentally accessible strain rates, which presents challenges when comparing models to experimental measurements. Bottom-up coarse-graining provides a means for systematically mapping atomistic information to lower resolution models to increase the length and time scales achievable by simulation. Cellulose is an abundant carbohydrate biopolymer with applications to many fields of research, such as materials science and renewable energy, due to its desirable mechanical properties and viability for conversion into biofuel. The effect of moisture content on the Young's modulus of cellulose is of special interest due to its native environment often being in the hydrated secondary plant cell wall and the grinding energy requirements for biomass feedstock preprocessing. The current work investigates the effects of water solvent on the Young's modulus of cellulose calculated from coarse-grained MD mechanical stress simulations. The coarse-grained model was parametrized from atomistic MD calculations of cellulose-cellulose potentials of mean force using umbrella sampling techniques under vacuum and solvated conditions. The Young's moduli of the coarse-grained cellulose assemblies parametrized from cellulose in vacuum or solvated in water were computed via mechanical stress simulations to highlight the importance of capturing solvent interactions for modeling the mechanical behavior of cellulose.

BASIC BIOLOGICAL SCIENCES,RADIATION PROTECTION AND↗

mesoflow [SWR-22-56]

Mesoflow is a continuum scale simulation tool developed specifically for modeling transport and chemistry at the mesoscale. Our solver utilizes Cartesian block-structured adaptive mesh refinement to resolve complex surface morphologies (of catalysts/biomass particles among others) directly obtained from X-ray tomography data. An immersed boundary based formulation enables rapid representation of complex geometries prevalent in most mesoporous interfaces. The solver is developed on top of open-source performance portable library, AMReX, providing parallel execution capabilities on current and upcoming high-performance-computing (HPC) architectures. Our flexible software framework enables integration of complex chemical mechanisms at heterogenous interfaces and time-split algorithms for circumventing highly disparate reaction and flow time-scales. Our current studies indicate a ten-fold performance gain by using graphics-processing-units (GPU) compared to a single processor for representative problem sizes (2 million cell mesh).

Sitaraman, Hariswaran↗

Coupling of Flavonoid Initiation Sites with Monolignols Studied by Density Functional Theory

Lignin recalcitrance presents a challenge for the development of a bioeconomy that employs lignocellulosic feedstocks. The efficiency of lignin deconstruction is improved by a reduction of molecular weight, and given the discovery that flavones serve as initiation sites in lignin biosynthesis, these molecular weight reductions could potentially be achieved with plant metabolic engineering to over-express flavonoids. Upon increasing the flavonoid content in lignin, the bond strengths and properties of flavonoid-monolignol linkages become increasingly important. To that end, the current work applies density functional theory calculations to elucidate the bond dissociation enthalpies (BDEs) of flavonoid-monolignol linkages, including dimers with oxidized monolignols. Specifically, the dimer bond strengths and monomer hydrogen abstraction energies for the flavonoids tricin, chrysoeriol, luteolin, apigenin, catechin, epicatechin, epigallocatechin, and epigallocatechin gallate are calculated, when coupled to seven natural and engineered monolignols. Results indicate that 4'-O-ß linkage strengths between flavonoids and monolignols are of comparable strength to inter-monolignol ß-O-4 linkages, with average flavonoid-monolignol BDEs of 70.7 kcal/mol relative to ~69.3 kcal/mol in analogous canonical monolignols. Epigallocatechin yielded the lowest 4'-O-ß bond strength of 52.3 kcal/mol when coupled to an oxidized monolignol, while the flavones overall produced lower average BDEs, relative to the flavanols. Substituents at the 3'-C and 5'-C positions on flavonoids affected the dimer linkage strengths to a greater extent than glycosylation or substituents further from the linkage. Erythro and threo stereochemistry across the flavonoid-monolignol linkage library exhibited only small energetic differences and no pronounced correlations. Taken together, the predictions from this work support the concept that higher concentrations of flavonoid initiation sites in lignin may afford linkage properties conducive to more facile lignin depolymerization.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

m2p (Monomers to Polymers)

The software is Python-based code that takes monomers, represented as SMILES, and "reacts" them together to form long polymer chains, represented as SMILES, based on known polymer chemistries. The software is used in a machine-learning pipeline where a database and machine-learning approaches are used to predict polymer properties base on molecular structure. This software is necessary to represent the polymer chain, instead of simple repeat units, as the software statistically represents different molecular architectures that occur during reactions (e.g. head-to-head, head-to-tail, and tail-to-tail). It is capable of building polymer chains with five different chemistries: vinyl/olefin, polyamide, polycarbonate, polyester, and polyimides. The user is able to specify different monomer pairings and degree of polymerization for different polymer chains.

Wilson, Nolan↗

A Molecular-MNIST Dataset for Machine Learning Study on Diffraction Imaging and Microscopy

An image dataset of 10 different size molecules, where each molecule has 2,000 structural variants, is generated from the 2D cross-sectional projection of Molecular Dynamics trajectories. The purpose of this dataset is to provide a benchmark dataset for the increasing need of machine learning, deep learning and image processing on the study of scattering, imaging and microscopy.

BASIC BIOLOGICAL SCIENCES,OTHER INSTRUMENTATION↗

Lignin Forcefield Source Datafiles

Three archives are provided here comprising the development of a CHARMM36-compatible lignin forcefield suitable for use in biomolecular simulations. One archive parameters.zip contains the output topology and parameter files in CHARMM format suitable for use in constructing simulation systems. The remaining two hold the logic of the parameterization featuring scripts that do the optimization including the source for the GPU-accelerated objective function evaluation and target data creation scripts.zip as well as minimal outputs needed to recreate the work datafiles.zip. Due to licensing restrictions Gaussian log files that created the target data are not provided here. Instead only binary-formatted parsed target data as well as the input decks used to create the target data are provided in datafiles.zip. Readme files for individual directories created within the directory structure are contained within scripts.zip.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