“Charge transfer complexes in polymer:fullerene bulk heterojunction solar cells.”
For an efficient conversion of a flux of solar photons into an electric current by organic materials, the presence of a material interface between an electron donating and electron accepting material is crucial. Most successful active layers for organic solar cells comprise a blend of conjugated polymers as electron donors and fullerenes as electron acceptors, exhibiting power conversion efficiencies higher than 6%, nowadays. In order to find pathways to increase this efficiency further, properties of the electronic states at the donor/acceptor interface and their role in determining the overall power conversion efficiency, are investigated in this work.
To probe these interfacial properties, the fast and highly sensitive technique Fourier-Transform Photocurrent Spectroscopy (FTPS) is used to detect the weak absorption caused by ground state interaction of polymers and fullerenes, forming a charge transfer complex (CTC). Optical excitation of this donor/acceptor CTC by light with photon energies lower than the optical gap of both the donor and acceptor materials, results in the creation of a charge transfer (CT) exciton or CT state, comprising an electron in the acceptor phase, coulombically bound to a hole on the donor phase.
In our study, such a CT transition within the optical gap of both pure materials was detected in all polymer:fullerene solar cells exhibiting a significant photovoltaic effect. In these cases, the energy of the CT state is lower than the energy of the excited states of the pure blend constituents, and can efficiently be populated. The competition between geminate recombination and field dependent dissociation of CT excitons can still limit photocurrent production in some polymer:fullerene material combinations.
The origin of the open-circuit voltage
It is further shown that next to the free charge carrier and photocurrent generation, the open-circuit voltage (Voc) is also affected by donor/acceptor CTC formation. Voc is determined by the balance between free carrier generation and recombination processes in the active layer. These recombination processes can proceed through the formation of a CT exciton with subsequent emission of low energy photons, a process that is visible in electroluminescence experiments. Although the electroluminescence spectrum is dominated by this CT emission, the external quantum efficiency of this process is very low, in the 10-6 to 10-9 range.
In order to quantitatively investigate the role of CTC formation on the photovoltage production in polymer:fullerene solar cells, a reciprocity relation between Voc and the photovoltaic and electroluminescent actions of a generalized solar cell is used. This theory is established for various types of both inorganic and dye sensitized solar cells. In this work, it is shown to be valid, also in the case of polymer:fullerene solar cells, on the condition that the sub-gap absorption and emission due to the CT states is taken into account. Because both absorption and emission can only be detected by highly sensitive techniques, this has been overlooked in the first ten years of research on polymer:fullerene photovoltaics.
As predicted by the reciprocity relations, a linear correlation between Voc and the spectral position of the CT band is observed for a range of polymer:fullerene blends, comprising different donor polymers. The energy of the CT state (ECT) is known to correlate with the difference between the HOMO energy of the polymer donor and the LUMO energy of the fullerene acceptor. This explains the widely observed correlation between Voc, measured under solar conditions, and this energetic difference.
Influence of polymer:fullerene stoichiometry and crystallization on ECT and Voc
We also investigate the influence of the preparation conditions on ECT and thus Voc. Increasing the concentration of the fullerene derivative PCBM from 5 % to 80 % in MDMO-PPV:PCBM photovoltaic devices, results in a redshift of the CT band, of ~0.15 eV. The reason for these redshift could be due to an increasing degree of PCBM crystallinity upon increasing the PCBM content. However, the slight increase in overall dielectric constant of the blend upon increasing the PCBM content, could also cause the observed redshift. The relative contributions of dielectric constant changes and crystallinity changes however, are not determined yet, and will be the subject of future work.
The effect of the donor polymer crystallinity on the spectral position of the CT band has been investigated in more detail for polythiophene:PCBM solar cells. The crystalline fiber to total polymer weight ratio in the polythiophene:PCBM blends was varied between ~ 0.1 and ~ 0.9. We observed that ECT decreases about ~20 meV, when increasing the fiber to total polymer weight ratio by 0.1. It was also found that Voc always follows roughly the same trend as ECT.
Energetic losses between qVoc and ECT
For the solar cells investigated in this work an energetic difference between ECT and qVoc measured under solar illumination conditions of ~0.6 eV is found. The origin of this difference is twofold. One part is energetic loss due to the radiative recombination through the CT state and can be reduced by reducing the number of CT states present in the device or by reducing the electronic coupling between donor and acceptor, changing the CT absorption cross section, but also the CT emission rate constant. For the investigated devices, this radiative loss is ~0.25 eV. The second part of the loss, about 0.35 eV is due to non-radiative recombination mechanisms. Further investigations, identifying these non-radiative recombination paths, are necessary in order to find possible pathways to minimize this part of the energetic loss.
Additionally, Voc was investigated under conditions different from solar conditions. Also ECT was determined at different temperatures, between 150 K and 300 K. For the investigated polymer:fullerene blends, it was found that the difference between ECT and qVoc decreases linearly with decreasing temperature and logarithmically with decreasing illumination intensity. Furthermore, irrespective of the used illumination intensity, the extrapolation of qVoc to 0 K equals the extrapolation of ECT to 0 K.
Upper limits for power conversion efficiency of organic donor/acceptor solar cells
Upper limits of attainable power conversion efficiency for organic donor/acceptor photovoltaics are derived for AM1.5 conditions at room temperature. The obtained maximum efficiency values are a function of two parameters: The optical gap of the main absorber Eg and the energetic losses due to the presence of the CT state. This efficiency limit is between 25 % and 33 %, if the optical gap of the main absorber is in the 1.2-1.7 eV region and if the losses through the CT state are below 0.2 eV. However, best performing polymer:fullerene solar cells nowadays have power conversion efficiency of 6 %. In these devices, EQEPV values of 70-80 % are reached above the optical gap of the polymer. The main source of energy loss in the best device lies thus in the low values of Voc as compared to the upper limit.