Research

Previous work on connecting quantum-mechanical and coarse-grained scales relied on computationally-costly and complex workflows that involve backmapping and repeated quantum chemical calculations [Adv. Funct. Mater. 2020]. Due to their poor scaling, such methodologies cannot be used for accelerating materials discovery and optimization. To address this issue, we showed how coarse-grained models can be parametrized in order to retain electronic properties, and supervised machine learning used to retrieve electronic structure information at coarse-grained resolutions. We showed this for radical-containing redox-active polymers, and compared the obtained results to the traditional backmapping-based approach [arXiv 2022]. The comparison shows that the method can achieve accuracy on par with backmapping-based approaches, while speeding up the workflows by orders of magnitude, thereby opening the way for exploration of structure-electronic structure relationships across the thermodynamic phase space of soft electronic materials.

We modeled the morphologies of organic solar cells with coarse-grained simulations based on the Martini model [J. Am. Chem. Soc. 2017]. The approach explicitly simulates the solvent evaporation process that takes place during the formation of such morphologies. Moreover, the approach allows to retain chemical detail while permitting to reach relevant length scales. The method also proved useful in elucidating host-dopant miscibility in organic semiconductor blends for thermoelectric applications [J. Chem. Mater. 2017], as we teamed up with the groups of Prof. Koster and Prof. Hummelen. In this context, a minimalistic Martini-based approach also proved very insightful in predicting the miscibility of a small-molecule dopant in environments of different polarity [Adv. Mater. 2018].

The morphologies obtained with the coarse-grained simulations described above can be promptly backmapped to atomistic resolution, allowing for detailed studies of electronic properties by quantum chemical calculations [J. Am. Chem. Soc. 2017, Adv. Funct. Mater. 2020]. In particular, we studied the impact that functionalizing the electron acceptors with polar side chains has on the energy levels of such materials [Adv. Funct. Mater. 2020]. By leveraging a similar combination of quantum-chemical and microelectrostatic calculations, with Anna we investigated the effect of the environment on the energy levels of dye molecules in a complex molecular aggregate [Chem. Sci. 2020, J. Am. Chem. Soc. 2020].

We individuated and characterized the limits of the Martini 2 coarse-grained model [J. Chem. Theory Comput. 2019]. This work opened the way for the development of a new version of the force field, dubbed Martini 3 [Nat. Methods 2021, Adv. Theory. Simul. 2022]. Among other things, we paid particular attention, unlike in the previous version of the model, to aspects of the model important for describing soft materials. Recently, we put together a perspective of the state of the Martini model in materials science - an emerging, exciting field of application of this model [Adv. Mater. 2021].