Ceramics and Glasses Modeling by Simulations and Machine Learning: Simulation of Glass and Ceramics
Sponsored by: ACerS Glass & Optical Materials Division
Program Organizers: Mathieu Bauchy, University of California, Los Angeles; Peter Kroll, University of Texas at Arlington; N. M. Anoop Krishnan, Indian Institute of Technology Delhi
Monday 2:00 PM
October 10, 2022
Room: 408
Location: David L. Lawrence Convention Center
Session Chair: Mathieu Bauchy, UCLA; Peter Kroll, The University of Texas at Arlington; Anoop Krishnan, IIT Delhi
2:00 PM Invited
Machine Learning to Design and Discover Sustainable Cementitious Binders: Learning from Small Databases and Developing Closed-form Analytical Models: Aditya Kumar1; Taihao Han1; 1Missouri University of Science and Technology
To reduce the carbon footprint of Portland cement, the prevailing practice embraced by concrete technologists is to partially replace the cement in concrete with supplementary cementitious materials (SCMs). Chemistry of the SCM profoundly affects all subprocesses leading up to property development in concrete. Owing to the substantial diversity in SCMs’ compositions, computational models are unable to produce a priori predictions of properties of [cement + SCM] mixtures. This study presents a deep learning (DL) model capable of producing a priori, high-fidelity predictions of cement hydration kinetics and phase assemblage development in [cement + SCM] mixtures. The DL is coupled with: (1) A fast Fourier transformation algorithm that reduces the dimensionality of training datasets; and (2) A thermodynamic model that constrains the DL. The training of the DL is leveraged to develop a closed-form analytical model capable of predicting cement hydration kinetics in [PC + SCM] mixtures.
2:30 PM
Developing ReaxFF for Simulation of Silicon Carbonitride Polymer-derived Ceramics: Shariq Haseen1; Peter Kroll1; 1University of Texas at Arlington
Polymer-derived ceramics (PDCs) exhibit desirable properties such as enhanced mechanical properties at high temperatures, oxidation resistance, and use as prospective anodes in lithium-ion batteries. In order to investigate the polymer-to-ceramic conversion process, we develop a reactive force field (ReaxFF) for large-scale simulations of the pyrolysis of polysilazanes into SiCN ceramics. The approach promises to achieve quantum-chemical accuracy while maintaining fast calculations. Parameter optimization of ReaxFF is done through energy and force matching of static structures, for which we built an extensive library of crystalline and amorphous models. Trajectories obtained through ab initio molecular dynamic simulations are also used for parameter optimization. We use our final Si-C-N-H ReaxFF parameters to investigate the thermal conversion of polymers to PDCs. During elevated temperature simulations, we monitor the formation of gaseous species and identify chemical reactions. The resulting amorphous SiCN ceramics are analyzed to further elucidate the genesis of embedded carbon structures.
2:50 PM
Molecular Dynamics Simulation of Tellurite Glasses: Amreen Jan1; N M Anoop Krishnan1; 1Indian Institute of Technology Delhi
Tellurium oxide-based glasses are among the most promising candidates for integration in non-linear optical devices. However, these glasses are known to be prone to fast devitrification and hence, it is necessary to understand their formation and properties to be able to take full advantage of their peculiarity. Though a lot of experimental work has been carried out in the field of tellurite glasses but these glasses have not been explored much using atomistic modelling. In this work Tellurium oxide glass has been simulated, using the interatomic potentials based on the framework of Born model of ionic solids and further, core shell model. This work explores the effect of system size, quenching rate, and ensemble (NPT and NVT) in terms of connectivity, short-range and medium range order. The glass is estimated to be sensitive to the quenching rate and ensemble choice but not much to the system size.
3:10 PM
Molecular Dynamics Study of Domain Switching Dynamics in KNbO3 and BaTiO3: Rajan Khadka1; Pawel Keblinski1; 1Rensselaer Polytechnic Institute
We use molecular dynamics simulations to investigate the polarization switching dynamics in the single domain and 180o bidomain (i.e., including the preexisting domain wall) models in BaTiO3 and KNbO3. In a single domain study, for both materials, we observed that the hysteresis loop is essentially non-existent in the highest temperature non-cubic phase. We attribute this behavior to the observation of spontaneous local polarization fluctuations leading to the elimination of the nucleation barrier. Interestingly, in the case of the bidomain structure, while we observe domain migration driven by the electric field, at high fields new domain nucleation is severely suppressed by comparison with the single domain simulations. This behavior is explained by the suppression of simulation dimension fluctuations due to the presence of the two domains. We further demonstrate that artificial suppression of the simulation cell dimension fluctuations in the case of a single domain switching also suppresses new domain nucleation.
3:30 PM Break
4:00 PM
In-Silico Simulations of Polymer Pyrolysis: Peter Kroll1; 1University of Texas at Arlington
Chemical reactions during thermal processing of hybrid organic-inorganic polymers transform the precursor into a ceramic. Here we present details – atom trajectories, principal mechanisms, and their outcomes – of fundamental reactions in pyrolysis of silicon-based polymers obtained from quantum-chemical (ab-initio) Molecular Dynamic Simulations. We observe intra-chain and inter-chain coupling, cross-linking, and elimination reactions. Furthermore, we observe Kumada-type reactions that incorporate carbon into the polymer backbone.
4:20 PM
Pore-resolved Simulations of Chemical Vapor Infiltration in 3D Printed Preforms and the Kinetic Regimes: Mengnan Li1; Vimal Ramanuj1; Ying She2; Ramanan Sankaran1; 1Oak Ridge National Laboratory; 2Raytheon Technologies Research Center
With the rapid advancement of computing capability, pore-resolved direct numerical simulations (DNS) have become feasible for studying the transport and kinetic regimes encountered in chemical vapor infiltration (CVI). We present DNS simulations using a level set approach to capture the complex topology between the vapor and solid phases during the densification process. The simulations were performed using a finite rate kinetic model for the deposition of silicon carbide (SiC) from a methyltrichlorosilane (MTS) precursor. We present the computational model and the simulation approach. The model was applied to simulate densification of 3D printed preforms and conduct a parametric study to investigate the role of competing transport and reaction processes. Results are presented to show the evolving non-uniform porosity, structure functions and the distribution of residual voids.