Some DSSCs' problems

   General considerations on DSSCs' learning curve

   The jump in solar energy conversion efficiency of DSSCs motivated significant optimism with respect to the feasibility of cheap and long-term stable dye sensitization solar cells. It was, however, not generally shared, first, because the cell included noble metals (ruthenium/platinum) and it had to be perfectly sealed because of its toxic organic components. Conductive glass is also a non-negligible cost factor [13].

   In addition, the expected long-term stability of the dye was deduced from spectroscopic and short-term photochemical evidence. It had not been tested in long-term operating dye sensitization solar cells [13].

   In the nearly one and a half decades following the discovery of Grätzel’s group, many research teams (approximately 80) have tried to improve and understand the dye sensitization solar cell. Many different dyes and transition metal complexes have been tested as sensitizers [13,25, 26,27]. Many types of electrolytes have also been examined including gel-polymer and molten-salt-electrolytes. Many new preparation techniques have been tried for nano-crystalline titanium dioxide. Numerous spectroscopic analytical or electrochemical techniques have been applied to characterize nano-structured dye sensitization cells and chemical components were added or replaced. Further, in the search for the origin of the photopotential the SnO2/ TiO2/ electrolyte contact received special attention [13].

   Efforts have been developed industrially to produce sensitization solar cells in a reproducible, professional way to obtain very reproducible solar cell prototypes. Nevertheless, the efficiency never exceeded a 7-8% limit and the long-term stability was not satisfactory for supporting straightforward industrial production effort [13].

   Progress may depend on the number of researchers involved, on the availability of new materials, or on jumps in knowledge or creativity. It may be argued that silicon solar cells have received much more attention, especially from industry. But, interestingly, CdTe or CuInSe2/S2 (CIS) solar cells, which have received much less industrial attention show similar learning curves [13].

   Industry concentrates on industrial prototypes, which require more adapted production strategies, the efficiency of which typically is 30% lower than the laboratory efficiencies. Analysis of research progress until now, in comparison with the development of other, classical, solar cells, demonstrates that the development of the dye sensitization solar cell is stagnating to some extent. This can be seen from the learning curve, which is depicted in Fig. 1 (bottom) and compared with that experienced for classical solar cells (c-Si, a-Si, CdTe, CIS). The development of the dye sensitization solar cell first followed a learning curve somewhat slower than that found for classical solar cells, during its first two decades, when still only a few research groups participated [13].

Fig. 1. Comparison of the learning curve for classical crystalline and thin layer laboratory solar cells (c-Si, a-Si, CdTe, CIS) with the learning curve for dye sensitization cells.

   After 1991, the learning curve for efficiency practically stagnated in spite of the involvement of many research groups. The efficiency record of 10.4% obtained with a very small cell of area only 0.18 cm2 remains a singularity produced and apparently takes advantage of a minimal internal resistance due to small cell dimensions. It appears that for solar cells of cm2 dimensions efficiencies of 7-8% are realistic, with cells, which reach 9%, being exceptions obtained under especially favorable circumstances [13].

   In all efficiency measurements the temperature effect, the phenomenon that efficiency may significantly decrease with temperature, is not considered (when heating from 10 to 70 °C a silicon solar cell, for example, may lose one-third of its efficiency). Typically the measurements are performed at lower ambient temperature at conditions where the solar cell is not allowed to heat up [13].
   The best sensitization cells with purely organic dyes apparently reached efficiencies of 6%, and more recently with coumarine and polyene sensitizers even 7.7% [13].

   We emphasize that the learning curve is reasonably correct, for dye solar cell efficiencies in Fig. 1b, only for its later phase (starting with 1995) when efficiencies could be confirmed. Before, efficiencies were probably lower measured not with simulated solar cell efficiencies, but measured with laboratory light sources, therefore downward pointing arrows were introduced into Fig. 1 (bottom) to indicate effectively lower efficiencies [13].
   The earlier efficiency reports from the Lausanne group with TiO2-based cells, on the other hand, sometimes communicated 10% and even more efficient cells, but later only confirmed with very small area cells [13].

   The solar cell costs were then estimated to be only 10% of the silicon cell costs, a too optimistic expectation for many researchers, given that encapsulation of classical solar cells is already more expensive and calculated for lifetimes of 20-30 years [13].

   Sensitizer stability

   One of the big accomplishments of Grätzel and his group was the identification of physical-chemical conditions, which allowed Ru complexes to tolerate a surprisingly large number of sensitization cycles. This is confirmed now. But is the number of electron transfer cycles as high as proposed? Since industrial efforts started in dye sensitization solar cell development, it was claimed that the ruthenium complex can survive 108 electron transfer steps and would therefore be stable for 20 years. These conclusions were apparently derived from open-circuit experiments, performed with dye sensitization solar cells, illuminated with high laser light intensities. In contrast, when photocurrent imaging techniques were applied to selectively illuminated dye sensitization cells, clear photo-degradation was observed [13].

   Photo-degradation is limited to selectively illuminated areas of a dye sensitization cell and depends linearly on the light intensity [13].

   The stability of dye molecules apparently depends on the specific adsorption sites. There are adsorption sites where dye molecules have a greater chance to degrade and others where they have a smaller chance [13].

   Some molecules photoreacted rapidly, others slower and the rest very slowly so that it could be concluded that the adsorption sites were responsible for the stability behavior. Where the chemical bonding is optimal, the oxidized sensitizer may survive until it is regenerated. Where it is perturbed and non-ideal, the oxidized sensitizer may irreversibly react [13].

   Only when the ruthenium complex is well bonded to TiO2 interfaces, does it show reasonably good performance. This definitely means that the interfacial bonding of the ruthenium complex to the sensitizer substrate is of most critical importance [13].

   The conclusion must be that the photo-degradation of ruthenium complexes relate to the presence of unfavorable adsorption sites [13].

   Electrolyte instability and electrolyte contamination

   As explained before, the iodide/triiodide redox couple has proven to be highly successful in dye sensitization solar cells because of its inherent kinetic irreversibility. Electrons are much more easily donated by iodide than recaptured by triiodide. Replacement of I-/I3- by redox systems with similar redox potentials (Fe2+/3+, Fe(CN)3+/4+, quinone/hydroquinone) leads to a significant drop of solar cell efficiency.

   Unfortunately, iodine is known to be photochemically reactive. It can engage in photochemical mechanisms especially by dissociating into atomic iodine, which is a radical and participating in many reactions. It can also photochemically react with oxygen or other redox species. Water is excluded from the electrolyte when using Ru complexes in sensitization cells due to its negative effect on bonding to TiO2. However, sealing techniques for liquid dye sensitization solar cells are far from perfect so that problems arise. Infrared studies show that gradually water molecules diffuse from the atmosphere into the cell. The consequence is a reaction with triiodide leading to iodate with intermediates of the iodine electrochemistry reacting further with oxygen and water. The regeneration of the oxidized sensitizer S+ by iodide in water- and oxygen-free organic electrolyte [13].

2S+ + 3I- → I3- + 2S

will, with a certain probability, change in presence of water and oxygen:

2S+ + 3I- + 4O2 + H2O → 2S + 3IO3- + 2H+

   Electrolyte degradation is responsible for a frequently observed deterioration of dye sensitization cell function. It is not a photo-degradation but a dark degradation, which will influence photo-degradation, because the oxidized sensitizer cannot be reduced by iodide [13].