Tell me the light you absorb, I will tell you the light you emit
This law has since been extended in several ways. Nowadays, one can deal with near-field sources, to which the concept of “direction of illumination” is not relevant. It has also been extended to deal with other light sources such as luminescent dyes and electroluminescent devices (e.g. LEDs).
When an object is illuminated, it will absorb part of the impinging light and heat up. Similarly, if an object is hot, it will emit light and cool down. The more efficient is an object in absorbing light, the more efficient it will be at emitting ligth if it is heated up. This relation between emission and absorption has been observed and quantified by Kirchhoff and Planck long time ago: for each frequency, the intensity of the light emitted in a given direction by an object whose temperature is uniform is directly proportional to amount of light the same object absorbs when illuminated from that direction.
In this project, we went one step further by considering two problems that have received little attention so far. First, we studied the coherence of the light emitted by luminescent and incandescent sources. Indeed, the emission of light by such sources arises from random processes. Despite the randomness of the source, the light emitted in different direction or at different positions may be partially or even fully correlated. Hence, we studied the link that exists between the correlation function of the light emitted and the correlation function of the light absorbed by such emitters.
Second, we focused the study on a specific type of emitters: electroluminescent devices such as Light Emitting Diodes (LEDs) or photovoltaic cells. If an electric current is passing throught the devices, part of the electric power dissipated by the device will be emitted as light. And if the device is illuminated by light, part of the absorbed power can be collected as an electrical current. Relations between the emission and absorption has already been derived for particular devices under restrictive hypotheses. The second point of our study consisted in identifying more general conditions under which such reciprocity is valid.
The details of our findings are available in the following papers: