Luminescence Imaging Explained
What is luminescence?
Luminescence is the emission of light that is observed upon an external excitation of specific materials such as semiconductors. Different types of luminescence are distinguished by the way that external excitation is achieved; a common application of luminescence are light emitting diodes (LED's) in which luminescence is stimulated by applying an external voltage. This specific form of luminescence is called Electroluminescence (EL). In Photoluminescence (PL) the luminescence is achieved by illuminating a sample with an external light source. The key advantage of PL over EL is that PL is a contact-less technique. PL also does not require any specific electronic device structure to achieve meaningful measurement results; this means that PL can be applied to samples at all stages of solar cell processing including unprocessed raw wafers and blocks.
Luminescence Imaging
In PL imaging the entire sample area is illuminated uniformly and the emission is captured with one single shot taken with a specialised camera. In other words, a photograph of the luminescence emission is taken. In contrast, existing and commercially available PL systems, that are used for example in the microelectronics industry for the characterisation of silicon wafers are based on PL mapping (i.e. pixel by pixel raster scanning) produce a two-dimensional distribution of the luminescence intensity. PL imaging is advantageous, since it is able to provide a high resolution image of the sample with very short total data acquisition times (see TYPICAL MEASUREMENT TIMES below).
Since the detected luminescence light is emitted from the bulk of the sample we say the luminescence image “looks into” the wafer and is more like an X-ray image rather than a conventional photograph of the surface of a sample. All luminescence imaging applications discussed here are performed under so-called steady state conditions, which means that the external light source is turned on and at constant intensity during the image acquisition (i.e. while the camera shutter is open).
Basics
The luminescence spectrum i.e. the distribution of colours in the emitted light depends strongly on the material. Most current BTi luminescence imaging applications do not distinguish the different colours in the emission spectrum but rather measure an integral intensity, i.e. the intensity summed over all colours to which the camera is sensitive. Theoretical analysis allows this integrated intensity to be correlated in a quantitative way with various important physical and electrical cell parameters. The camera that is used for PL imaging in the LIS-R1 series is sensitive only to the spectral range of luminescence corresponding to band-to-band recombination from Silicon, and is therefore directly correlated with the minority carrier lifetime. Spectral components in the luminescence at longer wavelengths that are related to specific defects within the material and that are investigated in some of the microelectronics applications mentioned above are not measured. This is important since such contributions to the measured integral luminescence signal would cause inaccuracies in the interpretation of the data.
The elegance of PL imaging is that the integral luminescence intensity is measured with a single camera shot and with a megapixel camera, allowing the measurement of the intensity variation across the sample with high lateral spatial resolution. The above quantitative correlation with physical and electrical cell parameters is then performed for each pixel in the image, allowing e.g. high resolution images of the minority carrier lifetime to be obtained. Minority carrier lifetime is one of the most important solar cell material parameters since it is directly related to cell efficiency. Importantly PL imaging allows such quantitative measurements to be performed on silicon wafers with exceptionally short measurement times. Other benefits adding to the simplicity of applying this method include: the sample can remain at room temperature, no specific sample preparation is required and images can be measured on samples with full rear metallisation.
Typical measurement times
After the “emitter diffusion” the effective minority carrier lifetime in solar grade multicrystalline silicon wafers is typically >10 microseconds. Using full frame readout (i.e. maximum spatial resolution of 1024 x 1024 pixels) BT Imaging's LIS-R1 tool measures a 6-inch 1-Ωcm p-type wafer with τeff >10μs with an exposure time of only 1s.
In a simplified model the luminescence intensity is proportional to the background doping ND/A of the wafer and to the minority carrier lifetime τeff. In as-cut wafers the effective minority carrier lifetime is substantially shorter than in diffused wafers and data acquisition times required to achieve sufficient signal to noise therefore increase to typically ten seconds. For high-resistivity wafers the required data acquisition times increase inversely proportional to the background doping level, i.e. the lower the doping concentration the longer the required data acquisition time. For the same reason data acquisition times are roughly three times higher for n-type wafers compared to p-type wafers with the same resistivity (an n-type wafer with the equivalent resistivity has about three times lower doping concentration).
In cases where long measurement times are required with full frame readout, the measurement times can be significantly reduced by pixel binning (e.g. combining 4x4 pixels into one single pixel) which however also reduces spatial resolution. Such binning is conveniently achieved via the LIS-R1 software and allows measurement times to be reduced to 1s in most practical cases.