Page 257 - Mirjam-Theelen-Degradation-of-CIGS-solar-cells
P. 257
The impact of atmospheric species
High local concentrations of CO could have been caused by a fast diffusion rate of CO
2
2
through the grain boundaries. It is also possible that an additional reaction between
the atmospheric species and the CIGS solar cell components was involved, thereby
lowering the pH.
The local pH in the grain boundaries could have become below 6, while the global pH
of the complete aqueous solution was still safely above 6.5. Since zinc oxide dissolves
at a pH below 6, this would led to local dissolution starting from the grain boundaries.
Since the H O/O sample that did not contain CO also showed traces of grain boundary
2
2
2
dissolution, it is naturally proposed that the additional reactions with elements from
the solar cells must have played a role. Unfortunately, more knowledge about the exact
nature of the dissolution should be obtained by pH or elemental measurements at
specified positions within the grain boundaries, which was not possible.
However, it is shown on the SEM pictures (Figure 8.5) that dissolution often starts near
the CdS/i-ZnO region. Furthermore, a new layer seems to have formed in this region, so
it is therefore likely that more elements, like copper, indium, gallium, selenium, cadmium
or sulphur play a role in the dissolution of ZnO. It is proposed that one or more of these
elements is involved in the change of the pH. One likely candidate is the sulphur present
in CdS, since sulphur containing species like H SO are known for their acidity.
4
2
8.4.2 Shunt resistance
The samples that degraded completely (H O/CO /N and H O/air) showed a decrease in
2
2
2
2
shunt resistance at the same time as the decrease of the J , which might be explained
sc
by the formation of shunt paths. At a later point in time, the shunt resistance increased
again - however, this was not necessarily a disappearance of the shunt paths, but more
likely a complete change in the morphology, which then dominated over the shunt
effects.
A more surprising result was the decrease in the shunt resistance for the sample
degraded in H O/N . This repeatedly observed phenomenon might be explained by
2
2
the formation of conductive materials along the edge of the samples. These materials
(for example binary phases like Cu Se) will probably oxidise when oxygen is present
2-x
in combination with water, leading to a non-conductive oxide. Therefore, no shunt
resistance deterioration is observed for the H O and the H O/O samples.
2
2
2
8.4.3 Impact on CIGS modules
This study has shown that CIGS solar cells degrade rapidly in the presence of both wa -
ter and large quantities of CO . In order to protect the solar cells, the ingression of the
2
polar water molecules can be prevented by water barriers, but one can also consider
a barrier that hinders the ingression of the non-polar CO molecules, even based on
2
255
High local concentrations of CO could have been caused by a fast diffusion rate of CO
2
2
through the grain boundaries. It is also possible that an additional reaction between
the atmospheric species and the CIGS solar cell components was involved, thereby
lowering the pH.
The local pH in the grain boundaries could have become below 6, while the global pH
of the complete aqueous solution was still safely above 6.5. Since zinc oxide dissolves
at a pH below 6, this would led to local dissolution starting from the grain boundaries.
Since the H O/O sample that did not contain CO also showed traces of grain boundary
2
2
2
dissolution, it is naturally proposed that the additional reactions with elements from
the solar cells must have played a role. Unfortunately, more knowledge about the exact
nature of the dissolution should be obtained by pH or elemental measurements at
specified positions within the grain boundaries, which was not possible.
However, it is shown on the SEM pictures (Figure 8.5) that dissolution often starts near
the CdS/i-ZnO region. Furthermore, a new layer seems to have formed in this region, so
it is therefore likely that more elements, like copper, indium, gallium, selenium, cadmium
or sulphur play a role in the dissolution of ZnO. It is proposed that one or more of these
elements is involved in the change of the pH. One likely candidate is the sulphur present
in CdS, since sulphur containing species like H SO are known for their acidity.
4
2
8.4.2 Shunt resistance
The samples that degraded completely (H O/CO /N and H O/air) showed a decrease in
2
2
2
2
shunt resistance at the same time as the decrease of the J , which might be explained
sc
by the formation of shunt paths. At a later point in time, the shunt resistance increased
again - however, this was not necessarily a disappearance of the shunt paths, but more
likely a complete change in the morphology, which then dominated over the shunt
effects.
A more surprising result was the decrease in the shunt resistance for the sample
degraded in H O/N . This repeatedly observed phenomenon might be explained by
2
2
the formation of conductive materials along the edge of the samples. These materials
(for example binary phases like Cu Se) will probably oxidise when oxygen is present
2-x
in combination with water, leading to a non-conductive oxide. Therefore, no shunt
resistance deterioration is observed for the H O and the H O/O samples.
2
2
2
8.4.3 Impact on CIGS modules
This study has shown that CIGS solar cells degrade rapidly in the presence of both wa -
ter and large quantities of CO . In order to protect the solar cells, the ingression of the
2
polar water molecules can be prevented by water barriers, but one can also consider
a barrier that hinders the ingression of the non-polar CO molecules, even based on
2
255
