Measurement of Triple Junction PV Cells

Categories: Solar & Photovoltaics, R&D

Whilst standards1 exist for the evaluation of the spectral response of monolithic multiple junction (MJ) solar cells, the measurement procedure is far from standard.

Without such guidance, however, it is likely that in the case of subcells having low shunt resistance or low reverse breakdown voltage, significant measurement errors will result.2

In seeking to measure the spectral response/EQE of monolithic MJ solar cells, such as the III-V monolithic GaInP/GaInAs/Ge triple junction solar cell depicted here, the measurement of component subcells on an individual basis is not possible since they are epitaxially grown on one substrate and interconnected by tunnel diodes.

Subcell spectral response must therefore be determined by making use of the effect of current limitation, achieved through light biasing.

Light Biasing

In addition to biasing the subcell under test at a level of one sun to simulate use conditions, it is necessary to light bias the non-tested subcells at a higher intensity such that the former generates the smallest photocurrent and is therefore current limiting. It follows that good control of the light bias source spectrum is required. 

Subcell Interactions

Since the current of the MJ cell is limited by the subcell under test, it follows that the non-tested cells with excess photocurrent operate close to their Voc. If the MJ cell is tested under short circuit conditions, then a negative voltage approximately equal to the sum of the Voc of the other cells is placed across the tested subcell.

Whilst the photocurrent generated by light biasing may be considered as a constant, the presence of the monochromatic probe gives rise to changes in subcell operating voltage, depending on the probe wavelength and the response range of the current limiting subcell under test.

Where the subcell under test has non-ideal properties such as a low shunt resistance or a low reverse breakdown voltage, often exhibited in low bandgap materials such as germanium, this can lead to changes in the measured photocurrent, and the incorrect reporting of spectral response/EQE: the outcome depends largely on the true nature of the I-V curves of both the subcell under test and that of the non-tested subcells.

Outside the response range of the subcell under test, the monochromatic probe gives rise to an increase in operating voltage of the non-tested cells, compensated for by a reduction in the operating voltage of the subcell under test. This may give rise to an increase in Jsc, showing a response where one is not expected.

Within the subcell response range, the presence of the monochromatic probe will directly lead to an increase in Jsc, shifting the non-tested cells to lower operating voltage, and compensated by an increase in the voltage of the tested subcell. The resulting current may be less than that expected in response to the monochromatic probe, leading to a lower reported spectral response/EQE.

Both effects are commonly observed phenomenon in the measuring the bottom cell of the GaInP/GaInAs/Ge triple junction solar cell. 

Voltage Biasing

Shifting the external voltage of the cell will tend to minimise both of the above effects: in moving to higher gradient regions of the I-V curve of the non-tested cells, lower shifts in subcell operating voltage are encountered, giving rise to less variation in the current of the cell under test.

Optimisation of Light Biasing

Increasing the photocurrent generated by the non-tested subcells leads to increased gradient of the I-V curve in the proximity of their operating voltage, whilst reducing the photocurrent generated by the subcell under test will lead to a reduction of operating voltage closer to Voc where the gradient of the I-V curve is steepest.

Both will give rise to less variation in the current of the cell under test, and is achieved through appropriate filtering.

MJ Cell Example

For the correct measurement of the GaInP/GaInAs/Ge triple junction solar cell, the following procedure is recommended.

Current limiting may be verified by spectral response measurement at various levels of non-tested cell light bias intensity.

GaInP top junction

The GaInP junction responds ~300-700nm. The device under test should be illuminated simultaneously by a solar simulator at one sun bias, and a second simulator filtered with a red long-pass filter. The spectral response may then be measured directly over the extended range 300-800nm.

GaInAs middle junction

The GaInAs junction responds ~500-900nm. The device under test should be illuminated simultaneously by a solar simulator at one sun bias, and a second simulator filtered with a blue band pass filter, transmitting also in the infrared. The spectral response may then be measured directly over the extended range 300-1100nm.

Ge bottom junction

The Ge junction responds ~900-1800nm. The device under test should be illuminated simultaneously by a solar simulator at one sun bias, and a second simulator filtered with an IR rejection filter

To take this subcell to short circuit condition, an external voltage should be applied. Under dual solar simulator irradiation, the operating voltage, Vop of the top two junctions is recorded (typically ~2-2.5V). The voltage across the device should then be set to to -Vop to take the third junction into short circuit.

The spectral response may then be measured directly over the extended range 800-1800nm. Where the correct bias voltage is used, a maximum spectral response should be recorded. This may be verified by repeating the measurement at other levels of voltage bias.