Tuesday, May 22, 2007

Quantum Mechanics for Dummies #2: Observation

Well, it's been a while, but an e-mail exchange not long ago prompted me to get back to this series. If you haven't yet read it (or need a refresher), I recommend you go back and read my first post, on the Wave Nature of Matter. This time, I'm going to be talking about something that can happen to waves and is a very important part of Quantum Mechanics: the collapse (which is closely tied into the concepts of a measurement and an observation).

But before that, we need to cover the concept of eigenstates. (For anyone who already knows about it, this is a vastly simplified explanation, so you're probably safe just skipping past.) Unlike particles, waves don't have discrete positions they'll be in, but rather a range of possible positions they could be found at at any time. Since waves travel, the probability distribution of where it will end up will often change with time. However, there are ways to trap waves, such as a beam of light between two mirrors. In these cases, the wave can only take on discrete stable patterns so that it doesn't end up interfering with itself. These discrete patterns are what are known as eigenstates.

But the probability distribution for a wave won't always fall perfectly into one eigenstate. Often its probability distribution will be a linear sum of multiple eigenstates. Due to interference between the waves of the different states, this pattern won't be perfectly stable and will change somewhat with time, but it will generally do so in a periodic fashion.

This also isn't limited to cases such as light waves. For instance, all particles have a property known as "spin" (if you take that exactly as it sounds, you're close enough). If you measure the spin of a particle such as an electron, you will always get one of only two values, regardless of the orientation of your measuring device: +h/(4π) and -h/(4π) (called "spin up" and "spin down." Think of it as spinning clockwise or counterclockwise).

Both spin up and spin down are eigenstates of the electrons spin, but it's not necessary for the electron to be exactly in one of these states. To generate this, say you take an electron that you just measured to be spin up, then made a second measurement at a right angle to the first. Classically, you would expect to measure a spin of zero, but this isn't one the allowable results. Instead, you'll end up with a 50% chance to measure spin up and a 50% chance to measure spin down.

And here's where things get a bit strange: After you measure it once, if you go back and measure it the same way, you'll always get the same result. It's no longer a 50/50 split. However, if you go and measure it back in the initial direction, it's a 50/50 split here. The implication here is that by measuring the spin of the electron, you somehow changed it - in this case so that it was in a spin up eigenstate with respect to your new measurement.

What's happened here is known as wavefunction collapse. After your first measurement, the electron was spin up in the first direction, which corresponded to a 50/50 split when measuring from the second direction. This 50/50 split was a combination of two eigenstates for the electron. When you then measured it in this direction, one of those states was randomly selected and the electron then became 100% in that state.

And this when the misinterpretations start to happen. The reason for the misinterpretations is the fact that quantum physicists happened to use one particular word in describing it: "observation." In the sense of what happens, an observation simply entails measuring the wavefunction of something, causing said wavefunction to collapse.

But that's not how the word "observation" sounds to the layman. When many hear it, they then think, "So, does this mean that reality is unresolved until I look at it?" People started to believe that quantum states wouldn't resolve until the information from them had filtered its way to a human mind. Even when it came to quantum physics, people wanted to put the human mind on some special pedestal in the universe.

This argument wasn't limited to laymen however. At first, the best quantum theorists couldn't decide themselves what exactly caused the wavefunction to collapse. All they knew was that if they weren't measuring particles, the wavefunction wouldn't collapse, and if you were, then it would. (They tested this by means of the double-slit experiment, which I mentioned in my previous post in this series.)

So, you have scientists not knowing the answer and using a word which heavily implies that human consciousness actually is the answer, and what do you expect happens? People pick it up and start extrapolating, saying that we then must create reality with our minds and be able to control it as we see fit. Even if we were to accept the premise (that being processed by a human mind is what caused the collapse of a wavefunction), this in no way implies that the human mind actually creates or can control reality. The results are still inherently random, whatver the processing mind may wish.

Beyond that, there's actually good reason to believe that it isn't a human mind that causes the collapse of a wavefunction. Let's go back to the case of the double slit experiment, which is our best way of determining whether or not a wavefunction has collapsed. In this case, we'll be shooting electrons from an initial source through one of two slits. One foot beyond the slits is a detector screen. We know that if the electron's wavefunction is uncollapsed at the slits, we'll end up seeing an interference pattern on the screen, while if it's collapsed, we'll see the sum of two diffraction patterns.

If we just let the experiment run, without any detectors, we end up seeing the interference pattern (nothing's causing the wavefunction to collapse). If, instead, we put in detectors at each slit that will notify us if the electron passes through (say by blinking a light on the left or right side), we see the sum of diffraction patterns on the detector screen. Now, what if you were to try this: Have the detectors at the slits and turned on, but don't look at the lights. You could simply disable the feature that has it flashing the lights on the detectors and not store data of which slit the electron passed through, so no human could ever know which way it went.

In this case, we can expect to see one of two outcomes: Either we see an interference pattern, which means that without a human observing it, the wavefunction wouldn't collapse, or we could see a sum of diffraction patterns, which would mean the interaction of the electron with the detector (or some process within the detector after the detection) caused the collapse. This has in fact been done, many times. Very frequently, scientists did experiments using a detector but didn't care which slit was detected, and so they didn't set it up to tell them this data. The result of these tests? The wavefunction collapsed anyways, so human consciousness is not necessary to cause the collapse of a wavefunction.

