Device architectures for molecular semiconductor PV devices that deliver promising performance contain design features that deal with the strong electron-electron interactions present in these low dielectric constant materials.  I will discuss two examples where this plays a critical role.   First, the process of charge separation at the donor-acceptor heterojunction in bulk heterojunction devices requires that electron and hole move rapidly beyond their coulomb (re)capture range, of around 5 nm.  We have shown this is achieved in polymer fullerene blends through access to delocalised conduction band states formed in fullerene clusters [1], and we now find that this process seems more general, present also in structures with non-fullerene acceptors.  Second, large spin exchange energies allow scope for multiple exciton generation for materials for which the triplet exciton energy is less than one half of the singlet exciton energy, since this favours energetically the fission of a photogenerated singlet to a pair of triplet excitons. If this process can be used in tandem with a lower energy gap semiconductor that harvests singlet excitons directly then this may enhance solar energy conversion beyond the single-junction Shockley-Queisser limit.  I will discuss recent work in Cambridge that explores a number of the steps that would be required to achieve this.  Though the initial process of singlet fission can be extremely rapid (80 fsec for pentacene), we now find clear evidence from transient optical spectroscopy that this can produce a bound triplet-triplet pair that takes considerably longer, beyond 10 nsec, before this dissociates to free triplet excitons.   We have observed this for TIPS-tetracene in solution [2] and I will discuss more recent observations in solid films.  It is necessary to use the energy in these triplet excitons to generated separated electron hole pairs.  I will report some recent observations of the process of ionisation of triplet excitons at all-organic donor acceptor heterojunctions formed with pentacene and fullerene.  An alternative approach is to transfer the triplet exciton itself to a luminescent material and we have explored transfer of triplet excitons from pentacene to lead selenide nanocrystals [3]. 

1          Gélinas, S. et al. Science 343, 512-516, doi:10.1126/science.1246249 (2014).

2          Stern, H. L. et al. PNAS 112, 7656-7661, doi:10.1073/pnas.1503471112 (2015).

3          Tabachnyk, M. et al. Nature Materials 13, 1033-1038, doi:10.1038/NMAT4093 (2014).