Page 115 - Mirjam-Theelen-Degradation-of-CIGS-solar-cells
P. 115
Temperature dependency of the electrical parameters of CIGS solar cells
and n, should be executed with care. Additionally, the resistances are calculated over
different areas, so comparison of the absolute values of the resistances is not possible.
The temperature dependencies as shown in Figure 4.3 give a representative picture
for most measured samples. Next to the trend (negative or positive) and the magnitude
of the change, these figures also shows how precisely the temperature dependency
can be determined for the PV parameters: While the temperature – open circuit
2
voltage relationship is completely linear with a very high correlation factor (R>0.998
for all samples) in this temperature zone, other relationships like temperature - series
resistance are less linear. It was observed that the series resistance can in most cases
be better fitted with a quadratic (aT+bT+c) than a linear function (aT+b), while the
2
saturation current can be best fitted with an exponential curve. The shunt resistance
showed a linear decrease for all samples, although the PI samples in some cases showed
o
an increase for temperature between 20 and 40C. Based on these observations, the
temperature dependency of all 42 samples were determined by fitting of the curves.
In most cases, a linear relationship was assumed, thereby calculating the temperature
coefficient (a) following:
X(T )=aT +b (4.1)
cell
cell
where X can be any of the electrical parameters of the cell and T cell is cell’s temperature.
When an quadratic or an exponential function is assumed, then a and b are shown
following:
X(T )=aT 2 cell +bT +c (used for series resistance) (4.2)
cell
cell
bTcell
X(T ) = a x e (used for saturation current density) (4.3)
cell
For the linear fits, the average temperature coefficient a plus the standard deviation
per batch are shown in Table 4.2. For the quadratic and exponential fits, both a and
b and their standard deviations are depicted. For the series resistance, parameters
based on both a quadratic and a linear fit are shown. For this parameter, both fits are
purely empirical and since we cannot determine the physical relevance of the fit, so
we cannot judge which fit is better.
The temperature dependencies of the efficiency, open circuit voltage, short circuit
current density and fill factor plotted against their values at room temperature are
depicted in Figure 4.4.
Based on Figure 4.3, Table 4.2 and Figure 4.4, it can be concluded that reproducible
and reliable trends can be observed for the temperature coefficients of the open
113
and n, should be executed with care. Additionally, the resistances are calculated over
different areas, so comparison of the absolute values of the resistances is not possible.
The temperature dependencies as shown in Figure 4.3 give a representative picture
for most measured samples. Next to the trend (negative or positive) and the magnitude
of the change, these figures also shows how precisely the temperature dependency
can be determined for the PV parameters: While the temperature – open circuit
2
voltage relationship is completely linear with a very high correlation factor (R>0.998
for all samples) in this temperature zone, other relationships like temperature - series
resistance are less linear. It was observed that the series resistance can in most cases
be better fitted with a quadratic (aT+bT+c) than a linear function (aT+b), while the
2
saturation current can be best fitted with an exponential curve. The shunt resistance
showed a linear decrease for all samples, although the PI samples in some cases showed
o
an increase for temperature between 20 and 40C. Based on these observations, the
temperature dependency of all 42 samples were determined by fitting of the curves.
In most cases, a linear relationship was assumed, thereby calculating the temperature
coefficient (a) following:
X(T )=aT +b (4.1)
cell
cell
where X can be any of the electrical parameters of the cell and T cell is cell’s temperature.
When an quadratic or an exponential function is assumed, then a and b are shown
following:
X(T )=aT 2 cell +bT +c (used for series resistance) (4.2)
cell
cell
bTcell
X(T ) = a x e (used for saturation current density) (4.3)
cell
For the linear fits, the average temperature coefficient a plus the standard deviation
per batch are shown in Table 4.2. For the quadratic and exponential fits, both a and
b and their standard deviations are depicted. For the series resistance, parameters
based on both a quadratic and a linear fit are shown. For this parameter, both fits are
purely empirical and since we cannot determine the physical relevance of the fit, so
we cannot judge which fit is better.
The temperature dependencies of the efficiency, open circuit voltage, short circuit
current density and fill factor plotted against their values at room temperature are
depicted in Figure 4.4.
Based on Figure 4.3, Table 4.2 and Figure 4.4, it can be concluded that reproducible
and reliable trends can be observed for the temperature coefficients of the open
113
