The impact of the deposition temperature on the conduction band in ZnSnO for CIGS-solar cells
|Bandgap||Energy difference between the valence band and the conduction band which the electron is excited between.|
|Crystallites||Very small crystals.|
In CIGS-solar cells there is a buffer layer for improving the electronic and optical properties in the interface between the CIGS layer and the window layer. In this layer, a good balance between the energy levels in the material is needed, both to the absorber and the front contact. The most common buffer layer is CdS, although this has a couple of disadvantages such as the low bandgap that enables absorption of some of the sunlight and the use of a chemical bath deposition method to deposit it, which is not in vacuum and therefore not optimal for industrial manufacturing of CIGS-solar cells. Atomic layer deposition (ALD) is a vacuum-based technique which has been successful for depositing other buffer layers. This is a chemical technique where the layer is deposited from gas to solid with alternating, self-limiting reactions, which gives a controllable deposition of uniform films at relatively low temperatures. ZnSnO (ZTO) is an example of buffer layers deposited by ALD, for which the conduction band has been shown to decrease in energy with increasing temperature. This is probably due to nanometer-sized crystallites formed in the films that have quantum mechanical effects on the bandgap. The higher temperature will increase the size of the crystallites, which leads to a lower conduction band. It is desirable to have a conduction band that is somewhat higher than the conduction band of the absorber layer. In this study, the performance of solar cells with ZTO deposited by ALD at different temperatures was investigated.
Solar cells were manufactured as a stack of glass/Mo/CIGS/buffer layer/window layer, where ZTO was used as a buffer layer. The ZTO was deposited with ALD at temperatures between 90 and 180°C. The number of pulse cycles for the ALD was altered for the different temperatures to make ZTO films with the same thickness. The composition was seen to be only marginally influenced by the temperature. The solar cells were characterized to evaluate the structure and performance. A numerical simulation was also done in the program SCAPS to explain the measured values.
The current was not significantly affected by the temperatures, but the voltage was highest at 105°C and then decreased with increasing temperature. Crystallites were found in the ZTO-layer with microscope, for 120°C these were relatively few and 3-4 nm in diameter while for 180°C they were quite many and 5-7 nm in size. Spectroscopy showed that the temperature did not give rise to diffusion between the materials, which therefore cannot explain the difference in voltage. The simulation showed that the highest efficiency was acquired when the conduction band for ZTO was 0.1-0.3 eV higher than the conduction band for the absorber layer, and a higher and lower value would lead to a decreased efficiency for example due to a lower voltage.
The measured change in voltage cannot be explained with another composition or diffusion between the materials, since these have been excluded. It is also unlikely that for example the concentration of Na would be significantly affected by the temperature in the ALD since the reference samples with CdS did not display the change in voltage due to temperature. A probable explanation is therefore the change in energy level of the conduction band, which changes with temperature and in its turn then leads to the change in voltage. The lower voltage at high and low temperatures are probably explained with recombination at the interface between absorber layer and ZTO, for high temperatures and a low conduction band due to defects and for low temperatures and a high conduction band due to blocking of electrons to be transmitted into the front contact. The study showed that appropriate temperatures to deposit ZTO with ALD was 105-135°C and that changing the bandgap for ZTO could be used to match the conduction band in CIGS and fit as a buffer layer. This could also be used for matching with CIGS layers with other bandgaps as in this study.
Read the original article here