(89f) High Pressure Reaction Kinetic Modeling: Leveraging Single Data-Rich Experiment

High pressure reactions play a pivotal role in various industries, ranging from pharmaceuticals to petrochemicals, where the manipulation of pressure conditions can significantly alter reaction kinetics, selectivity, and product yields.¹

Safety should not be taken lightly when considering high pressure reactions. A number of accidents has already shown how dangerous these types of reactions can be ². Another challenge in high pressure reactions is the potential for side reactions or unexpected reaction pathways under extreme pressure conditions. These reactions can result in the formation of undesirable secondary products, reduced selectivity and decreased overall process efficiency. Understanding the underlying chemistry and reaction mechanisms is important to mitigate the risk associated with these challenges and optimizing reaction conditions to maximize productivity. Performing high pressure reactions additionally creates operational difficulties, namely during sampling, illustrating the urge to implement more PAT tools and models to reduce operational burden.

In this study, an efficient approach for developing kinetic models for high-pressure reaction systems is described. Substantial value of employing just one data-rich experimental reaction allied with computational capability is demonstrated. Leveraging computational expertise, a digital twin was developed, comprising kinetics of a high-pressure reaction and Vapor-Liquid Equilibrium (VLE), with a key emphasis on temperature-dependent behavior.

(Figure 1. Process workflow from a digital twin to a complete model)

Described methodology aimed to build a kinetic model with just one reaction experiment. Through advanced computational systems, this single reaction becomes the cornerstone for adjusting vital kinetic parameters, such as rate constant and activation energy. Furthermore, to fully incorporate temperature dependency in the model, detailed influence on generated pressure was achieved with good accuracy. This temperature-dependent digital twin stands out by its predictive capability of simulating reaction behavior across a wide range of operating temperatures.

Due to reaction and experimental setup complexity, discrete sampling during reaction was challenging. Alternatively, in-line infrared (IR) spectroscopy was employed as Process Analytical Technology (PAT) tool, using ReactIR from Mettler Toledo. Prior to the reaction, an IR calibration curve was performed to determine concentration of starting material, 2-(methylthio)pyrimidin-4(3H)-one, and measure consumption during reaction execution.

(Figure 2. Reaction Scheme)

Reaction profile was previously predicted using available kinetic parameters of the same reaction with a different substrate. Reaction was slower than predicted, but after sequential temperature increases, higher reaction rates were observed. The kinetic constant was thereafter fitted, exhibiting a twenty-fold decrease compared to the initial proposed value. This effect was somewhat anticipated based on the structural differences between the studied molecule and the one used as reference. Following this experiment, the next step is to forecast the reaction at a desirable isothermal temperature and validate the model.

(Figure 3. Temperature and ReactIR profiles during reaction)

Our methodology represents an unconventional approach, demonstrating that a single experimental reaction coupled with computational capabilities can yield comprehensive insights into the kinetics of reactions, namely under pressure. This streamlined approach not only minimizes experimental resources but also accelerates the development of kinetic models, offering unprecedented efficiency and accuracy in understanding complex reaction systems.