Pobreza Monetaria

Given the proven validity of the force-field parameterisation scheme we have devel- oped, an interesting first continuation of this work would be in further exploring the role of solvent, side-chains, and conjugated structures to obtain further insight into the conformational properties of conjugated polymers in solution. Based on the existing parameters for fluorene and thiophene, it would be interesting to explore the solvent and side-chain interplay for molecules longer than 2mers as well as the role of poor solvents in molecules of lengths greater than their respective persis- tence lengths. Particularly, experimenting with solvent additives, such as DIO or phenyl-napthalene, as well as applying our parameterisation scheme to co-polymers such as PTB7 [57–59] or PffBT4T-2OD [80, 81] could offer insight into the solution

aggregation behaviour which has made both of these such promising OPV materials. Another route, which is the primary goal of our future development of this work, is in simulating multi-molecular systems in solution and in solid state formations. This is would begin with investigating solution aggregation behaviour. It would be interesting to observe how aggregation occurs with respect to solvent and ad- ditives as well as how backbone length and side-chain length enhances or hinders these processes. This could then be expanded to incorporate fullerene-type (PCBM) molecules to study selective solvation processes. Ultimately, one could imagine mod- elling evaporation processes by, at least in the first instance, a method of gradually removing solvent molecules from the system, and observing the packing formations which occur.

Multi-molecular simulation would also be of particular importance for further validation and development of our force-field parameterisation scheme by leading to calculations of properties, such as densities, which depend both on the possible dihedral disorder as well as the strength of the conjugated backbone π-stacking interactions. While we have shown the validity of a force-field with respect to conformational properties, further validation and improvement of parameterisation approaches is crucial to unlocking the full scope of these methods for studies of conjugated molecular systems.

Further exploration of the behaviours of absorption and other transition be- haviour based on the methods employed here could also shed more light on the role of conformational disorder. As above, utilising poorly-solvated systems, alternative molecules, or aggregates as input for quantum chemical calculations could be inter- esting both in further generalising the conclusions we have drawn for well-solvated fluorenes and thiophenes as well as uncovering the effects of other sources of disor- der such as intermolecular interactions and packing. Also, calculations of additional ground-to-excited state properties utilising the geometries obtained here is possi- ble. One such example is calculations of ionisation energies for which preliminary calculations have been performed.

To go beyond ground-to-excited state transitions, it would also be worthwhile ex- ploring the use of ab-initio MD methods which are currently applicable at the short (L . lp) molecule scale. Further development of these methods (as well as available

computational power) will surely be of great importance in understanding excited- state relaxation processes such as exciton formation and transfer. Additionally, it would also be interesting, from the perspective of classical force-field validation, to develop something akin to a subtraction procedure based on comparing calcu- lations of e.g. potentials of mean force for various fixed dihedrals in the ab-initio MD and classical MD regimes as opposed to static quantum chemical and force-field scans. This could allow a more definite correspondence of dynamical methods and, in principle, a more realistic classical force-field description.

It is clear that this work can be the beginning of much interesting research at the heart of organic semiconductor theory as well as into the general applicability of MD simulations for these materials. The results we have obtained so far and the methods we have developed can surely serve as foundational blocks for further investigations of these principles and, thus, the development of a better understanding of the physics of conjugated material conformation and morphology.

Length variations of partial

charges

In this appendix, we provide full details of partial charge convergence following the overview presented in Section 3.3. We begin by demonstrating the progression of total monomer charges for increasing backbone lengths. This is followed by a discussion of charge convergence at the individual atom level.

A.1 Total atomic charge convergence

In Figure A.1, the total monomeric charges, Q, of each unit of fluorene, dioctyl- fluorene, thiophene, and hexyl-thiophene molecules is shown for backbone lengths ranging from 2mer to 9mer. The three symmetric molecules (those other than hexyl-thiophene) have monomers which have zero total charge for a 2mer and only slight (_{' 0.02e) fluctuations about zero charge in individual Q values with increasing} length. Hexyl-thiophene, on the other hand, has total charges which are consistently greater than zero charge near the molecular end-points while zero charge units only appear at the 5mer scale. While the absolute value of the non-zero charges on the end units is small (_{∼ 0.05e), these are representative of significant differences in the} individual atomic charge distributions (as we discuss in the following section). As such, it follows that a molecule of at least a 5mer backbone length is necessary to obtain a set of converged hexyl-thiophene charges.