Photovoltaic (PV) cells and Light Emitting Diodes (LEDs) are already part of our lives. Due to their widespread applications, they are still being intensively studied, looking for improved performance, reduced costs or new applications. Photovoltaic cells absorb light to generate electricity and have a major role in the generation of green electricity. LEDs are powered using electricity to efficiently generate light. LEDs are appreciated for their long lifetime and efficient energy conversion from electricity to light. Their very fast response to an injected current is also much appreciated for telecommunications.
What fewer people know is that a PV cell can be used to generate light if a current is forced through (be it in the invisible part of the spectrum), while LEDs can generate electricity if light is shone on them. The similarity between both technologies has already been quantified for some particular types of LEDs and PV cells: the light generated in a given direction by a device when it is switched on is related to the current it generates when light is shone on it from that direction. Such similarities are extremely useful in science since a vast amount of the work that has been done to improve light absorption by PV cells can be adapted to improve light emission by LEDs, and vice-versa.
The goal of this project was to extend the existing relation between photo-generated current and electroluminescent fields and use this extension to study light emission by LEDs. Indeed, it is known that the power absorbed by a body (and thus a PV cell) illuminated by two sources may be different if the sources are lit on sequentially or simultaneously, due to interference effects. To fully characterize an absorber, one needs to consider every possible pair of illuminations. The function describing the response of the structure to pairs of illuminations is called “mixed losses”. Similarly, when an electric current is forced through an LED, the light is randomly emitted from different locations within the active zone of the LED. These randomly generated fields are adding up to form the total generated light. To fully characterize the total fields, one needs to use a statistical description to handle both the randomness of the sources and the deterministic evolution of the fields: the fields are said to be partially coherent and can be described using a second order correlation tensor, the “mixed emission”. During this project, we theoretically showed that the existing relation between the absorption and emission of fields can be extended to the mixed losses and mixed emission of a device, providing a much richer information. We also investigated the conditions under which this relation is valid.
In order to validate our predictions, we have built a measurement system and experimentally observed the correspondence between the mixed losses and the mixed emission in commercial LEDs.
This result has important implications, especially concerning the modelling and design of luminescent sources such as LEDs. Indeed, most electromagnetic simulation software can compute the power absorbed by a structure when it is illuminated by a pair of sources. However, very few can compute the fields that are being emitted by luminescent sources. Using the reciprocity relation, we can predict the fields emitted by an LED and optimize them using traditional electromagnetic solvers. It opens new avenues to tune the frequency and spatial distribution of the fields emitted by luminescent sources, with important applications, among others, in telecommunications.
