Page 82 - Mirjam-Theelen-Degradation-of-CIGS-solar-cells
P. 82
Chapter 2
metallic matrix. This is especially the case if the protective MoSe layer is removed
2
in the scribing process, leading to decreased protection and thus damaging of the
molybdenum surface. This oxidation can be detrimental for the module, but this is
not necessarily the case, since conductivity in the P3 area is only relevant in the later-
al direction. Therefore, a back contact with a non-conductive top layer and a highly
conductive molybdenum bottom layer can still function as back contact as long as the
conductive bottom layer is still thick enough for a sufficent conductivity. Furthermore,
the length of the exposed molydenum in the scribe is only tens of micrometres.
Wennerberg et al. [27] calculated that the sheet resistance of the molybdenum that is
present in P3 has to increase by three orders of magnitude (from approx. 0.3 Ω/ to
500 Ω/) to cause a fill factor loss of 10 %. Therefore, no effect of molybdenum degra -
dation will be seen in P3 until the contact is completely oxidised and thus interrupted.
Westin et al. [104] used a special test structure to study the influence of damp heat
exposure on P3 conductivity. The authors reported the very large impact of P3 corro-
sion. The back contact remained conductive for 100 to 200 hours, even when strong
evidence of a corrosive attack on the Mo surface was visible. After this time, the initial -
ly 450 nm thick molybdenum film was completely corroded, leading to a very large
resistance increase. Something similar was seen for test structures prepared to anal-
yse the loss of conductivity in CIGS modules: After 400 hours exposure to a damp heat
test, conductivity could no longer be measured. This coincided with the point in time
when the P3 isolation scribes became completely transparant. For several modules,
loss of connection in the P3 scribe area was detected. It was noticed that the surface
corrosion was the strongest along the centre of the scribe line, where the mechanical
tip had damaged the Mo.
Furthermore, it should be noticed that the popularity of laser scribing is increasing,
compared to the conventionally applied mechanical scribing. Laser scribing might in -
troduce new starting points for degradation.
2.5.1.4 Impact of module design
The design of the mini-module also impacts its degradation behaviour: Kempe et al.
looked at damp heat treated encapsulated mini-modules and observed that the main
changes occurred on the side of the cell adjacent to the P1 scribe [108]. These changes
were not attributed to ohmic losses, but to the module design ( Figure 2.14): as the cur-
rent crosses the cell in the zinc oxide front contact, its voltage drops, while the voltage
in the Mo back contact remains more constant because of the unchanged conductivity.
This results in the reduction in voltage drop across the cell when moving away from
the P1 scribe. Because of the strong voltage-current relationship of a diode, this small
voltage drop change creates a much large change in current crowding adjacent to the
80
metallic matrix. This is especially the case if the protective MoSe layer is removed
2
in the scribing process, leading to decreased protection and thus damaging of the
molybdenum surface. This oxidation can be detrimental for the module, but this is
not necessarily the case, since conductivity in the P3 area is only relevant in the later-
al direction. Therefore, a back contact with a non-conductive top layer and a highly
conductive molybdenum bottom layer can still function as back contact as long as the
conductive bottom layer is still thick enough for a sufficent conductivity. Furthermore,
the length of the exposed molydenum in the scribe is only tens of micrometres.
Wennerberg et al. [27] calculated that the sheet resistance of the molybdenum that is
present in P3 has to increase by three orders of magnitude (from approx. 0.3 Ω/ to
500 Ω/) to cause a fill factor loss of 10 %. Therefore, no effect of molybdenum degra -
dation will be seen in P3 until the contact is completely oxidised and thus interrupted.
Westin et al. [104] used a special test structure to study the influence of damp heat
exposure on P3 conductivity. The authors reported the very large impact of P3 corro-
sion. The back contact remained conductive for 100 to 200 hours, even when strong
evidence of a corrosive attack on the Mo surface was visible. After this time, the initial -
ly 450 nm thick molybdenum film was completely corroded, leading to a very large
resistance increase. Something similar was seen for test structures prepared to anal-
yse the loss of conductivity in CIGS modules: After 400 hours exposure to a damp heat
test, conductivity could no longer be measured. This coincided with the point in time
when the P3 isolation scribes became completely transparant. For several modules,
loss of connection in the P3 scribe area was detected. It was noticed that the surface
corrosion was the strongest along the centre of the scribe line, where the mechanical
tip had damaged the Mo.
Furthermore, it should be noticed that the popularity of laser scribing is increasing,
compared to the conventionally applied mechanical scribing. Laser scribing might in -
troduce new starting points for degradation.
2.5.1.4 Impact of module design
The design of the mini-module also impacts its degradation behaviour: Kempe et al.
looked at damp heat treated encapsulated mini-modules and observed that the main
changes occurred on the side of the cell adjacent to the P1 scribe [108]. These changes
were not attributed to ohmic losses, but to the module design ( Figure 2.14): as the cur-
rent crosses the cell in the zinc oxide front contact, its voltage drops, while the voltage
in the Mo back contact remains more constant because of the unchanged conductivity.
This results in the reduction in voltage drop across the cell when moving away from
the P1 scribe. Because of the strong voltage-current relationship of a diode, this small
voltage drop change creates a much large change in current crowding adjacent to the
80