Why then, do we still see people claiming that it is? Mostly it's do to a poor understanding of the subject. They see words like "observation" and interpret it to mean human observation. Even some professors of quantum physics (including one I had as an undergrad) made this mistake, and then taught it as accepted fact to their students. The only way you can really know for sure on something like this is to go to the experiments themselves. Fortunately, this particular phenomenon is testable, and it has been tested.* Unfortunately, this won't stop some people; I raised this issue up with my professor at one point, and he said testing it was a waste of time because he knew that human consciousness had to be involved. Well, you can't convince everyone.

*Edit to add: Unfortunately, no one seems to have a link to the studies that actually test this, most likely because they were done so long ago that they were never published online anywhere. Actual science has moved on far past this point, while popular science is just starting to get interested in it. I've asked a few people in the know about it, and while some, like the professor mentioned above, still hold to the view that it's human consciousness that does the collapsing, the consensus is pretty clear that the collapse happens long before any human mind looks at whether the particle was detected in either slit.

In fact, even when humans are looking at this, there's so much processing that goes on in the technology that interprets the data that the particle as already traveled and hit the plate at the end before information about which slit it passed through reaches the mind of a human (the "flashing light" is just a metaphor, it's not what's actually done in these experiments). So if it was human thought that was controlling these outcomes, then this result would also have to reach back in time to collapse the particle's wavefunction at some point before it hit the plate.

* * * * *

Other posts in this series:

Quantum Mechanics for Dummies #1: Wave Nature of Matter


Nochte Elphi said...

Very informative and concise. I was not aware that the double slit experiment was re-run without concious observation.

Do you know if there is any head-way on what 'observation' means?

I also wonder what, if anything, effect these changed states of matter have on their environment. Sure they look different and produce different data, but does a particle care if its neighbor has coalesced or is still in a cloud of probability?

If the answer is no then the observations are not really changing anything at the fundamental level. Sure we are producing bands of interference on some developement paper, but that may only be relevant in a purely intellectual pursuit. The subatomic world may still go about its business and be none the wiser.

Simply put, the numbers we see may be different in these situations, but are the laws of the subatomic world dependant on the numbers we see?

Infophile said...

Do you know if there is any head-way on what 'observation' means?

I can't say for sure, but my guess guess is that an actual "observation" (wavefunction collapse) occurs when a non-virtual particle is created or destroyed. (Virtual particles are exchanged in the mediation of forces between real particles, and simply forces acting between particles doesn't cause collapse.) For instance, this could occur when a photon hits an electron and is absorbed by it or bounces off (bouncing off actually requires the absorption and reemission of a photon, so this would qualify as well).

I also wonder what, if anything, effect these changed states of matter have on their environment. Sure they look different and produce different data, but does a particle care if its neighbor has coalesced or is still in a cloud of probability?

I'd suspect that it does indeed matter, and that something like this would lead to a manner of entanglement between different particles. For instance, let's take an electron which is unresolved between spin-up and spin-down. This spin causes the electron to create a magnetic field which would interact with another electron which passes by, deflecting it in one of two directions. Since it's unresolved, the direction the second electron is deflected will also be unresolved.

Then, you can go and try to detect the second electron and determine what location it's in to see which way it was deflected. Since this was at first unresolved, you'd expect to see it sometimes going one way and sometimes going the other. Alternatively, you could directly measure the spin of the first electron and then determine which way the second must have gone.

Now, what if you measure both of them at the same time, but wait long enough after the interaction such that they're a great distance away from each other? It turns out that you'll get consistent results; the second electron was always deflected the "right" way, even though neither wavefunction was resolved. The interpretation here is that the two wavefunctions are what we call "entangled," that is, they're both determined by the same probability distribution and a measurement on one forces the other to collapse into a corresponding state. I went into this issue a bit more in-depth here.

As for your last question, it does indeed seem relevant that things work this way. The reason for this comes down to the fact that wavefunctions can "cheat" and do things particles can't do. If a classical particle encounters a barrier (a potential energy greater than its kinetic energy), it can never pass it. However, waves propagate differently then particles, and although they won't perfectly propagate through such a barrier, some of it will indeed leak through to the other side. Since particles that are left unobserved act as waves, this opens the door for particles to slip through these barriers. (This is what's commonly known as "Quantum Tunneling.")

Where is this relevant? Well, the phenomenon of tunneling is used often in circuit design in order to proportionately cut down the current flow, since we can predict with great accuracy how many electrons will travel through a given potential barrier. Tunneling also occurs within stars in order to allow nuclear fusion to take place. There's a huge potential barrier created by electromagnetic forces between particles, and a small volume behind it where the strong nuclear force can take effect and drop the potential off. Classical particles would almost never pass this barrier to fuse, but waves can and do, allowing stars to burn.

Nochte Elphi said...

So it seems to me that the world does potentially behave differently when it is 'being observed' as opposed to when it is not. We do have a comparable effect on these particles just by looking at them.

