618
A. Jariwala et al.
the SPV vector depends on the type of carriers with higher mobility in the case of
Dember effect domination.
The vector model is of great interest when more than one mechanism are involved
for generation of SPV in the spectrum run. For example, a sample with multi-layered
structure can have multiple sources of the SPV signals. All these sources can generate
small SPV vectors with various amplitude and phase and the effective SPV measured
by the experiment would be the vector sum of all these small vectors. All these small
vectors basically represent different absorption mechanisms in different layers and
substrate. It is possible to model such a situation by considering multiple SPV vectors
with different amplitudes and angels with the x-axis. The measured SPV signal then
would be the vector sum of all the different vectors corresponding to different SPV
generation processes at different region of the multilayered structures. Thus, a shift
in the phase of the measured SPV signal generally corresponds to either inclusion or
removal of one or more process contributing to the photovoltage.
Here it must be noted that the experimentally measured SPV signal is represented by the resulting overall vector, that is the vector sum of the all the vectors
representing different active processes. Hence, extracting quantitative information
about any individual processes is quite complicated and requires careful analysis
with other supporting characterizations. Also, it relies on an assumption that all the
active processes are independent of each other, which is a prerequisite for superposition of the vectors. However, in case if these conditions are not met, the model can
still be used to carry out the qualitative analysis and the same may be very useful in
quite a few instances. Also, it is important to calibrate the system for correct phase
measurements carefully before employing the same for analysis of real samples.
Particularly the inherent parasitic phases shifts occurring between the SPV signal
and the measurement by lockin or the phase shift between the optical chopper and
the reference signal fed to the lockin must be carefully normalized in order to make
reliable conclusions from the SPV phase spectra [56, 67].
We have also used the SPV phase spectra successfully to investigate the
effects of laser annealing on the Fermi-level pinning at the surface of CdTe films
[68]. The experiments were carried out on superstrate structure of CdTe device
(glass/transparent conducting oxide (TCO)/CdS/CdTe). The samples used in the
experiments have not been subjected to CdCl 2 treatment. Back contacts have also
not been applied to the samples. An Nd:YAG laser in CW mode was used to carry
out the laser annealing. The operating wavelength and the power density used for
the experiments were 1064 nm and 250 W/cm
2 , respectively. It must be noted here
that the CdTe samples showed strong absorption at this photon energy (1.17 eV) at
1064 nm despite of the fact that the energy is below the bandgap of CdTe. This could
be attributed to the absorption related to the defect states. SPS was carried out in
soft-contact mode at room temperature as illustrated in Fig. 5. The TCG is coupled
to the CdTe surface capacitively through an air gap. The TCG film was used as the
counter-electrode on the front (CdS) side. Figure 8 shows the SPV phase spectra of
the samples described above for un-annealed and laser annealed samples. An SPV
amplitude spectrum has also been plotted for unannealed film for references.
A. Jariwala et al.
the SPV vector depends on the type of carriers with higher mobility in the case of
Dember effect domination.
The vector model is of great interest when more than one mechanism are involved
for generation of SPV in the spectrum run. For example, a sample with multi-layered
structure can have multiple sources of the SPV signals. All these sources can generate
small SPV vectors with various amplitude and phase and the effective SPV measured
by the experiment would be the vector sum of all these small vectors. All these small
vectors basically represent different absorption mechanisms in different layers and
substrate. It is possible to model such a situation by considering multiple SPV vectors
with different amplitudes and angels with the x-axis. The measured SPV signal then
would be the vector sum of all the different vectors corresponding to different SPV
generation processes at different region of the multilayered structures. Thus, a shift
in the phase of the measured SPV signal generally corresponds to either inclusion or
removal of one or more process contributing to the photovoltage.
Here it must be noted that the experimentally measured SPV signal is represented by the resulting overall vector, that is the vector sum of the all the vectors
representing different active processes. Hence, extracting quantitative information
about any individual processes is quite complicated and requires careful analysis
with other supporting characterizations. Also, it relies on an assumption that all the
active processes are independent of each other, which is a prerequisite for superposition of the vectors. However, in case if these conditions are not met, the model can
still be used to carry out the qualitative analysis and the same may be very useful in
quite a few instances. Also, it is important to calibrate the system for correct phase
measurements carefully before employing the same for analysis of real samples.
Particularly the inherent parasitic phases shifts occurring between the SPV signal
and the measurement by lockin or the phase shift between the optical chopper and
the reference signal fed to the lockin must be carefully normalized in order to make
reliable conclusions from the SPV phase spectra [56, 67].
We have also used the SPV phase spectra successfully to investigate the
effects of laser annealing on the Fermi-level pinning at the surface of CdTe films
[68]. The experiments were carried out on superstrate structure of CdTe device
(glass/transparent conducting oxide (TCO)/CdS/CdTe). The samples used in the
experiments have not been subjected to CdCl 2 treatment. Back contacts have also
not been applied to the samples. An Nd:YAG laser in CW mode was used to carry
out the laser annealing. The operating wavelength and the power density used for
the experiments were 1064 nm and 250 W/cm
2 , respectively. It must be noted here
that the CdTe samples showed strong absorption at this photon energy (1.17 eV) at
1064 nm despite of the fact that the energy is below the bandgap of CdTe. This could
be attributed to the absorption related to the defect states. SPS was carried out in
soft-contact mode at room temperature as illustrated in Fig. 5. The TCG is coupled
to the CdTe surface capacitively through an air gap. The TCG film was used as the
counter-electrode on the front (CdS) side. Figure 8 shows the SPV phase spectra of
the samples described above for un-annealed and laser annealed samples. An SPV
amplitude spectrum has also been plotted for unannealed film for references.