Page 170 - Mirjam-Theelen-Degradation-of-CIGS-solar-cells
P. 170
Chapter 5
assigned to the conduction band, more specifically to inserted electrons occupying
the otherwise empty Mo 4d and 5s orbitals.
The presence of a non-zero density of states at E is a direct evidence of the metallic
F
character of these compounds. Similar structure at E to those previously observed for
F
KMoO compounds are clearly detected for degraded Mo25/2 and Mo25/15 (Figure
x
5.28a and b). These occupied states indicate a metallic behaviour which is typical
for the purple bronze materials. It can be noticed that a comparable structure has
been observed for titanium oxide [42]. The wide depleted region lying between the
conduction and valence bands is also typical for these transition metal oxides and has
already been detected in other molybdenum and tungsten bronzes [43]. According
to references [41,43], the largest structure extending below 4 eV corresponds to the
valence band which is largely representative of the Mo 4d states.
Nevertheless, a valence band signal corresponding to the insulator MoO oxide could
3
overlap the NaMoO one. In order to determine the fraction of Na intercalated MoO ,
+
x
3
one can consider the Mo 3d spectra ( Figure 5.27). Swiatowska et al. [44] have observed
6+
5+
that the intercalation of lithium in MoO leads to a partial reduction of Mo to Mo .
3
Based on Mo 3d-Se 3s signals (Table 5.9 and Figure 5.27), where a Mo signal is clearly
5+
detected for both types of degraded molybdenum, it is assumed that a comparable
intercalation process takes place, confirming the observation of a conduction band for
the degraded molybdenum (Figure 5.27). The average valence state of molybdenum
5+
6+
for Na Mo O is closed to 5.5 meaning a r=Mo /(Mo +Mo ) ratio close to 0.5. Our
5+
0.9
6
17
5+
5+
6+
calculated Mo /(Mo +Mo ) ratio is ~ 0.25. This r-value would indicate that the
+
samples do not have the standard structure, but contain less Na. This would imply
that the formation of a material with the composition NaMoO occurred. However,
4
23
another more likely explanation is that only a part of MoO is intercalated. Therefore,
3
we propose that the degradation product is MoO with and without intercalated
3
Na . As the r values are similar for both molybdenum degraded samples (Mo25/2
+
and Mo25/15), the formed intercalated molybdenum oxides have the same average
composition. is present in both samples.
It should be noted that conductivity measurements showed the disappearance of
o
conductivity after 150 hours at 85C/85% RH. This can be explained by either a low
Na content in the complete or part of the molybdenum oxide layer: since XPS is
+
very sensitive for the material on the top nanometers and does not measure the bulk
material, the bulk might contain higher quantities of MoO which is not or only slightly
3
168
assigned to the conduction band, more specifically to inserted electrons occupying
the otherwise empty Mo 4d and 5s orbitals.
The presence of a non-zero density of states at E is a direct evidence of the metallic
F
character of these compounds. Similar structure at E to those previously observed for
F
KMoO compounds are clearly detected for degraded Mo25/2 and Mo25/15 (Figure
x
5.28a and b). These occupied states indicate a metallic behaviour which is typical
for the purple bronze materials. It can be noticed that a comparable structure has
been observed for titanium oxide [42]. The wide depleted region lying between the
conduction and valence bands is also typical for these transition metal oxides and has
already been detected in other molybdenum and tungsten bronzes [43]. According
to references [41,43], the largest structure extending below 4 eV corresponds to the
valence band which is largely representative of the Mo 4d states.
Nevertheless, a valence band signal corresponding to the insulator MoO oxide could
3
overlap the NaMoO one. In order to determine the fraction of Na intercalated MoO ,
+
x
3
one can consider the Mo 3d spectra ( Figure 5.27). Swiatowska et al. [44] have observed
6+
5+
that the intercalation of lithium in MoO leads to a partial reduction of Mo to Mo .
3
Based on Mo 3d-Se 3s signals (Table 5.9 and Figure 5.27), where a Mo signal is clearly
5+
detected for both types of degraded molybdenum, it is assumed that a comparable
intercalation process takes place, confirming the observation of a conduction band for
the degraded molybdenum (Figure 5.27). The average valence state of molybdenum
5+
6+
for Na Mo O is closed to 5.5 meaning a r=Mo /(Mo +Mo ) ratio close to 0.5. Our
5+
0.9
6
17
5+
5+
6+
calculated Mo /(Mo +Mo ) ratio is ~ 0.25. This r-value would indicate that the
+
samples do not have the standard structure, but contain less Na. This would imply
that the formation of a material with the composition NaMoO occurred. However,
4
23
another more likely explanation is that only a part of MoO is intercalated. Therefore,
3
we propose that the degradation product is MoO with and without intercalated
3
Na . As the r values are similar for both molybdenum degraded samples (Mo25/2
+
and Mo25/15), the formed intercalated molybdenum oxides have the same average
composition. is present in both samples.
It should be noted that conductivity measurements showed the disappearance of
o
conductivity after 150 hours at 85C/85% RH. This can be explained by either a low
Na content in the complete or part of the molybdenum oxide layer: since XPS is
+
very sensitive for the material on the top nanometers and does not measure the bulk
material, the bulk might contain higher quantities of MoO which is not or only slightly
3
168
