Modeling light-matter interaction in complex embedded systems

Systems in the condensed phase are characterized by a high degree of complexity due to their large size and variety of intermolecular interactions at the microscopic level. Computational spectroscopy is a valuable tool for the investigation of complex systems, but it requires ad hoc modeling able to reduce the computational cost to simulate effectively realistic scenarios. To this end, I have developed a set of computational tools for the calculation of spectroscopic properties of chemical systems embedded in the external environment (solvents or nanostructured plasmonic materials). The first part of the thesis focuses on solvated systems. Starting from the classical, atomistic, and fully polarizable model based on the Fluctuating Charges (Fluctuating Dipoles) force field, I propose two complementary multiscale approaches for the simulation of spectroscopic properties: the first is based on a hybrid quantum-classical description of the environment for the first-principles description of short-range non-electrostatic interactions between solute and solvent. The second aims at cost-effective simulations of absorption spectra of large embedded substrates through the coupling of a polarizable embedding with the Time-Dependent Density Functional Tight-Binding approach. In the second part of the thesis, I focus on the description of surface-enhanced spectroscopies, in which the environment consists of plasmonic nanostructured materials that may impressively amplify the electromagnetic field experienced by the adsorbed target thanks to the surface plasmon resonance phenomenon. In this framework, I extend the classical, atomistic frequency-dependent Fluctuating Charges model to the simulation of the optical properties of noble metal nanoparticles and graphene-based materials composed of millions of atoms. The agreement with ab initio and experimental data is remarkable, also in the case of non-homogeneous plasmonic silver-gold alloyed nanoparticles. The model is then coupled to a DFT Hamiltonian for the simulation of Surface-Enhanced Raman Scattering spectra.