Page 252 - Mirjam-Theelen-Degradation-of-CIGS-solar-cells
P. 252
Chapter 8
profiles were detected for the reference air sample (not depicted).
The sample degraded in non-purged water, which contains small quantities of among
others nitrogen, oxygen and CO, is shown in Figure 8.c. In this figure, it seems that
2
the thickness of the ZnO:Al layer has decreased slightly, while the CIGS and the zinc
oxide layer have intermixed slightly. The sodium profile is still the same, but it is
possible that the content in the CIGS layer and especially near the pn-junction has
decreased. The potassium content might have increased in this region. As is shown
in Figure 8.5, the thickness of none of the layers has reduced greatly, while the layers
also do not seem intermixed. It is therefore more likely that the reduced sputter time
can be explained by the dissolution of part of the zinc oxide, resulting in a very porous
material which has an increased average sputter rate compared to the dense original
zinc oxide. Since the sputter rate in zinc oxide is locally very high (at the empty grain
boundaries) and locally lower (at the still existing grains) and many grains and empty
areas are included in the measurement, the intermixing is probably largely an artefact
of the SIMS measurement. The intermixing between the CIGS and the molybdenum
also becomes larger, which can probably be explained by the same reason.
Additionally, the amount of hydroxide has increased, especially near the ZnO:Al/air
interface, while the sulfur and the aluminium content were also higher very close to
this surface. A small change was also observed at the CIGS/Mo interface, where the
oxygen concentration likely has slightly increased.
The sample degraded in HO with O shows an even larger intermixing between
2
2
the ZnO:Al and CIGS signals, which might indicate less homogeneity or just more
dissolution of the zinc oxide. Further trends are similar to the non-purged H O sample.
2
The picture changes greatly when the samples degraded in HO/air and HO/CO /
2
2
2
N are studied. As expected from the SEM pictures, for the HO/air sample, the
2
2
sputtering time for the ZnO:Al has been reduced even further, while the amount of
foreign species near the ZnO:Al/air interface becomes higher. These species include
hydroxide, sulfur and to a lesser account chlorine. There seems to be a clear border
in between the thinner ZnO:Al and the CIGS layer, while the Mo/CIGS interface is also
quite well defined, which might indicate a very homogeneous dissolution. In this
sample, the highest increase in oxygen content at the Mo/CIGS interface was found.
The sodium profile in this sample is similar to the profiles in the reference samples.
The H O/CO /N gave largely the same picture as the HO/air samples, but the border
2
2
2
2
between the CIGS and the ZnO:Al signals is less well defined. Furthermore, the
increase in oxygen at the Mo/CIGS interface can barely be observed.
250
profiles were detected for the reference air sample (not depicted).
The sample degraded in non-purged water, which contains small quantities of among
others nitrogen, oxygen and CO, is shown in Figure 8.c. In this figure, it seems that
2
the thickness of the ZnO:Al layer has decreased slightly, while the CIGS and the zinc
oxide layer have intermixed slightly. The sodium profile is still the same, but it is
possible that the content in the CIGS layer and especially near the pn-junction has
decreased. The potassium content might have increased in this region. As is shown
in Figure 8.5, the thickness of none of the layers has reduced greatly, while the layers
also do not seem intermixed. It is therefore more likely that the reduced sputter time
can be explained by the dissolution of part of the zinc oxide, resulting in a very porous
material which has an increased average sputter rate compared to the dense original
zinc oxide. Since the sputter rate in zinc oxide is locally very high (at the empty grain
boundaries) and locally lower (at the still existing grains) and many grains and empty
areas are included in the measurement, the intermixing is probably largely an artefact
of the SIMS measurement. The intermixing between the CIGS and the molybdenum
also becomes larger, which can probably be explained by the same reason.
Additionally, the amount of hydroxide has increased, especially near the ZnO:Al/air
interface, while the sulfur and the aluminium content were also higher very close to
this surface. A small change was also observed at the CIGS/Mo interface, where the
oxygen concentration likely has slightly increased.
The sample degraded in HO with O shows an even larger intermixing between
2
2
the ZnO:Al and CIGS signals, which might indicate less homogeneity or just more
dissolution of the zinc oxide. Further trends are similar to the non-purged H O sample.
2
The picture changes greatly when the samples degraded in HO/air and HO/CO /
2
2
2
N are studied. As expected from the SEM pictures, for the HO/air sample, the
2
2
sputtering time for the ZnO:Al has been reduced even further, while the amount of
foreign species near the ZnO:Al/air interface becomes higher. These species include
hydroxide, sulfur and to a lesser account chlorine. There seems to be a clear border
in between the thinner ZnO:Al and the CIGS layer, while the Mo/CIGS interface is also
quite well defined, which might indicate a very homogeneous dissolution. In this
sample, the highest increase in oxygen content at the Mo/CIGS interface was found.
The sodium profile in this sample is similar to the profiles in the reference samples.
The H O/CO /N gave largely the same picture as the HO/air samples, but the border
2
2
2
2
between the CIGS and the ZnO:Al signals is less well defined. Furthermore, the
increase in oxygen at the Mo/CIGS interface can barely be observed.
250
