complimentary article
I have been thinking about this all day and I am mindfucked on the matter. There's two points of views I have developed -
The first one is simple, from a classical perspective light MUST always be pure work because you can focus its direction at more or less no energy cost, for example the old kerosene lamp. From this perspective it seems like light must by necessity be identified as having no heat content.
On the other hand we have a system where heat enter and work leaves, this is by definition an entropy lowering activity meaning that somehow the rest of the universe must have an entropic increase. So naturally we get the researcher's conclusion that light must be considered in some way entropic.
The best explanation I can come up with then is since the light is going to illuminate radially it must increase multiplicity overall by causing heating in a larger volume than the LED itself. Can it be regarded as entropic in itself then, or causing entropy within the universe?
At any rate this effect must not be large since this is only caused by thermal energy being taken in which means that light's entropic "effect" doesn't need to be large for this to be okay.
1) the junction is energized just to the conductance point, and no higher, so that thermal excitations can push electrons over the gap. (At rest state, they wouldn't be anywhere close.)
Right so the voltage is pulling just hard enough for thermal noise to knock the electron up. Thus why this only works at low voltages, makes sense.