Page 50 - Mirjam-Theelen-Degradation-of-CIGS-solar-cells
P. 50
Chapter 2
X-ray emission spectroscopy). The authors observed the formation of a sulphate spe-
cies at or near the ZnO/CdS and the ZnO/CIGSSe interfaces after damp heat treatment.
In both cases, sulphur atoms diffused into the ZnO layer and formed a sulphate species
there (such as ZnSO). In contrast, a series of similar samples that were tested under
4
dry heat conditions showed no evidence for sulphur oxidation or sulphate formation.
Thus Heske et al. concluded that the source of oxygen for the sulphate formation is
predominately the ambient humidity rather than the ZnO layer. The authors propose
that the losses in fill factor observed by [40] and the commonly observed increase in
ZnO resistivity are enhanced by a change in the electronic and chemical structure at
and near the ZnO/(CdS)/CISSe interface, as well as by the altered electronic properties
of the ZnO film containing a detectable fraction of ZnSO 'contamination' in the vicin-
4
ity of the interface [48].
Ramanathan et al. [53] described the impact of exposure to high temperature air and ar -
gon on borosilicate glass/i-ZnO/ZnO:Al and borosilicate glass/CdS/i-ZnO/Al:ZnO stacks.
The stack containing CdS showed an increase in sheet resistance from 18 Ω/ to 60 Ω/
o
due to 2 minutes annealing at 200 C, while the stack without CdS has a constant series
resistance. Optical measurements showed a decrease in absorption in the wavelengths
where CdS absorbs (380 up to 540 nm) due to the heat treatments. It is proposed that a
reaction between i-ZnO and CdS occurred, resulting in the formation of an intermediate
material with a higher band gap. Suggested candidates are Cd 1-x Zn S and ZnO S .
x
1-x x
This effect was also seen for full CIGSe cells, which showed an increase in blue re-
o
sponse of up to 20% in EQE due to heat treatment (2 minutes at 250 C). The treatment
also led to shift of the absorption edge to lower wavelengths, suggesting that the
absorber itself has an increased band gap as well, maybe due to sulphur diffusion
from CdS to CIGSe.
2.3.3.2 Degradation of alternative buffers
A growing number of alternatives for CBD CdS buffers have been introduced [52]. In
this chapter, the stability of these alternative buffers is described.
Allsop et al. [54] studied the damp heat stability of CIGS solar cell with spray-ILGAR
(ion layer gas reaction) deposited indium sulphide buffers. Encapsulated mini-mod-
ules without edge sealing, like PVB laminates, were exposed to approximately 1000
hours to a damp heat test. Mini-modules that had lower initial efficiencies due to poor
buffer layer homogeneity from an non-optimised process showed an approximately
30% higher degradation compared to those with initially higher efficiencies due to an
optimised homogeneous buffer.
Spiering et al. [55] compared CIGS mini-modules (15 x 30 cm²) with ALD indium sul-
phide buffers to those with CBD CdS buffers. The mini-modules were exposed to
48
X-ray emission spectroscopy). The authors observed the formation of a sulphate spe-
cies at or near the ZnO/CdS and the ZnO/CIGSSe interfaces after damp heat treatment.
In both cases, sulphur atoms diffused into the ZnO layer and formed a sulphate species
there (such as ZnSO). In contrast, a series of similar samples that were tested under
4
dry heat conditions showed no evidence for sulphur oxidation or sulphate formation.
Thus Heske et al. concluded that the source of oxygen for the sulphate formation is
predominately the ambient humidity rather than the ZnO layer. The authors propose
that the losses in fill factor observed by [40] and the commonly observed increase in
ZnO resistivity are enhanced by a change in the electronic and chemical structure at
and near the ZnO/(CdS)/CISSe interface, as well as by the altered electronic properties
of the ZnO film containing a detectable fraction of ZnSO 'contamination' in the vicin-
4
ity of the interface [48].
Ramanathan et al. [53] described the impact of exposure to high temperature air and ar -
gon on borosilicate glass/i-ZnO/ZnO:Al and borosilicate glass/CdS/i-ZnO/Al:ZnO stacks.
The stack containing CdS showed an increase in sheet resistance from 18 Ω/ to 60 Ω/
o
due to 2 minutes annealing at 200 C, while the stack without CdS has a constant series
resistance. Optical measurements showed a decrease in absorption in the wavelengths
where CdS absorbs (380 up to 540 nm) due to the heat treatments. It is proposed that a
reaction between i-ZnO and CdS occurred, resulting in the formation of an intermediate
material with a higher band gap. Suggested candidates are Cd 1-x Zn S and ZnO S .
x
1-x x
This effect was also seen for full CIGSe cells, which showed an increase in blue re-
o
sponse of up to 20% in EQE due to heat treatment (2 minutes at 250 C). The treatment
also led to shift of the absorption edge to lower wavelengths, suggesting that the
absorber itself has an increased band gap as well, maybe due to sulphur diffusion
from CdS to CIGSe.
2.3.3.2 Degradation of alternative buffers
A growing number of alternatives for CBD CdS buffers have been introduced [52]. In
this chapter, the stability of these alternative buffers is described.
Allsop et al. [54] studied the damp heat stability of CIGS solar cell with spray-ILGAR
(ion layer gas reaction) deposited indium sulphide buffers. Encapsulated mini-mod-
ules without edge sealing, like PVB laminates, were exposed to approximately 1000
hours to a damp heat test. Mini-modules that had lower initial efficiencies due to poor
buffer layer homogeneity from an non-optimised process showed an approximately
30% higher degradation compared to those with initially higher efficiencies due to an
optimised homogeneous buffer.
Spiering et al. [55] compared CIGS mini-modules (15 x 30 cm²) with ALD indium sul-
phide buffers to those with CBD CdS buffers. The mini-modules were exposed to
48
