Computer keyboard and laboratory glass in blue light. Science concept. Copyright: isak55@Shutterstock

Decarbonisation from a variety of industrial and power emission sectors highlights a marked need for capture technologies. Advanced tailored sorbent-based technologies offer higher capture capacities than the state-of-the-art systems. The PrISMa project addresses this challenge by integrating materials design with process design and environmental considerations.

The choice of enabling materials in sustainable chemical processes directly influences the overall performance. Computer-aided molecular and process design is the perfect tool to determine the optimal process configurations, operating conditions, and material structures. It's current state-of-the-art is discussed here.

This work presents a perturbed-chain statistical associating fluid theory (PC-SFAT)-based solvent selection and process optimization workflow for solution crystallization processes with readily executable computational tools that are usable with minimal expertise. The application of the workflow is illustrated via a case study on aspirin crystallization.

Molecular simulations are used to calculate energy profiles of the cavities in methane hydrates and to derive thermodynamic material parameters. The aim is to obtain non-measurable material parameters comparatively easily in this way and thus to be able to parameterize material data models more universally.

The article presents an evaluation method using an experimental and simulation-based analysis of refrigerant circuits to generate more reliable design knowledge for low refrigerant charge applications. The results show promising steps towards a reliable prediction of capacity, coefficient of performancet, and refrigerant charge for circuits with 100 g to 300 g of propane.

Application of machine learning to membrane polymer design is presented. Artificial neural networks were used to predict the physical properties of polymers. This study found that these networks were capable of predicting these properties with high accuracy.

The integrated design of solvents and processes links molecular structures to process performance by predictive assessment across multiple scales. The presented method solves the inverse problem by combining process flowsheet optimisation including heat integration with molecular design. Thereby, we can select molecules that are optimal regarding process-level objectives.

Novel algorithms for exploring Pareto sets are applied to distillation processes. Product purities are maximised while total heat duties are minimised. The illustrated Pareto set consists of both convex (blue dots) and non-convex (red dots) regions. This article shows how such Pareto sets can be navigated interactively using a ray tracing technique.

To solve the simultaneous process and fluid optimization problem for industrial applications, the two-stage CoMT-CAMD approach is implemented in CHEMASIM, the in-house flowsheet simulator of BASF SE. The required modifications of the method and the extension of the user interface of CHEMASIM results are presented for single and multiobjective optimization problems of process-fluid selection and design.

The working domain of different NH₃/H₂O heat pump cycles is investigated to provide temperatures up to 150 °C. An algorithm is developed, which maximizes coefficient of performance and volumetric heat capacity and minimizes compressor discharge temperature for given external temperatures and internal pressure limits.

A computer-aided molecular and process design framework is successfully extended with a detailed process model for a hybrid extraction-distillation process. The framework is capable to design solvents with superior performance compared to solvents identified by common methods.

This work proposes a systematic framework to search for phase change materials that are sustainable-by-design without compromise on their functional performance. Computer-aided molecular design tools, a simple process simulation model, and QSPR models are applied to predict thermophysical properties, ecotoxicity and climate change indicators.

Fragrance is a desirable and often essential attribute in various consumer products. To predict fragrance attributes of molecules, a novel machine learning based methodology has been developed. Rough set-based machine learning is used to link fragrance to molecular structure. The factors affect the perception of fragrance has also been analysed.

A novel computer-aided crosslinked polymer design framework is proposed, in which machine learning and molecular dynamics are integrated for the design of an optimal polymer structures. Two case studies are presented to illustrate the application of the proposed CAPD framework, and the optimized results are verified.

Suitable working fluids for high temperature heat pumps are identified from a COMT-CAMD approach. PC-SAFT and a realistic model for the ideal gas heat capacity are applied as thermodynamic model. Results are presented for a one-stage and a two-stage heat pump cycle. Cyclobutane and butene isomers are identified as promising working fluids.