5
$\begingroup$

In electronics a diode is a component allowing current passing in only one direction, and blocking the other side. I'm wondering if something similar exists for visible light or other EM waves, like radiowaves. Thanks

$\endgroup$
3
  • $\begingroup$ This is known as an optical isolator $\endgroup$
    – Brionius
    Feb 1, 2015 at 0:02
  • $\begingroup$ I think this is known as "Maxwell's demon". $\endgroup$
    – Hot Licks
    Feb 1, 2015 at 4:12
  • $\begingroup$ A surface that allows light to pass in one direction but not the other? I believe that is known as an event horizon $\endgroup$
    – Jim
    Feb 3, 2015 at 14:56

2 Answers 2

4
$\begingroup$

The optical isolator component is active. It consumes energy and so is no different (thermodynamically speaking) from the heat-pump in a refrigerator.

If you are talking about a passive component that draws no power then, a surface that allowed light pass in one direction only would violate the second law of thermodynamics.

To see why, imagine two rooms, perfectly insulated from each other and from the rest of the universe but connected via a window made of this surface. Now start off with both rooms at the same temperature. Heat radiation from room 1 would pass through the window into room 2 but could not pass back through. So room 1 would cool down and room 2 would heat up. You would have a perfect heat pump that did no work and that violates Clausius' statement of the 2nd Law, just for a start.

This process would continue until room 1 reached absolute zero and that would violate the 3rd Law as well.

You might be wondering about a "one-way mirror"? That is not the same thing. A one-way mirror is just a window with a high reflective co-efficient (e.g. it reflects 60% of the light that falls on it). If it is very bright on one side (the interrogation room) and very dim on the other (the viewing room) then from the bright side, the reflection off the window is much brighter than the transmission from the dim side and so that is washed out.

$\endgroup$
5
  • $\begingroup$ No, a ideal diode doesn't consume any power. It is simply a one-way valve for current, which doesn't require any external power to operate. Some specific implementations do take external power, but that's a engineering issue. $\endgroup$ Feb 3, 2015 at 14:41
  • $\begingroup$ Sorry. I meant that the light-valves are active. As you correctly state, a diode is a one-way valve for current and doesn't violate any thermodynamics laws since it is not pumping heat. I corrected my answer to be more precise. $\endgroup$ Feb 4, 2015 at 11:04
  • $\begingroup$ +1 Great answer Owen. I would be a little careful about the word "Optical Isolator", though: the optical isolator commonly used to shield lasers from back reflexions is not the same thing as the heat pump you rightly discuss: the former is passive - it only works with zero entropy (étendue) light, to wit purely polarised, so it doesn't violate the second law. Thermalised light passes through both ways freely. It blocks only the light in the polarisation of the outgoing laser's wave. $\endgroup$ Feb 4, 2015 at 11:42
  • $\begingroup$ I can't say how great your first sentence is though: you really cut to the chase categorising the device as a heat pump. $\endgroup$ Feb 4, 2015 at 11:42
  • $\begingroup$ Thanks indeed for the answer. Read it as soon you wrote it but I happened to log into SE from desktop only now, sorry for the late upvote. $\endgroup$ Jan 10, 2017 at 20:51
0
$\begingroup$

Something similar does exist. It's called a "directional light filter":

https://www.sciencenews.org/article/light-filter-lets-rays-through-only-one-direction

Note: I said "similar" as you asked in your question. This isn't exactly what you're looking for but it's step in the right direction.

This might be more on track. A "wave diode":

http://www.wired.com/2011/04/one-way-light/

http://physics.aps.org/story/v27/st15

$\endgroup$
2
  • $\begingroup$ Your first reference, at least was way off. $\endgroup$
    – Hot Licks
    Feb 1, 2015 at 4:09
  • $\begingroup$ @Hot Licks I'm sure the OP can figure that out for himself. I have no idea why he's asking and I'm sure he's open to other possibilities he may not have thought of. $\endgroup$ Feb 1, 2015 at 10:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.