Special Session C-8
Exploiting Computational Tools in Materials Manufacturing and in the User Industry
ABSTRACTS
Session C-8.1 Metal and metal alloys manufacturing and user industry
C-8.1:L01 Phase Selection in 316L Processed by Laser-powder Bed Fusion
C.-A. GANDIN, G. GUILLEMOT, P. MARTIN, MINES Paris, PSL University, CEMEF UMR CNRS 7635, CS10207, Sophia Antipolis, France; P.W. VOORHEES, C.A. HARELAND, Department of Materials Science and Engineering, Northwestern University, Evanston, IL, USA
A generalized model for the kinetics of a dendrite tip with a non-equilibrium interface is presented for multicomponent alloys. The model includes full coupling with thermodynamic databases to account for non-dilute non-ideal solutions, in addition to a full diffusion matrix in the liquid. The consequences of the computed non-equilibrium phase diagram boundaries on the dendrite tip kinetics are considered, and substantial deviations from existing theories are observed, especially in the case of zero solute drag. The model is applied to the rapid solidification of the 316L stainless steel containing five solute species. The phase diagram properties (i.e., partition coefficients and liquidus slopes) are directly visualized as a function of velocity and observed to vary non-linearly and, in some cases, non-monotonically due to the combination of non-linear phase diagrams and kinetic effects. Recent reports in the literature on the competition between the growth of ferrite and austenite from the melt are revisited with these new developments.
Session C-8.2 Ceramics, glass, cement and user industry
C-8.2:IL01 Machine Learning of Phase Diagrams
J. Lund, H. Wang, R. Braatz, R.E. García, Purdue University, West Lafayette, IN, USA
By starting from experimental- and ab initio-determined phase diagrams (PDs) of materials, a machine learning (ML) method is developed to infer the free energy function for each phase. The ML method samples the multidimensional space of Gibbs free energy parameters and user-defined physical constraints into a database of millions of PDs in order to identify the target material properties. The method presented herein is 1000x to 100,000x faster than currently available approaches, and defines a new paradigm on the quantification of properties of materials and devices. The developed methodology is combined with the most widely used thermodynamic models – regular solution, Redlich-Kister, and sublattice formalisms– to infer the properties of materials for a few lithium-ion battery applications, reconstructing without human bias, well-established CALPHAD formulations while identifying previously missed stable and metastable phases and associated properties.
C-8.2:IL02 Chemical-Reaction-Induced Wear Process Simulations of Carbon- and Silicon-based Solid Materials
YANG WANG, Research Institute of Frontier Science, Southwest Jiaotong University, Chengdu, China
Wear of carbon- and silicon-based solid materials are of importance for a large variety of fields such as solid lubrication, semiconductor, and manufacturing tools. For example, amorphous carbon is a promising solid lubricant coating in many engineering surfaces, whereas the wear in high vacuum is innegligible obstructing it from utilizing in space crafts; on the other hand, manufacturing of silicon chips should experience many rounds of chemical-mechanical polishing processes in which materials’ wear is the fundamental issue. For either carbon- or silicon-based solids, there is a common feature that the both friction and wear properties are dominated by the interfacial chemical reactions, including interfacial bond formation, shear-induced structural transformation, and etc. Experimental approaches are difficult to fully reveal these reaction dynamics and chemical-reaction-induced wear mechanisms, whereas atomistic simulation is a promising tool to well handle the chemical reaction dynamics during the complicated friction/wear process. This talk would like to show how powerful atomistic simulation is for revealing the chemical reaction dynamics and the relevant wear mechanisms, using some recent research advances as examples.
C-8.2:L03 Multi-scale Simulation Approach for Exploring Optimized Electrode Structure of Dye-sensitized Solar Cell Devices
M. ONODERA1, M. Kubo1, 2, 1Institute for Materials Research, Tohoku University, Sendai, Japan; 2New Industry Creation Hatchery Center, Tohoku University, Aramaki, Aoba-ku, Sendai, Japan
Features of Dye-Sensitized Solar Cells (DSSC) are low-cost production, wide variety of designs, independence of installation sites, etc. In order to contribute for the better understanding of the energy conversion mechanisms and precising design optimization, we have developed our original multi-scale DSSC simulator, and then have continuously extending the function and predictability of this simulator. This method enables the evaluation of the I-V characteristics for DSSC devices at macro-scale from the nano-scale properties and meso-scale structures of applied materials, which is achieved by integrations of each scale of simulators; The nano-scale simulation reveals physical properties and electronic structure of materials through quantum chemical calculations. The meso-scale simulation evaluates diffusion coefficients of electrons and ionic species inside the network of TiO2 particles and the electrolyte. The macro-scale simulation gives the I-V characteristics for DSSC devices using the data from the lower-scale simulations. In this study, we performed simulations on several electrode materials with changing micro- and meso-parameters. Consequently, the insights for designing the structure of TiO2 network were obtained to improve electron transport characteristics.
