Page 162 - Mirjam-Theelen-Degradation-of-CIGS-solar-cells
P. 162
Chapter 5
In the non-degraded Mo25/2 spectrum, the Se 4s core level is clearly present, as well
as an valence band signal (0-10 eV) which is similar to the signals previously recorded
for MoSe [13]. For the degraded porous Mo25/15, the information from the survey
2
spectrum was confirmed: the Se 4s peak has vanished and the O 2s signal is raised,
confirming the almost complete oxidation of the MoSe layer. When the dense Mo25/2
2
is considered after degradation, the Se 4s signal is still detected, which indicates that
part of the MoSe layer has not been oxidised.
2
The identification of the valence band part for the degraded samples was more
complicated. In order to separate the signal of MoSe from the other present species,
2
the MoSe signals as obtained from the non-degraded sample were subtracted from
2
these spectra by the following treatment:
1. The contribution of MoSe has been determined through the Se 4s area relative
2
to the pure MoSe spectrum (non-degraded Mo25/2).
2
2. Subtraction of the MoSe contribution from the full spectra.
2
3. Normalisation of the spectrum of Mo2 to get similar counts as Mo25/15
The new spectra of Mo25/2 and Mo25/15 are plotted together as shown in the Figure
5.28b. The resulting graphs are very similar for Mo25/2 and Mo25/15, which indicates
the reaction products have the same chemical composition. The main part of the
valence band between 3 and 10 eV binding energy is built from a mixture of O 2p
states with Mo 4d. Since the photoelectric cross-section of the Mo 4d states at Al
K energy is 16 times greater than that of O 2p states, the valence band spectrum
α
obtained by XPS is largely representative of the Mo 4d admixture into the valence
band. The signals between 3 and 10 eV can therefore be attributed to molybdenum
and oxygen in molybdenum oxide.
However, it must be highlighted that a small peak is detected at 1 eV, which is near
the Fermi level E. This indicates the presence of a conductive material. This latter
F
structure cannot be related to MoO or another molybdenum oxide due to their
3
insulating nature. Reference [31] showed that the valence band spectrum of MoO
3
showed no intensity near the Fermi level. Furthermore, it cannot be MoSe, since the
2
signal for this material has been subtracted. Therefore, it was concluded that another
conductive species must also be present in these samples.
160
In the non-degraded Mo25/2 spectrum, the Se 4s core level is clearly present, as well
as an valence band signal (0-10 eV) which is similar to the signals previously recorded
for MoSe [13]. For the degraded porous Mo25/15, the information from the survey
2
spectrum was confirmed: the Se 4s peak has vanished and the O 2s signal is raised,
confirming the almost complete oxidation of the MoSe layer. When the dense Mo25/2
2
is considered after degradation, the Se 4s signal is still detected, which indicates that
part of the MoSe layer has not been oxidised.
2
The identification of the valence band part for the degraded samples was more
complicated. In order to separate the signal of MoSe from the other present species,
2
the MoSe signals as obtained from the non-degraded sample were subtracted from
2
these spectra by the following treatment:
1. The contribution of MoSe has been determined through the Se 4s area relative
2
to the pure MoSe spectrum (non-degraded Mo25/2).
2
2. Subtraction of the MoSe contribution from the full spectra.
2
3. Normalisation of the spectrum of Mo2 to get similar counts as Mo25/15
The new spectra of Mo25/2 and Mo25/15 are plotted together as shown in the Figure
5.28b. The resulting graphs are very similar for Mo25/2 and Mo25/15, which indicates
the reaction products have the same chemical composition. The main part of the
valence band between 3 and 10 eV binding energy is built from a mixture of O 2p
states with Mo 4d. Since the photoelectric cross-section of the Mo 4d states at Al
K energy is 16 times greater than that of O 2p states, the valence band spectrum
α
obtained by XPS is largely representative of the Mo 4d admixture into the valence
band. The signals between 3 and 10 eV can therefore be attributed to molybdenum
and oxygen in molybdenum oxide.
However, it must be highlighted that a small peak is detected at 1 eV, which is near
the Fermi level E. This indicates the presence of a conductive material. This latter
F
structure cannot be related to MoO or another molybdenum oxide due to their
3
insulating nature. Reference [31] showed that the valence band spectrum of MoO
3
showed no intensity near the Fermi level. Furthermore, it cannot be MoSe, since the
2
signal for this material has been subtracted. Therefore, it was concluded that another
conductive species must also be present in these samples.
160
