Continuous-Molecular Targeting for Integrated Solvent and Process Design

Abstract

An integrated design of processes and solvents is a prerequisite for achieving truly optimized solvent-based processes. However, solving the full integrated problem in a single optimization is usually not possible even for a predefined process topology due to the required discrete choices between molecular structures. Current approaches therefore mostly decompose the integrated problem into a process design and a molecular-design subproblem. The interaction between these subproblems is usually limited in practice, and a direct link between process performance and molecular characteristics of the solvent is not achieved. In this work, a novel methodology for the integrated process and molecular design problem is suggested where the discrete molecular decisions in the integrated design problem are circumvented by defining a hypothetical molecule. The approach is building upon a molecular-based thermodynamic model, where the parameters representing a molecule are treated as continuous. These parameters are optimized together with other process parameters, leading to an ideal hypothetical target molecule (represented by a set of parameters) and a corresponding optimized process. Only in a subsequent step, the parameters of the thermodynamic model representing the hypothetical molecule are mapped onto an existing optimal solvent. The method is illustrated for the design of solvents for carbon dioxide capture where the benefits of the integrated design approach are demonstrated. The perturbed-chainpolar-statistical-associating-fluid theory (PCP-SAFT) equation of state is used as a thermodynamic model. The framework introduced is generic in nature and thus applicable beyond the study of solvents to the integrated design of materials and processes in general.

Group contribution methodology based on the statistical associating fluid theory for heteronuclear molecules formed from Mie segments

“…Electronic mail: a.galindo@imperial.ac.uk led to novel applications, such as the integrated design of solvents and processes, where molecular characteristics of solvents are determined as part of the optimization of the process. [4][5][6][7] An important aspect in the development of thermodynamic methodologies is a predictive capability, which is commonly perceived as the ability to provide predictions of phase behavior and other bulk properties without the need for experimental data for the determination of molecular model parameters.…”

Conceptual Design of Azeotropic Distillation Processes

“…Entrainer selection should not only be based on thermodynamic properties and heuristics, which may even contradict each other [21], rather, the selection of promising entrainer candidates should be integrated with the evaluation of candidate process structures [22]. Recently reported simultaneous approaches to the design of entrainer-based processes include property clustering [23] and hypothetical molecules [24]. In the latter approach, for example, a thermodynamic model, which links molecular structure to the properties of a hypothetical entrainer molecule, allows for a simultaneous optimization of process structure and entrainer.…”

Outer approximation algorithm with physical domain reduction for computer‐aided molecular and separation process design

“…Within this class of approaches, the CAMD step can be performed via the use of "molecular property clusters," 21 by using algebraic methods, 22 or molecular signatures. 23 In another method first proposed by Bardow et al, 24,25 continuous molecular targeting (CoMT), the continuous descriptors of an optimal (hypothetical) solvent, representing the parameters of the PC-SAFT equation of state, 26 were first determined based on process performance. This was then used to identify an optimal molecule with similar descriptors, from a list of compounds 25,27 or more recently by deploying CAMD techniques to derive a CoMT-CAMD methodology.…”