How to use QSPR‐type approaches to predict properties in the context of Green Chemistry

Abstract

Green Chemistry is an active field of chemical research in which the ultimate targets are to promote high‐value, safe, clean products from renewable resources using inherently safer and clean processes. Until recently, the main efforts focused on the production of chemicals based on renewable resources by cleaner processes, notably by eliminating or substituting solvents. Quantitative Structure‐Property Relationships (QSPR) offer the opportunity to also take into account safety issues (in particular fire and explosion risks) in the early steps of development of chemicals and processes in the context of Green Chemistry. Based on robust methods for their development and validation, these predictive approaches allow access to the properties of substances based only on the knowledge of their molecular structures, even before their synthesis. QSPR models have been developed for various kinds of properties, including physico‐chemical hazards, for diverse families of compounds. They can be used as virtual screening tools to identify the best candidate among a series of possible compounds (within databases of products) for a target application and even in computer‐aided molecular design to propose alternative molecular structures, for instance for substitution purposes. So they represent very relevant tools to take into account the physico‐chemical hazards of substances together with targeted functional properties from the early stages of R&D projects toward safe‐by‐design “green” products and processes. © 2016 Society of Chemical Industry and John Wiley & Sons, Ltd

Chemoinformatics for the Safety of Energetic and Reactive Materials at Ineris

“…To solve this problem, a first approach tested by Ineris consisted in using QSPR models to provide the flash point of the pure compounds. For instance, the combination of QSPR models like the one of Carroll [90] with the mixing rule of Liaw demonstrated reliable estimations of flash points for binary mixtures, [91] including miscible as partially miscible ones, and revealed also relevant for ternary mixtures [92] …”

“…For instance, the combination of QSPR models like the one of Carroll [90] with the mixing rule of Liaw demonstrated reliable estimations of flash points for binary mixtures, [91] including miscible as partially miscible ones, and revealed also relevant for ternary mixtures. [92] In a second approach, the QSPR methodology for pure compounds was adapted to develop new QSPR models for the flash point of mixtures by introducing mixture descriptors in the models. [93] These mixture descriptors D mix were defined using mixture formula combining the molecular descriptors d i of the different constituents of the mixture and their respective molar fraction x i in the mixture as presented in Eq.…”

Quantitative structure‐property relationship modeling of phosphoric polyester char formation

“…These materials are of interest for a variety of fire‐resistant properties . In silico studies afford the opportunity to consider safety matters (eg, fire and explosion risks) in the initial steps of progress of substances and operations . Quantitative structure‐property relationship (QSPR) approaches, by their robust and predictive results, can give valuable information on chemical properties using their molecular structures, even prior to their synthesis.…”

Investigating the impact of sugar-based surfactants structure on surface tension at critical micelle concentration with structure-property relationships

“…To evidence the best sugar-based surfactant for target applications, large experimental screening should be performed. However, since such intensive experimental syntheses and characterizations are time and cost expensive, estimations of the relevant properties of possible candidates using predictive methods could be of particular interest to select the most promising ones for detailed experiment campaigns [6].…”

“…AIC the Average Information Content of order 1, 2 ACIC the Average Complementary Information Content of order 2, η the hardness, and Te the topographic electronic index calculated from all atomic pairs using Zefirov's partial charge model[38].The second one (in Eq. 6) is a simpler fragment-based model including only constitutional descriptors with a slightly higher error in prediction with RMSEIN = 0.36 (log).log CMC = -20.00 nrel,S,h -2.65•10 -2 Mw,c -63.78 nrel,single,c + 64.75(6) …”