Abstract
To comply with the Kigali amendment to the Montreal Protocol in 2016, it becomes an urgent matter to develop new refrigerants with low global warming potential with simultaneously meeting conventional requirements of cooling performance, safety, and non-destructive to the ozone layer. Because each requirement links to different chemical property, proper control of various chemical properties is necessary to achieve the requirements. However, it seems to be extremely difficult to satisfy all the requirements simultaneously due to the tradeoffs among the properties. For this reason, we need to correctly recognize how these chemical properties behave when the composition of molecule is changed. Here we have done in silico screening that combines quantum chemical calculation, machine learning, and database search, where 10,163 molecules were exhaustively investigated within the properly imposed constraints, then we have found several candidates for new refrigerants. It should be noted that the synthesis of refrigerants is more difficult than that of ordinary organic molecules because glassware cannot be used for the synthesis of fluorine-containing molecules that most refrigerants apply. This makes in silico screening a more useful approach in the design of refrigerants.