Page 65 - Mirjam-Theelen-Degradation-of-CIGS-solar-cells
P. 65
Stability of Cu(In,Ga)Se 2 Solar Cells
ingression and therefore the efficiency loss of the solar cells due to exposure to damp
heat. This layer was not described as a water barrier, but its presence can make the cells
more intrinsically stable, so the costs of the barrier layers can be reduced.
SnO has also been used as TCO for CIGS solar cells. Selin-Tosun et al. [61] described
2
the difference in degradation behaviour between cells with a top layer of SnO 2 /ITO and
with i-ZnO/ZnO:Al. The unencapsulated solar cells had comparable initial efficiencies,
but it was observed that SnO /ITO solar cells showed an efficiency decrease of only 5%
2
after 120 hours of exposure to damp heat, while the efficiency of the standard i-ZnO/
ZnO:Al samples decreased by more than 70%. The parameters mainly responsible for
this difference are the V and the R . Since the V and R decreases were associated
oc
sh
sh
oc
with water permeation into the pn-junction, it was concluded that the SnO 2 window lay -
er hinders this migration. The least degradation was observed for a cell with a SnO 2 layer
with a semi-crystalline nature: it consisted of nano-crystals embedded in an amorphous
matrix. Therefore, no grain boundaries were present, so the migration of atmospheric
species through the SnO to the p-n junction was hindered.
2
As described above, ITO and other alternative TCOs are more expensive than the stan -
dard ZnO:Al, but they are also more stable. Therefore, the financial choice between low
costs ZnO:Al compared to the expensive ITO is not as straightforward as expected, since
the barrier materials required for the ZnO:Al based CIGS modules are more expensive.
Since the barrier materials add greatly to the cost of the modules, ITO based CIGS mod -
ules can end up to be more cost-effective than ZnO:Al based modules.
2.3.4.4 Summary on TCO degradation
Several types of TCOs are candidate as front electrode for CIGS solar cells, of which sput -
tered ZnO:Al is most often used, while sputtered ITO can be implemented as well. In-
creased resistivity of ZnO:Al is often found to be the main cause for the loss of efficiency
of solar cells. This increased resistivity is primarily driven by mobility decrease and is
typically caused by the diffusion of ‘foreign’ species from the atmosphere into the grain
boundaries. The migration of among others water and CO can lead to the formation
2
of molecules like Zn(OH) and Zn (CO ) (OH) , which can form a potential barrier at the
2
5
6
3 2
grain boundaries. Adsorption of atmospheric species in the grain boundaries is also
possible.
More stable ZnO:Al layers can be obtained by thicker layers, higher deposition tempera -
tures or doping concentrations or post-deposition treatments at elevated temperature.
Furthermore, the increase in resistivity can largely be reversed by annealing at vacuum
or a reducing atmosphere at elevated temperatures. Furthermore, it was found that
ZnO:Al on rough substrates show a faster increase in resistivity than on smooth sub-
strates in the presence of humidity and elevated temperatures. Therefore, rough under -
63
ingression and therefore the efficiency loss of the solar cells due to exposure to damp
heat. This layer was not described as a water barrier, but its presence can make the cells
more intrinsically stable, so the costs of the barrier layers can be reduced.
SnO has also been used as TCO for CIGS solar cells. Selin-Tosun et al. [61] described
2
the difference in degradation behaviour between cells with a top layer of SnO 2 /ITO and
with i-ZnO/ZnO:Al. The unencapsulated solar cells had comparable initial efficiencies,
but it was observed that SnO /ITO solar cells showed an efficiency decrease of only 5%
2
after 120 hours of exposure to damp heat, while the efficiency of the standard i-ZnO/
ZnO:Al samples decreased by more than 70%. The parameters mainly responsible for
this difference are the V and the R . Since the V and R decreases were associated
oc
sh
sh
oc
with water permeation into the pn-junction, it was concluded that the SnO 2 window lay -
er hinders this migration. The least degradation was observed for a cell with a SnO 2 layer
with a semi-crystalline nature: it consisted of nano-crystals embedded in an amorphous
matrix. Therefore, no grain boundaries were present, so the migration of atmospheric
species through the SnO to the p-n junction was hindered.
2
As described above, ITO and other alternative TCOs are more expensive than the stan -
dard ZnO:Al, but they are also more stable. Therefore, the financial choice between low
costs ZnO:Al compared to the expensive ITO is not as straightforward as expected, since
the barrier materials required for the ZnO:Al based CIGS modules are more expensive.
Since the barrier materials add greatly to the cost of the modules, ITO based CIGS mod -
ules can end up to be more cost-effective than ZnO:Al based modules.
2.3.4.4 Summary on TCO degradation
Several types of TCOs are candidate as front electrode for CIGS solar cells, of which sput -
tered ZnO:Al is most often used, while sputtered ITO can be implemented as well. In-
creased resistivity of ZnO:Al is often found to be the main cause for the loss of efficiency
of solar cells. This increased resistivity is primarily driven by mobility decrease and is
typically caused by the diffusion of ‘foreign’ species from the atmosphere into the grain
boundaries. The migration of among others water and CO can lead to the formation
2
of molecules like Zn(OH) and Zn (CO ) (OH) , which can form a potential barrier at the
2
5
6
3 2
grain boundaries. Adsorption of atmospheric species in the grain boundaries is also
possible.
More stable ZnO:Al layers can be obtained by thicker layers, higher deposition tempera -
tures or doping concentrations or post-deposition treatments at elevated temperature.
Furthermore, the increase in resistivity can largely be reversed by annealing at vacuum
or a reducing atmosphere at elevated temperatures. Furthermore, it was found that
ZnO:Al on rough substrates show a faster increase in resistivity than on smooth sub-
strates in the presence of humidity and elevated temperatures. Therefore, rough under -
63
