The Substituent Effect in Chemistry and Machine Learning: Applications to Organic Photovoltaics and Energetic Materials Itamar Borges Jr. (IME - Dept. de Química) Resumo: The substituent effect – the modification of the electronic properties of molecules by changing groups of atoms - is ubiquitous in Chemistry and Materials Science. Machine learning (ML) techniques, a set of statistical methods that take advantage of recent developments in hardware and larger volumes of more accurate data, have become a handy tool for basic research. Combining these two approaches can reveal unknown phenomena and new physical insights. In this talk, we present Hammett’s theory, which quantifies via numerical constants, the electron-donor or -withdrawal power of substituents. We discuss a new comprehensive and systematic set of Hammett’s constants for general use determined with ML . Afterward, we examine diketopyrrolopyrrole (DPP) molecules with one or two substituents, which can be used as a fine-tuned donor material in organic electronic devices. Several photovoltaic properties, including power conversion efficiencies in bulk heterojunctions, exciton binding energies and intramolecular charge transfer effects, are discussed in the framework of Hammett’s theory . Finally, we examine the molecular origin of the impact sensitivity of an energetic material, which was rationalized with ML. In particular, we found that a positive or negative contribution of the molecular properties to the sensitivity depends on the value of this property . Finally, we discuss some implications of combining quantum chemistry with ML and mention other research carried out by our group. References: - MONTEIRO-DE-CASTRO, GABRIEL ; […]