This thesis is concerned with quantification of non-geminate recombination
losses in organic bulk heterojunction solar cells. After description of
the context of this work, the theoretical background and the methodology
employed are presented. In this thesis, many different polymer:fullerene
systems are investigated.
In the next chapter, we show that the study of non-geminate losses using
charge extraction/transient photovoltage analysis can be applied to many
different systems away from P3HT/P3HS blends. We see to what extent
ideality factors can give a more precise description of the exact recombination
mechanism.
Then, the change of optoelectronic properties of a high performance
polymer:fullerene blend upon a blend ratio perturbation is investigated.
The resulting shifts in energetics and dynamics of the blends are quantified.
A quantitative agreement between two methods (charge extraction
and electroluminescence) probing the shifts in the energetics at the heterojunction
is presented.
In the next two chapters, two limits of the common vision of polymer:
fullerene systems are explored by combining experiment and 1D drift-diffusion
modelling. First, the impact of the variations of the spatial distribution
of carriers on the apparent reaction order is experimentally investigated.
The study reconciles the apparent contradictions currently in
the literature regarding the meaning of high reaction orders. In the following
chapter, the often underestimated effects of unintentional doping
in polymer blends are addressed experimentally. In particular, its effect on
device optimisation, understanding of carrier collection and recombination
dynamics are explored. This study suggests that many donor/acceptor
blends are not, contrary to common belief, intrinsic semiconductors.
Finally, an analysis of the Langevin and non-Langevin behaviour
of some efficient systems is presented. The different interpretations of
regular observations disproving non-geminate recombination following
the Langevin type mechanism are reviewed. We suggest the ratio ’recombination
over collision’ is often overlooked and question the underlying
assumption that it should be unity.