Page 165 - Mirjam-Theelen-Degradation-of-CIGS-solar-cells
P. 165
Degradation mechanisms of the molybdenum back contact
edges of a CIGS cell. The non-selenised samples already showed visual degradation
and a decrease of the reflectance after two hours.
Within the selenised samples, there was also an impact of the sputter pressure on the
reflectance. Mo2Se retained only spot-like degradation and was still quite reflective
after 105 hours of degradation, while on Mo10Se and Mo15Se the spots grew into a
mosaic-like structure on the surface, which led to a decrease in reflectivity. Mo2Se
was the only sample that retained a reflectance over 10%. Therefore, considering the
optical properties, dense selenised molybdenum layers are also the most suitable for
incorporation in CIGS solar cells.
5.4.1.4 Degradation route
Figure 5.29 shows a possible degradation route for molybdenum thin film layers. Due
to damp heat exposure of molybdenum (step 1), an oxide layer is formed on top of the
molybdenum sample. The same effect might occur in the grain boundaries (step 2).
It was observed in cross-section SEM that porous layers had thicker top layers than
dense samples, which indicates that these layers degrade the fastest. This can be
attributed to the faster diffusion of water and other molecules like O and CO in
2
2
molybdenum with a higher volume of intergranular material, since the mobility of for
example sodium is faster in MoOthan through the metallic molybdenum [32]. The
x
formation rate of the molybdenum oxide will depend on the deposition parameters.
It is proposed that the stable character of the dense selenised sample can be explained
by the presence of MoSe on the surface and in the grain boundaries, which protects
2
the metallic molybdenum from oxidation.
When the samples are removed from the climate chamber and their temperature as well
as the humidity is quickly decreased, or just in the climate chamber itself, the oxide layers
can crack (step 3). The cracking of the molybdenum oxide layer is proposed to be due to
-1
the difference in thermal expansion coefficients between the glass (α slg : 8.6·10 -6 K ) [33],
-6
-1
-1
the molybdenum (α : 4.8·10 K ) [33] and the molybdenum oxide (α MoO3 : 18-90·10 -6 K )
Mo
[34], which becomes important after a temperature change from 85C to room
o
temperature. Furthermore, evapouration of water diffused in the molybdenum layers
could also lead to breaking of the molybdenum oxide top layer. The cracking of the top
oxide layer then allows the exposure of fresh metallic molybdenum to the atmosphere,
so the degradation process can start over, leading to a multilayer stack consisting of
Mo/MoO /MoO (step 4). The needles on top of these layers, which may consist of
x
x
Na CO might have resulted from the migration of sodium from the glass through the
3
2
molybdenum surface, where it then reacted with oxygen and carbon from the air.
163
edges of a CIGS cell. The non-selenised samples already showed visual degradation
and a decrease of the reflectance after two hours.
Within the selenised samples, there was also an impact of the sputter pressure on the
reflectance. Mo2Se retained only spot-like degradation and was still quite reflective
after 105 hours of degradation, while on Mo10Se and Mo15Se the spots grew into a
mosaic-like structure on the surface, which led to a decrease in reflectivity. Mo2Se
was the only sample that retained a reflectance over 10%. Therefore, considering the
optical properties, dense selenised molybdenum layers are also the most suitable for
incorporation in CIGS solar cells.
5.4.1.4 Degradation route
Figure 5.29 shows a possible degradation route for molybdenum thin film layers. Due
to damp heat exposure of molybdenum (step 1), an oxide layer is formed on top of the
molybdenum sample. The same effect might occur in the grain boundaries (step 2).
It was observed in cross-section SEM that porous layers had thicker top layers than
dense samples, which indicates that these layers degrade the fastest. This can be
attributed to the faster diffusion of water and other molecules like O and CO in
2
2
molybdenum with a higher volume of intergranular material, since the mobility of for
example sodium is faster in MoOthan through the metallic molybdenum [32]. The
x
formation rate of the molybdenum oxide will depend on the deposition parameters.
It is proposed that the stable character of the dense selenised sample can be explained
by the presence of MoSe on the surface and in the grain boundaries, which protects
2
the metallic molybdenum from oxidation.
When the samples are removed from the climate chamber and their temperature as well
as the humidity is quickly decreased, or just in the climate chamber itself, the oxide layers
can crack (step 3). The cracking of the molybdenum oxide layer is proposed to be due to
-1
the difference in thermal expansion coefficients between the glass (α slg : 8.6·10 -6 K ) [33],
-6
-1
-1
the molybdenum (α : 4.8·10 K ) [33] and the molybdenum oxide (α MoO3 : 18-90·10 -6 K )
Mo
[34], which becomes important after a temperature change from 85C to room
o
temperature. Furthermore, evapouration of water diffused in the molybdenum layers
could also lead to breaking of the molybdenum oxide top layer. The cracking of the top
oxide layer then allows the exposure of fresh metallic molybdenum to the atmosphere,
so the degradation process can start over, leading to a multilayer stack consisting of
Mo/MoO /MoO (step 4). The needles on top of these layers, which may consist of
x
x
Na CO might have resulted from the migration of sodium from the glass through the
3
2
molybdenum surface, where it then reacted with oxygen and carbon from the air.
163
