Illumination Sources

Measuring solar cells requires a stable light source that closely matches the conditions of sunlight. Not only the intensity but also the spectrum must be matched to a standard. An obvious option is to simply use the sun itself. In locations where there is little cloud this is a good solution [1] but there are still variations in atmospheric conditions that require correction to compare measurements over time. The spectrum also changes through out the day and this further limits the time for testing.

The most common solution is to use an artificial light source that simulates the sun. The ideal illumination source would have following features [2];

  1. a spatial non uniformity of less than 1%.
  2. a variation in total irradiance with time of less than 1%,
  3. filtered for a given reference spectrum to have a spectral mismatch error of less than 1%.
  4. These requirements are essential in obtaining an accuracy of better than 2%

Testers are classified according to three criteria:

  1. Spectral match
  2. Irradiance inhomogeneity - spatial uniformity over the illumination area
  3. Temporal Instability - stability over time.

There are three classes within each of these criteria where 'A' is the top rating an 'C' is the lowest rating

Table: Solar simulator classification according to IEC 60904-9 Ed. 2.0.

Class Spectral Match Irradiance inhomogeneity Temporal Instability
  Long Term Short Term
A 0.75 - 1.25% 2% 0.5% 2%
B 0.6 - 1.4% 5% 2% 5%
C 0.4-2.0% 10% 10% 10%

For instance, a simulator with the designation ABA would have a spectral match < 1.25%, an inhomogeneity 2-5%, and stability < 0.5% in the long term and < 2% in the short term.

Many simulator related artifacts can be eliminated by using a reference cell that has the same spectral response as the cell under test.

The most common light source is a Xenon arc lamp with filters installed to approximate the AM1.5G spectrum. Simple testers often just use a halogen lamp with a dichroic filter. The lamp filament is much lower than the sun's 6000 K so it produces much more infrared light and much less UV. The reflector on the bulb is selective so that the visible and UV is reflected towards the cell but most of the infrared radiation isn't reflected and passes out the back of the bulb. Halogen lamps have the advantage of greater temporal stability compared to Xenon arc lamps.

ELH lamp with a dichroic reflector so that most of the infrared is not reflected. The ELH is an American Standards Institute (ANSI) code that describes this type of lamp.

Absorption Coefficient


The standard airmass 1.5 spectrum compared with the spectrums from typical solar simulator sources. Click on the graph for a scalable version.

There are also considerable variations between individual samples of the same type and with age.

Deviations from Air Mass 1.5

  • deviations from AM1.5 cause errors in Isc

It is difficult to make a light source that exactly matches the AM1.5 spectrum and with the necessary illumination intensity. As the previous graph shows there is often a considerable amount of variation between the spectrum of the lamp and the required AM 1.5 spectrum. There are two approaches for correcting for the diffirences between the AM1.5 spectrum and the actual spectrum from a solar simulator.

Calibration cell with the same spectral response.

The approach taken by most in-house testers is to use a calibration cell that has the same spectral response as the the cell under test. The light intensity of the tester is adjusted so that the cell Isc matches the Isc as measured at an external testing laboratory. However, slight changes in cell processing (e.g. the doping profile of the emitter, variation of anti-reflection coatings) cause changes in spectral response and the need for a new calibration standard.

Measure Spectral Response

Primary calibration labs use light sources that are much closer to the standard however differences still occur. To compensate for the differences, calibration labs measure the spectral response of the device under test and then use that to correct for the known difference between the spectrum of the light source and the standard spectrum [3]. Such corrections are time consuming and prone to error. In the early 90's an analysis of test error lead to an improved standard and also the down grading of some record efficiencies by up to 1% absolute [4].

There is an online calculator for determining the level of spectral mismatch from a light source at PV Lighthouse. It includes the ability to correct for an arbitrary spectrum.