Environmental, health, and safety assessment of chemical alternatives during early process design: The role of predictive modeling and streamlined techniques

Doctoral thesis, 2022

Industrial chemicals are important for many aspects of modern life, though they can be harmful to the environment and human health. Environmental or safety concerns identified during the early design and selection of chemicals could motivate choices as to safer alternatives and process setups. There is a growing interest in developing more rapid, and streamlined assessment methods to obtain a first indication of the potential impacts linked to the nature and use of industrial chemicals.

This work applies predictive modeling and streamlined techniques to estimate the potential environmental, health, and safety hazards associated with specific chemical structures. The assessment is performed during the design and selection of promising candidates for a particular process as part of the computer-aided molecular design (CAMD) and process setup. The case of phase-change solvents used for post-combustion carbon capture is examined. Furthermore, the refinement of predictive models through the incorporation of knowledge already existing in the field (prior knowledge) is investigated. A procedure for knowledge extraction from scientific articles that applies text mining is proposed.

The results show that incorporating impact assessment criteria into the CAMD facilitates the molecular design by enriching the Pareto front of candidates. The use of predictive models that estimate molecular properties, such as acute aquatic toxicity, bioconcentration, and persistency are found to support the identification of the optimal solvents for CO2 capture. Given the role of sustainability-related properties in tasks such as CAMD, the improved performance and the interpretability of the aquatic toxicity predictive models developed here and using prior knowledge are important. The process level assessment of the phase-change solvent systems indicated that phase-change solvent alternatives could provide benefits, not only in terms of reduced energy consumption but also lower impacts on human health and the environment.  However, the degradation behaviors of these compounds should be properly assessed and controlled to ensure beneficial performances compared to conventional carbon capture solvents. Overall, predictive modeling and streamlined life-cycle assessments (LCAs), as well as environmental, health, and safety evaluation methods were revealed to be valuable for defining the critical aspects that influence the potential impacts of chemicals and in supporting decisions concerning the molecular and process designs.