Production

Since its active layer largely determines device efficiency, this component’s morphology received much attention.

If one material is more soluble in the solvent than the other, it will deposit first on top of the substrate, causing a concentration gradient through the film. This has been demonstrated for poly-3-hexyl thiophene (P3HT), phenyl-C61-butyric acid methyl ester (PCBM) devices where the PCBM tends to accumulate towards the device’s bottom upon spin coating from ODCB solutions. This effect is seen because the more soluble component tends to migrate towards the “solvent rich” phase during the coating procedure, accumulating the more soluble component towards the film’s bottom, where the solvent remains longer. The thickness of the generated film affects the phases segregation because the dynamics of crystallization and precipitation are different for more concentrated solutions or faster evaporation rates (needed to build thicker devices). Crystalline P3HT enrichment closer to the hole-collecting electrode can only be achieved for relatively thin (100 nm) P3HT/PCBM layers.

The gradients in the initial morphology are then mainly generated by the solvent evaporation rate and the differences in solubility between the donor and acceptor inside the blend. This dependence on solubility has been clearly demonstrated using fullerene derivatives and P3HT. When using solvents which evaporate at a slower rate (as chlorobenzene (CB) or dichlorobenzene (DCB)) you can get larger degrees of vertical separation or aggregation while solvents that evaporate quicker produce a much less effective vertical separation. Larger solubility gradients should lead to more effective vertical separation while smaller gradients should lead to more homogeneous films. These two effects were verified on P3HT:PCBM solar cells.

The solvent evaporation speed as well as posterior solvent vapor or thermal annealing procedures were also studied. Blends such as P3HT:PCBM seem to benefit from thermal annealing procedures, while others, such as PTB7:PCBM, seem to show no benefit. In P3HT the benefit seems to come from an increase of crystallinity of the P3HT phase which is generated through an expulsion of PCBM molecules from within these domains. This has been demonstrated through studies of PCBM miscibility in P3HT as well as domain composition changes as a function of annealing times.

The above hypothesis based on miscibility does not fully explain the efficiency of the devices as solely pure amorphous phases of either donor or acceptor materials never exist within bulk heterojunction devices. A 2010 paper suggested that current models that assume pure phases and discrete interfaces might fail given the absence of pure amorphous regions. Since current models assume phase separation at interfaces without any consideration for phase purity, the models might need to be changed.

The thermal annealing procedure varies depending on precisely when it is applied. Since vertical species migration is partly determined by the surface tension between the active layer and either air or another layer, annealing before or after the deposition of additional layers (most often the metal cathode) affects the result. In the case of P3HT:PCBM solar cells vertical migration is improved when cells are annealed after the deposition of the metal cathode.

Donor or acceptor accumulation next to the adjacent layers might be beneficial as these accumulations can lead to hole or electron blocking effects which might benefit device performance. In 2009 the difference in vertical distribution on P3HT:PCBM solar cells was shown to cause problems with electron mobility which ends up with the yielding of very poor device efficiencies. Simple changes to device architecture – spin coating a thin layer of PCBM on top of the P3HT – greatly enhance cell reproducibility, by providing reproducible vertical separation between device components. Since higher contact between the PCBM and the cathode is required for better efficiencies, this largely increases device reproducibility.

According to neutron scattering analysis, P3HT:PCBM blends have been described as “rivers” (P3HT regions) interrupted by “streams” (PCBM regions).

Solvent effects

Conditions for spin coating and evaporation affect device efficiency. Solvent and additives influence donor-acceptor morphology. Additives slow down evaporation, leading to more crystalline polymers and thus improved hole conductivities and efficiencies. Typical additives include 1,8-octanedithiol, ortho-dichlorobenzene, 1,8-diiodooctane (DIO), and nitrobenzene. The DIO effect was attributed to the selective solubilization of PCBM components, modifies fundamentally the average hopping distance of electrons, and thus improves electron mobility. Additives can also lead to big increases in efficiency for polymers. For HXS-1/PCBM solar cells, the effect was correlated with charge generation, transport and shelf-stability. Other polymers such as PTTBO also benefit significantly from DIO, achieving PCE values of more than 5% from around 3.7% without the additive.

Polymer Solar Cells fabricated from chloronaphthalene (CN) as a co-solvent enjoy a higher efficiency than those fabricated from the more conventional pure chlorobenzene solution. This is because the donor-acceptor morphology changes, which reduces the phase separation between donor polymer and fullerene. As a result, this translates into high hole mobilities. Without co-solvents, large domains of fullerene form, decreasing photovoltaic performance of the cell due to polymer aggregation in solution. This morphology originates from the liquid-liquid phase separation during drying; solve evaporation causes the mixture to enter into the spinodal region, in which there are significant thermal fluctuations. Large domains prevent electrons from being collected efficiently (decreasing PCE).

Small differences in polymer structure can also lead to significant changes in crystal packing that inevitably affect device morphology. PCPDTBT differs from PSBTBT caused by the difference in bridging atom between the two polymers (C vs. Si), which implies that better morphologies are achievable with PCPDTBT:PCBM solar cells containing additives as opposed to the Si system which achieves good morphologies without help from additional substances.

Self-assembled cells

Supramolecular chemistry was investigated, using donor and acceptor molecules that assemble upon spin casting and heating. Most supramolecular assemblies employ small molecules. Donor and acceptor domains in a tubular structure appear ideal for organic solar cells.

Diblock polymers containing fullerene yield stable organic solar cells upon thermal annealing.Solar cells with pre-designed morphologies resulted when appropriate supramolecular interactions are introduced.

Progress on BCPs containing polythiophene derivatives yield solar cells that assemble into well defined networks. This system exhibits a PCE of 2.04%. Hydrogen bonding guides the morphology.

Device efficiency based on co-polymer approaches have yet to cross the 2% barrier, whereas bulk-heterojunction devices exhibit efficiencies >7% in single junction configurations.

Fullerene-grafted rod-coil block copolymers have been used to study domain organization.

Supramolecular approaches to organic solar cells provide understanding about the macromolecular forces that drive domain separation.