The question now is: why?

Infophile said...

"Why?" is one of those questions we can never really answer. Even if we do find the reasons for one layer of reality, it will just be based on the properties of a lower layer, which we can also ask "Why?" about. We can keep at this process, and we'll either have infinitely many layers and thus no end to the "Why?" or we'll have some end that just exists with no reason. Even this answer being God (or the deity of your choice) still leaves the unanswered question of "Why does God exist?"

In the end, things just are a certain way, and there can't really be any ultimate meaning to it.

Nochte Elphi said...

Perhaps that would be the highest example of Godel's Incompleteness Theorem.

He sure did stir up all sorts of trouble.

Anonymous said...

Now, what if you were to try this: Have the detectors at the slits and turned on, but don't look at the lights. You could simply disable the feature that has it flashing the lights on the detectors and not store data of which slit the electron passed through, so no human could ever know which way it went.

If you don't mind me saying so, that's a rather poor argument.

You spend three hours setting the experiment up, then, as long as you look away at the last moment, you've excluded consciousness?


The experiment is a conscious observation. A conscious decision is taken to set up the experiment - to detect the very phenomenon in question.


Infophile said...

Why should consciousness before the fact matter? The wave may be created by conscious interference, but it will collapse or not after that fact. You're really overextending what consciousness can supposedly do here, without any evidence that it can have these effects.

Let's look at a different argument: What was the universe like before consciousness first appeared? There were plenty of waves that would have needed resolution, but if consciousness is required for that, how could they ever resolve? And then at the end of that, how could consciousness form? The universe is a huge unresolved system, and it would require it to resolve to show consciousness before there would be consciousness to allow it to resolve.

Anonymous said...

>What was the universe like before consciousness first appeared?

What if it where the other way around - perhaps consciousness came before the universe? (ref Buddhism etc)

Infophile said...

Well, it's impossible to prove that it didn't work that way, but before legitimately considering it, we should get some decent evidence that it's true. As things are, all the best scientific evidence points to the conclusion that the universe was around a great deal before human consciousness arose (13.7 billion years or so).

Anonymous said...

Interesting, but it leaves me with more questions than answers. I often hear about this slit test, often accompanied by what I consider to be crazy theories about universes being created when I decide whether to have meatloaf or chicken for supper, but I have never seen any details of this test. What are these mysterious detectors? How do they work? At what speed are the particles being fired? Do they travel at the same speed after detection as before? What if detection occurs immediately after the particle leaves the firing device, rather than at the slits? Are the results the same? What happens if detection occurs before firing or midway down the "barrel", as it were? What happens if, instead of firing the particle, you drop single particles and let gravity do the "firing"? A "probability cloud" and particles which are, then aren't, then are once again particles for no other reason than they have or have not been measured is just too wacky a concept for me to accept until more simple explanations have been ruled out. A simpler explanation would be that the firing mechanism imparts some energy (which it definitely does in the case of an atom) on the particle, perhaps just momentum, perhaps a magnetic field, perhaps the particle rides on a wave also released by the firing device (I can only imagine magnetic fields being used to fire atoms and molecules) and the energy exchange cause by the detector (you simply cannot detect something without an exchange of energies, even in a passive way) cause a change in either the particle's orientation or energy or the wave energy in the immediate vicinity of the particle or, perhaps, in other particles hanging freely in space around the detector, which then interact with the particle directly. If I had to choose a more sane answer from either, a) it has no state before detection because it doesn't matter until it is checked and b) the process of detection interacts with it, causing a specific state as a side affect of the energy exchange necessary for detection, it's pretty much a no brainer which one is the saner answer here. So, as a laymen, why am I wrong? Why can't the explanation be just that simple?

Anonymous said...

'the process of detection interacts with it, causing a specific state as a side affect of the energy exchange necessary for detection'

Actually, I always wondered how this was taken into account when performing this type of experiment.

Anonymous said...

You have a great site. Waiting for more. Do we know WHY light travels the speed it does? I can understand why sound moves the speed it does, but c?

Unknown said...

I know nothing about this, I'm just a reader, I don't even speak english very well, but I wonder if anyone has studied how the mechanics of the human eye inteacts with the experiment in the observation process?

may the detector or the the human eye absorb part of the diffraction causing the wavefunction to collapse?

it maybe sounds stupid to you, but i'm just a reader, I'm here to learn.


Melynda said...

remember. observation does not mean human observation. and while your eyes do make measurements, anything that interacts with these particles will make 'observations'. humans have nothing to do with the outcome.

Anonymous said...

I read in another web site that when light is thrown over where the electrons form the interference pattern that the pattern disappears and that the electrons behave like particles. When the light is turned off, the pattern comes back. Is that you what are saying too? Also how do your detectors work? Do they throw lights onto the electrons for detection?

Anonymous said...

In this video http://www.youtube.com/watch?v=LW6Mq352f0E at about 2:40 man in the video explains that if detectrs are on, but no one collects the data, interference patters is get. So what he says, is that action of taking the data influences results.
I am confused. Is this man a nutjob? Thanks.