The work of REG(ISFH) was structured in two main scientific topics: a) study of the impact of the contacting scheme on the fill factor measurement of untabbed bare solar cells and b) determination of the uncertainty of the approximations in differential spectral responsivity measurements as defined in the IEC 60904-8 standard.
In order to study the impact of contacting geometries on the current-voltage (I-V) characteristic of bare untabbed solar cells, we performed analytical and numerical modelling on a wide variety of different contacting geometries. The advantage of analytical modelling is the speed of the calculation. This allowed us to study a larger number of different contacting schemes and busbar conductivities. Moreover, it enabled us to perform an optimisation approach to find an ideal contacting scheme. The use of numerical device simulation allowed us to take probe-to-probe variations of the contact resistance into account. These variations cause asymmetric potential distributions along the busbar that are difficult to treat analytically. In order to allow for an experimental verification we designed a freely configurable contacting bar. This contacting bar allowed the experimental verification of eight different schemes by using each contact probe either as a current or sensing probe. Modelling and measurement results consistently showed large FF differences between individual contacting layouts. These differences amount to up to 3 %abs in the case of high busbar resistivity of up to 40 Ω/m. We analysed the contacting geometries for their sensitivity on uncontrolled variations of the contacting resistances. In this analysis we found that using triplet rather than tandem configurations and using a larger number of test probes reduced the impact of varying contacting resistances to below 0.02 %abs. However, a large number of test probes also increased the mechanical stress applied to the cell. We proposed a contacting geometry that we consider suitable for calibrated I-V measurements. This contacting scheme is a triplet configuration with a total of 5 triplets consisting of two current probes and one sensing probe. The sensing probe is positioned to measure the average busbar potential in between the current probes. This is the optimal contacting geometry in terms of a low sensitivity to the busbar resistivity and as well as to variations of contacting resistances. In addition, this geometry does not impose unnecessarily large mechanical stresses to the cell under measurement.
In 2015 a new version of the IEC 60904-8 standard was published. The standard specifies the requirements for the measurement of the spectral responsivity (SR) of linear and non-linear photovoltaic devices. The spectral responsivity is used in solar cell calibration measurements to calculate the spectral mismatch factor (SMF). The SMF correction is required to avoid erroneous short circuit current values by accounting for deviations between the solar simulator and the solar reference spectral irradiance as well as for differences in the spectral responsivity of the reference cell and the cell under test. For the highest accuracy, the IEC 60904-8 standard demands us to measure the differential spectral responsivity (DSR) of a cell under at least 5 different bias light irradiances resulting in short circuit currents between 5 % and 110 % of the short circuit current under standard test conditions. Apart from this complete differential spectral responsivity approach, the new standard names four simplified procedures. For solar cells featuring a nonlinear response of the short circuit current with irradiance, the spectral responsivity determined from the complete and the approximated approach might deviate. In order to quantify potential deviations, we carried out solar cell device simulations using SENTAURUS devices. We assumed a non-linear crystalline silicon solar cell featuring a bias-dependent bulk and rear surface recombination. We realised this by simulating a front junction passivated emitter and rear (PERC) cell based on p-typ Cz-silicon. The analysis according to the approximations described in IEC 60904-8:2014 Edition 3.0 reveals a deviation between 0.2 % and 2.1 % depending on the wavelength used for the bias ramp between 900 nm and 1100 nm and assuming an uncertainty in the intensity of the bias light level below 0.1 %. The DSR setup at the ISFH Calibration and Test Centre (ISFH-CalTeC) was used to experimentally verify these results. As an example for a non-linear cell, we used a specially prepared industrial-type monocrystalline p-type Czochralski (Cz) silicon passivated emitter and a rear (PERC) type solar cell. The standard version of this cell structure comprised a rear side passivation stack consisting of an aluminium oxide and a silicon nitride layer. In contrast to this, the aluminium oxide layer was deliberately omitted forcing the rear surface recombination to become injection dependent. Consequently, the cell becomes considerably non-linear. The measurements showed full consistency with the simulation.
In order to study the impact of solar cell or light field inhomogeneity on the characteristic solar cell parameters extracted from I-V measurements, different long and short pass filters had been selected. 10 filters of each type of size (10x10) mm2 were ordered. The long and short pass filter allow the simulation of enhanced front and rear carrier generation whereas the neutral density filter allowed the simulation of an inhomogeneous light field. For a different pattern resulting in an inhomogeneity of the light field of approximately 9%, the fill factor variations are found to lie within the typical fill factor uncertainty only