Session C-8.3 Polymers and related materials and user industry
C-8.3:IL01 Tailoring Molecular Topology to Control the Mechanical Properties of Polymeric and Nanoparticle Networks
S. KETEN, Dept. of Mechanical Engineering, Dept. Civil and Environmental Engineering, Northwestern University, Evanston, IL, USA
Tailoring Molecular Topology to Control the Mechanical Properties of Polymeric and Nanoparticle Networks In this talk, I will summarize recent advances in computational design of new macromolecular materials that make use of nanoscale topologies, such as brushes, networks, and folded loops, that result in exceptional mechanical properties. I will first present physics-based and data-driven reduced order modeling approaches that were developed to describe molecular and mesoscale mechanics of polymers and polymer-grafted nanoparticle systems. Following this, I will present strategies for achieving higher strength, toughness, and impact tolerance in soft materials. The first strategy involves the use of polymer grafted nanoparticles to improve diametrically opposed mechanical properties such as modulus and toughness, while also controlling the time-dependent characteristics of the response. The second strategy involves creating nanoparticle interfaces with looped tethers that take inspiration from catch bonds in biological adhesion proteins, which results in molecular seat-belt type interfaces that self-strengthen at high strain rates like shear-thickening fluids. I will conclude with some thoughts on how to translate these findings to new material concepts.
C-8.3:IL02 Alternative Low Carbon Fuel: a Molecular Modeling Investigation on Corrosion Inhibition
S. LOEHLE, TotalEnergies OneTech, Solaize, France; A. SALCEDO, S. STEINMANN, C. MICHEL, ENS, Lyon, France
The use of ammonia as an alternative fuel in internal combustion engines for the maritime transport is gaining more and more interest due to the current climate objectives of reducing CO2 emission and for environmental concerns. Still ammonia combustion poses major challenges, one being the corrosion of metallic engine parts. The aim is to better understand and classify the performance of additives typically present in engine lubricants as metal corrosion inhibitors, by using a computational modeling approach. Computational modeling is a powerful tool since it allows to investigate at the molecular level the diffusion behavior of corrosive species and to perform a faster screening of different organic additives. Molecular Dynamics (MD) simulations and Thermodynamic Integration (TI) are employed, with improved potentials for the metal/heteroatom (nitrogen, oxygen) interactions description, based on Density Functional Theory (DFT) level of accuracy [1]. This enables an overall more realistic and accurate resolution of the metal/organic films structures, compared to already existing force fields. Based on the calculations it has been possible to determine which additive molecule performs best against diffusion of corrosive species, towards hematite (Fe2O3) and copper (Cu) surfaces.
[1] J. Rey et al, J. Phys. Chem. B 2021, 125, 10843−10853
C-8.3:L03 Meso-scale Proton and Oxygen Diffusivity Analysis in Cathode Catalyst Layer towards Boosting Polymer Electrolyte Fuel Cell Performance: Large-scale Reactive Molecular Dynamics Simulations
tetsuya nakamura, K. Mori, S. Shogo, Y. Su, Y. Asano, Y. Ootani, N.Ozawa, M. Kubo, Institute for Materials Research, Tohoku University, Sendai, Miyagi, Japan
Boosting polymer electrolyte fuel cell (PEFC) performance is required to reduce CO2 emissions. PEFC’s output depends on the electrode reaction activity, which involves the proton and oxygen diffusivity in catalyst layer (CL) consisting of carbon supports, Pt nanoparticles (NPs), Nafion, and water. To boost CL performance, molecular dynamics (MD) method is a promising approach for proposing design principles of CL structures. However, conventional MD method cannot simulate large-scale CL structures due to its high computational costs. Then in the present study we challenged to simulate the large-scale meso-scale CL structures by reactive MD methods and investigated the influence of meso pores in the carbon supports on the meso-scale proton and oxygen diffusivity. Consequently, more protons were adsorbed on the Pt NPs interior of the meso pores than those on the Pt NPs exterior of the meso pores. We found that less Nafion is distributed around the Pt NPs interior of meso pores, and then the protons desorbed from the Pt NPs do not adsorb directly to Nafion, but adsorb back to the Pt NPs. These results suggest that optimizing Nafion distribution leads to high proton diffusivity. Furthermore, we will discuss CL structures to accelerate meso-scale oxygen diffusivity in the conference.
