628
A. Jariwala et al.
a hydrogen plasma performed following the deposition of –Si:H layer in the same
PECVD reactor.
We have also reported the investigations on the electronic properties of the CdTe
surface due to continuous-wave laser annealing using the SPS [68]. As discussed
earlier in the Sects. 4.1 and 4.2, we interpreted our SPS results using a model
consisting of two space charge regions, one at the surface and the other one at
the CdTe/CdS interface. The SPV amplitude signals measured in the MIS configuration at different chopping frequencies, as described earlier, is shown in Fig. 9. It
is evident that the SPV signal obtained near the bandgap of E g ≈ 1.5 eV, at which
the absorption depth of about 10 μm is larger than the CdTe thickness, will have
the contribution from SPV signals generated at both the SCRs. On the other hand,
the SPV signal obtained in the super-bandgap regime (for energies > E g ) only SCR1
should contribute due to very strong absorption within a very short distance of SCR1.
We found that the SCR1 contained a little negative charge because of the ionized
dopants, which is compensated by equal positive surface charge to ensure charge
neutrality. Similarly, a dipole also forms at the SCR2 owing to ions near the junction,
which are negatively charged and the compensating positive charges balancing the
charge neutrality on the n-type CdS/TCO side. The induced electric fields due to
photo-generated charges across these two SCRs are thus opposite to one another.
This can be visualized as if two reverse polarity diodes are connected in series. Thus,
when sub- or near-bandgap light, which can penetrate deeply, illuminates the surface
of CdTe and the heterojunction simultaneously, the SCR1 drags the photo-generated
electrons towards the surface. This results in cancellation of some positive surface
charges. Exactly similar but opposite behaviors can be assumed in SCR2. The resultant effect is reduced band bending at both regions. Consequently the net SPV signal
decreases at long wavelengths due to partial cancellation of the SPV generated at the
two SCRs. In case of super-bandgap region, however, the high absorption coefficient
causes the light to get absorbed in a very thin region near surface. This means that
the measured SPS signal in this regime is dominated by the changes in the band
bending at surface SCR1 and we can determine the sign of the SPV generated by
the surface. In the third scenario, where the surface defects have been significantly
passivated after the laser annealing treatment and the bands at the surface have more
or less flattened, only the SCR2 is left as an effective charge separating region. Thus
the major contribution to the net SPV comes from SCR2 in this case.
Figure 9 shows the SPV spectra of non-laser-treated (non-LT) and laser-treated
(LT) samples. The SPV spectrum measured on non-LT sample resembles quite similar
qualitatively to the modeled curve reported earlier in the literature [67] for a situation
of two opposing space charge regions in series. We also noted that the overall SPV
signals of SPV spectra for LT samples are significantly larger as compared to the
non-LT samples. The increased SPV signal indicates increased carrier collection
and bulk material quality, the observations similar to the behavior reported earlier
in the SPS on CdTe solar cell devices before and after CdCl 2 treatment [107]. The
main peak is result of a dramatic rise in SPV signal due to exponential increase in
the absorption coefficient as the bandgap of CdTe (≈1.5 eV) is approached from
lower energies. The local minimum near 1.51 eV in the non-LT samples is a direct
A. Jariwala et al.
a hydrogen plasma performed following the deposition of –Si:H layer in the same
PECVD reactor.
We have also reported the investigations on the electronic properties of the CdTe
surface due to continuous-wave laser annealing using the SPS [68]. As discussed
earlier in the Sects. 4.1 and 4.2, we interpreted our SPS results using a model
consisting of two space charge regions, one at the surface and the other one at
the CdTe/CdS interface. The SPV amplitude signals measured in the MIS configuration at different chopping frequencies, as described earlier, is shown in Fig. 9. It
is evident that the SPV signal obtained near the bandgap of E g ≈ 1.5 eV, at which
the absorption depth of about 10 μm is larger than the CdTe thickness, will have
the contribution from SPV signals generated at both the SCRs. On the other hand,
the SPV signal obtained in the super-bandgap regime (for energies > E g ) only SCR1
should contribute due to very strong absorption within a very short distance of SCR1.
We found that the SCR1 contained a little negative charge because of the ionized
dopants, which is compensated by equal positive surface charge to ensure charge
neutrality. Similarly, a dipole also forms at the SCR2 owing to ions near the junction,
which are negatively charged and the compensating positive charges balancing the
charge neutrality on the n-type CdS/TCO side. The induced electric fields due to
photo-generated charges across these two SCRs are thus opposite to one another.
This can be visualized as if two reverse polarity diodes are connected in series. Thus,
when sub- or near-bandgap light, which can penetrate deeply, illuminates the surface
of CdTe and the heterojunction simultaneously, the SCR1 drags the photo-generated
electrons towards the surface. This results in cancellation of some positive surface
charges. Exactly similar but opposite behaviors can be assumed in SCR2. The resultant effect is reduced band bending at both regions. Consequently the net SPV signal
decreases at long wavelengths due to partial cancellation of the SPV generated at the
two SCRs. In case of super-bandgap region, however, the high absorption coefficient
causes the light to get absorbed in a very thin region near surface. This means that
the measured SPS signal in this regime is dominated by the changes in the band
bending at surface SCR1 and we can determine the sign of the SPV generated by
the surface. In the third scenario, where the surface defects have been significantly
passivated after the laser annealing treatment and the bands at the surface have more
or less flattened, only the SCR2 is left as an effective charge separating region. Thus
the major contribution to the net SPV comes from SCR2 in this case.
Figure 9 shows the SPV spectra of non-laser-treated (non-LT) and laser-treated
(LT) samples. The SPV spectrum measured on non-LT sample resembles quite similar
qualitatively to the modeled curve reported earlier in the literature [67] for a situation
of two opposing space charge regions in series. We also noted that the overall SPV
signals of SPV spectra for LT samples are significantly larger as compared to the
non-LT samples. The increased SPV signal indicates increased carrier collection
and bulk material quality, the observations similar to the behavior reported earlier
in the SPS on CdTe solar cell devices before and after CdCl 2 treatment [107]. The
main peak is result of a dramatic rise in SPV signal due to exponential increase in
the absorption coefficient as the bandgap of CdTe (≈1.5 eV) is approached from
lower energies. The local minimum near 1.51 eV in the non-LT samples is a direct