Thursday, May 11, 2006

Your dose of quantum awesomeness

For some reason, a lot of our musician friends are also scientists. Dan is of course a physicist, Might Could is all grad students (two physicists, one chemist), and keyboard wunderkind Sharif studies and teaches chemistry. So it wasn't particularly weird, last open mic night, to see Dan explaining the quantum Zeno effect to Sharif before their set. What was funny, I thought, was that when some friend came over to say goodbye to Sharif, she looked knowingly at me and said "are you smiling and nodding? That's what I do." I tried to find some way to say "um, no, I understand what they're talking about, and you probably would too if you'd been here the whole time"... actually I said something almost exactly like that, since I don't know this girl so who cares about nice. So for the unknown girls of the future, let me go on record: there is nothing all that hard about getting a lay grasp on quantum mechanics. I'll probably never be able to calculate a wave function or a path integral, but having a rough idea of what's going on and why it's so cool? Not a big hurdle.

I know what you're thinking... "but Jess, of course you had to go through some kind of rigorous training to be a physicist's girlfriend, and that's why you can talk about this stuff." Yes, of course, but listen: the secret is that physicists are actually closer to humanities people when they're talking about quantum than they are at any other time. It's mostly a matter of logic and metaphor.

The most important principle is that, at least until you measure them, photons don't or. They only and. Until you look at them they do all possible things. Then, of course, when you measure them, they're only doing one thing, so they must have been doing that. That's where the logic comes in -- if it's logically necessary that a photon was doing thing A to get the results you got, then the photon was doing thing A. The second most important principle is that mechanics at that level are so freaking weird that nobody can really discuss them without metaphors, except for nearly-crazy genius scientists. So if you have to talk about a photon "knowing" that you're looking at it (even though you're relatively sure it isn't conscious), or imagine it as a stream of light or a bouncy ball (even though it isn't), you're in good company.

Take the phenomenon Dan was explaining to Sharif, of quantum "seeing in the dark." The idea here is that you can "see" an object without any light ever touching it. How? Well, imagine that you have two beam splitters and two mirrors. They're set up in a square, such that a beam of light that goes in the beam splitter splits into two beams, at 90 degree angles from one another, e.g. one at zero degrees and one at 90, as in this image (call them the lower and upper paths respectively). Each of these beams hits a mirror, then bounces back at a 90 degree angle into another beam splitter, which combines them and sends out a single beam at the same angle at which it entered (zero degrees). The picture should help; I cribbed it, and the one below, from this reprint of a Scientific American article, because writing about quantum physics and refamiliarizing myself with Illustrator seemed like too much before lunch.

If the beam is not a beam but a single photon, it's basically the same. A photon entering the beam splitter has equal chances of being directed to the upper path or the lower path -- 50% either way. But of course a photon doesn't or; it ands. Since the chances are equal and you don't know which it will do, it actually goes on both the upper and lower paths, then combines with itself in the second beam splitter. There's a detector at the other end, so once it comes out, you know what it did. All fine and dandy -- photon in, photon out.

Now say there's an obstruction in the upper path. Physicists like to imagine it's explosive, just for kicks, so: there's an obstruction in the upper path, and if a photon hits it it will explode. Photon goes into beam splitter, same 50% chance of being directed each way... but if it goes on the upper path, BOOM. If, however, it goes on the lower path, it now hits the second beam splitter alone, instead of combining with itself there. So instead of combining, it splits, with again a 50% chance of going either way. As before, it actually does both until you find out which it did, but that's going to happen in just a second, because when it comes out of the beam splitter it goes into a detector and thus you know what it must have been doing. If it comes out at zero degrees (the "light port," or what the image calls "D-light," because it's the usual way for light to come out), you don't know anything about the exploding obstruction, since of course it would have done that anyway. BUT if it comes out at 90 degrees (the dark port), it must not have combined with itself at the second beam splitter, which means there was an obstruction. It never hit the obstruction -- if it had, there would have been an explosion. So it's not like it went in both directions but one was stopped or absorbed by the object. It must have gone on the lower path only, because that's the only way it could come out at the dark port. Note all the deductive reasoning here -- you didn't observe it, because you can't observe photons and have them do what they normally do (if you observe them, they only or). But you can reason it out, and whatever you deduce has to be what happened. In a certain sense, the photon went on both paths, because that is what photons do. But it can't have gone on both paths, it can't have gone on the upper path at all, because nothing exploded, and it must have gone on the lower path only because that's the only way it could come out the dark port. Reasoning equals history; reasoning equals observation.

The practical upshot here is that if a photon comes out the dark port, you can "see" that there is an obstruction, even though no light hit it. Unfortunately, that only works 25% of the time, and 50% of the time the whole thing blows up. How to improve the odds? It's too much for one post, so I'll address it in the next one.


Post a Comment

<< Home