C-8.3:IL04 Coarse-grained Modeling of Thermosets: A General Machine Learning Approach to Tunable Force-Fields
A. GIUNTOLI, University of Groningen, Groningen, Netherlands; A. van Beek, University College Dublin, Dublin, Ireland; Nitin Hansoge, 3M, Minneapolis, USA; T. W. Sirk, Army Research Lab, USA; S. Pal, K. Dansuk, W. Chen, S. Keten, Northwestern University, IL, USA
Predictive molecular simulations of thermosets for material design and manufacturing need both atomic-level resolution and large sizes to observe microscopic material behavior. Chemistry-specific coarse-grained models are suited for this task, but their force field development traditionally targets their atomic structure rather than dynamics and mechanical processes such as glass formation or deformations.
The Energy Renormalization coarse-graining scheme was developed to capture the glass-forming features of thermoplastics by introducing temperature-dependent parameters in the force field, but it is not easily transferred to the complex structure of thermosets, where crosslinks play an important role.
I will present our recent extension of the Energy Renormalization scheme to build chemistry-specific, transferable force fields for a coarse-grained model of epoxy resins with varying crosslink density and conversion rate. We employed Gaussian process metamodels to simultaneously calibrate the large number of force field parameters, developing a machine learning approach to coarse-graining that is easily generalizable, and can in principle be applied to different problems and classes of materials.
C-8.3:IL05 Drilling of CFRPs Using Single Layer Diamond Tools
ZHONGDE SHI, M.H. Attia, National Research Council Canada, Montreal, Quebec, Canada
An investigation into drilling CFRPs using single layer diamond tools was carried out aiming at developing a viable drilling process. Single grain scratching tests and FEM analysis were first conducted to identify material removal mechanisms. This was followed by grinding tests to investigate the cutting forces and develop force models for process optimization and prediction of drilling-induced damage. Orbital drilling tests were finally conducted to validate the force model and to investigate the performance of tools with different geometries and diamond grain sizes in terms of cutting forces, temperatures, surface roughness, and machined part quality. It was found that the material removal mechanism depends on fiber and matrix failure characteristics, which are affected by fiber orientations. The cutting force model based on cutting kinematics, material removal mechanism, and fiber strength can predict forces within ±20% error. It was proved that orbital drilling of CFRPs with single layer diamond tools is a viable process compared with drilling using carbide tools for reducing cutting forces, temperatures, and extending tool life. The drawback of high surface roughness with single layer diamond tools was overcome with a new hybrid oscillatory orbital process.
C-8.3:IL06 Large-scale Molecular Dynamics Simulations for Deformation and Fracture Processes of Crystalline Polyethylene
YUJI HIGUCHI, Research Institute for Information Technology, Kyushu University, Fukuoka, Japan
For the improvement of the mechanical properties of semicrystalline polymers such as polyethylene which is the most produced plastics, understanding the deformation and fracture mechanisms at the molecular scale is essential. However, revealing their mechanical properties using molecular simulations was difficult due to the complex structure of semicrystalline polymers. By using coarse-grained molecular dynamics simulations, we have successfully constructed the lamellar structure, which is the basic structure of semicrystalline polymers. The mechanical properties are consistent with the experiments, confirming the validity of the simulation results. By developing these results, we have revealed the movement of polymer chain ends causing void generation, void growth processes using large-scale simulations, the relationship between stress propagation and molecular-scale structures such as tie chains and entanglements, and changes of molecular structures and crystal structures in irreversible processes. Our results advance our understanding of the relationship between the structure and mechanical properties of semicrystalline polymers at the molecular scale and will contribute to the molecular design to improve the mechanical properties of plastics.