link to homepage

Institute of Energy and Climate Research

Navigation and service

Luminescence as a characterization technique for solar cells

S.Solar cells are semiconductor diodes converting optical energy (photons) in electrical energy while light emitting diodes convert electrical energy into photons. Every semiconductor diode can in principle be used as a solar cell or as a light emitting diode. However, a good light emitting diode is not necessarily a good solar cell or vice versa. The emission of a light emitting diode is termed electroluminescence (EL), because the free charge carriers are injected electrically. Even if a solar cell is a very bad light emitting diode, the EL spectrum will provide information about the photovoltaic properties of the diode. This is due to the principle of detailed balance stating that every elementary process needs to be in equilibrium with its inverse process if the whole system is in thermal equilibrium. Therefore, detailed balance is helpful when connecting two situations that are identical if we exchange the input and output terminals of the system.

The EL of a solar cell can be helpful in particular, when spatially or energetically resolved information about the solar cell is desired. Solar cells or modules typically are rather thin large area devices. Therefore it is both important and challenging to localize problems in photovoltaic devices in order to identify them. The most direct way to gain spatially resolved information about a solar cell is the light beam induced current measurement, where a laser scans over the solar cell while the short-circuit current is measured at the same time. The main disadvantage of this method is that it takes a long time (hours), especially for larger cells or modules. It is substantially faster to apply a forward voltage to the device and then to photograph the device at the same time (in an otherwise dark room/box). The camera will record the EL emitted from the solar cell and provide an image of the device that gives information on the electronic quality within typically less than a second.

Figure 1 shows electro- and photoluminescence images of a part of a multicrystalline Si solar cell featuring various different defects. The EL measurement in (a) shows dark areas around grain boundaries and dislocations because in these areas more non-radiative processes are happening, thereby reducing the percentage of electrons that will be able to recombine radiatively. In addition, there are vertical and horizontal stripes that originate from the metal front contacts that shadow part of the emission from the Si underneath. More interesting are the cracks in the lower and in the right part of the image and that lead to a roughly triangular shaped black area on the right. This area is electrically isolated due to the cracks in the wafer and the metal front contact. Due to these cracks, injected electrons and holes cannot get into that area to recombine there. The ability to detect these cracks is quite useful because they cannot be detected by eye or in a normal photograph without applied voltage. EL images are the fastest way to detect them and to avoid that affected solar cells are used for module fabrication (one defective cell would mean the whole module goes to the bin).

Figures 1b and c show two photoluminescence (PL) image of the same spot one of them taken under (b) open circuit and one under (c) short circuit conditions. PL in open circuit is similar to EL with the exception that the electrically isolated area on the right is still bright. This is because with optical injection, the cracks no longer inhibit the creation of electron hole pairs in that area. The short circuit situation is however substantially different. The image is roughly the opposite of the EL image shown in (a). This time, the isolated area is bright and everything else is dark. This is because in short circuit the optically created electron hole pairs are collected efficiently except for the isolated region where this is not possible due to the cracks.

µc-Si EL 001

Fig. 1a: EL image of a multicrystalline Si solar cell showing grain boundaries and dislocations as black shadows, the contact fingers and the busbar as horizontal and vertical black lines or stripes and the cracks as irregular black lines that isolate the large black triangular area on the right.

µc-Si PL 002

Fig. 1b: Photoluminescence image of the same spot in open circuit. Due to the optical excitation, the electrically isolated area on the right is bright here instead.

µc-Si PL 003

Fig. 1c: PL image at short circuit. In a good solar cell, at short circuit all photogenerated charge carriers should be collected and the PL therefore should be negligibly low. In this case, the cracks isolating the area on the right will make sure that collection from this region is impossible and therefore the electrons and holes will recombine because they cannot go anywhere. Thus, the PL at short circuit is close to a negative of the EL image in a.

EL imaging allows us to obtain spatially resolved information about the device but initially there is no spectral information accessible from the images shown in Fig. 1. In order to find out what energy the emitted photons have, we need to measure EL spectra. An ordered semiconductor emits luminescence mainly close to its band gap. Therefore, the EL will give information about the band gap of the semiconductor. This is interesting in particular for organic solar cells where the different energy levels at the donor acceptor interface make the different band gaps in the system a very important property to study. Figure 2 shows the type II heterojunction in an organic solar cell. The most important energy to measure is neither the band gap of the donor nor of the acceptor but the interfacial transition between the energy of an electron on the acceptor and a hole on the donor. This transition - labelled as charge transfer emission in Fig. 2 – controls the open circuit voltage and its difference to the optical gaps of donor and acceptor control the photocurrent of the solar cell.


Fig. 2: Schematic diagram of an interface between polymer and fullerene in an organic solar cell. EL measurements provide the EL spectrum of the blend that differs from the EL spectrum of the pure polymer (blue) and the pure fullerene (not shown) substantially. Because the EL spectrum of the polymer:fullerene blend has a much lower energy than that of the polymer and fullerene alone, the spectrum must originate from the interface.