Modeling Semi-Batch Vinyl Acetate Polymerization Processes

Abstract

Poly­(vinyl acetate) (PVAc) is a polymer of high industrial importance. The final properties of PVAc and products thereof are strongly dependent on its microstructure, which, in turn, is determined by the specific polymerization conditions and processes used for its production. In silico modeling approaches based on kinetic Monte Carlo simulations are of high interest since they can enable the prediction of the microstructural characteristics of the resulting polymer chains, as long as the model considers the specific (and complex) polymerization and process conditions. In this study, a robust and versatile kinetic Monte Carlo model was developed, allowing for the treatment of semi-batch radical polymerizations of vinyl acetate in methanol under reflux conditions. The kinetic model comprises a full kinetic scheme for the initiation, termination, propagation, and transfer reactions. The majority of the required rate coefficients for the elementary reactions is available from the literature. With respect to the diffusion-controlled termination reaction, the composite model, which represents the chain length dependence (CLD) of k t, has been extended to account for the polymer content in the system. The robustness of the kinetic model is demonstrated by the very good agreement between the experimental and calculated molar mass distributions of PVAc obtained at different reaction times and regardless of the feeding profiles used.

Kinetic Monte Carlo Simulations as a Tool for Unraveling the Impact of Solvent and Temperature on Polymer Topology for Self‐Initiated Butyl Acrylate Radical Polymerizations at High Temperatures

“…[ 2 ] To estimate the missing rate coefficients experimental polymerization results are compared to in‐silico derived data by Monte Carlo simulation and the Metropolis–Hastings algorithm is applied for parameterization. [ 50 ] The kinetic model is given in Scheme and the associated kinetic rate coefficients are listed in Table 1 . The complete set of elemental reactions required for kMC simulation of the self‐initiated polymerization of BA in bulk and solution comprises 50 elemental reactions.…”

“…The automated parameter search was executed by a Metropolis Hastings algorithm, which has already been used before. [50] A detailed description of the strategy is provided by Feuerpfeil et al [50] Table 1. Full set of kinetic coefficients used in kMC simulations.…”

“…Previously, the simulator was successfully applied to model, for example, butyl acrylate polymerizations, seeded emulsion polymerizations of styrene or reversible deactivation radical polymerization. [34,39,50,51] Additional details of the simulator were provided elsewhere. [52]…”

Polymer Reaction Engineering meets Explainable Machine Learning

“…Since then improvement in computing power and coding allowed for the application of the algorithm to more and more complex systems (Trigilio, Marien, van The kMC simulations were conducted using the in-house created open source kMC simulator mcPolymer (Drache & Drache, 2012), which is based on a full kinetic scheme with all elemental reactions occurring. Previously, the simulator was successfully applied to model, e.g., reversible deactivation radical polymerizations (Drache, 2009;Drache & Drache, 2012), acrylate polymerizations with backbiting and transfer to polymer reactions (Drache et al, 2015), or semi-batch vinyl acetate polymerization (Feuerpfeil et al, 2021).…”

Quantitative Structure–Property Relationship Model for Predicting the Propagation Rate Coefficient in Free-Radical Polymerization

“…The model-based regression analysis from kinetic data, as an alternative computational approach, enables researchers to calculate the intrinsic/apparent rate coefficients in FRP (Figure b). − The method includes deterministic and stochastic-based approaches. − The former is achieved through solving a set of algebraic/differential equations derived from mass balances, while the latter only requires the reaction probabilities . The method of moments (MoM) − and kinetic Monte Carlo ( k MC) − are the two most widely used kinetic modeling methods. More recently, Wu et al developed the distribution–numerical fractionation–method of moments (D–NF–MoM) model, which is capable of predicting bivariate distributions of macrospecies and bring the MoM deterministic solver a step forward .…”