Notes on quantum mechanics (under construction)
(updated 10/11/04)
NOTE: These are supplementary notes, you will still need to read your text
book.
Caution, I've used symbol font in a few places- if your browser does
not do symbol font these letters will show up as roman letters. I've tried
to avoid this as much as I could.
References:
- "Introduction to Quantum Mechanics" by David J. Griffiths.
- "The Fabric of the Cosmos" by Brian Greene.
The 'weirdness' of quantum mechanics.
Most of the difficulty in
understanding quantum mechanics comes because of its probabilistic
nature. For example, the electron wave function squared gives us the
probability of finding the electron at some location. Yet in some
cases we measure the particle like properties and actually find it in a
specific location (it arrives as a lump in a collision). How do we
reconcile the idea that the location is described by a probability yet
electrons can arrive in lumps as particles? The same problem is true of
photons and any other sub atomic particle (and even larger objects such
as buckyballs). The same is also true about other properties of a
subatomic particle such as polarization, spin, etc. (they have only a
probabilistic distribution of values until a measurement is made of
them, at which time we now know the specific value). As Griffiths and
Greene point out, when faced with dual wave/particle descriptions of an
electron we are tempted to take one of four positions:
- The realist position: The electron really is at a specific
location all the time, but quantum mechanics can't tell us precisely
where. Therefore quantum mechanics is incomplete, there must be hidden
variables which someday might be found and which will tell us where the
electron is all the time. This was Einstein's view and there is at
least one current group of physicists who are pursuing this idea, lead
by David Bohm (thus far inconclusively).
- The orthodox position: Until we actually make a measurement that
determines the particle like behavior the electron isn't anywhere in
particular. The observation of an electron in a particular location
causes its wave function to collapse; after measuring there is no
longer any probability but before measuring the electron doesn't exist
as a particle. This is called the Copenhagen interpretation and Bohr
was its main proponent.
- The agnostic position: Since we can't know the position of the
particle like behavior until we make the measurement we can just say
that it doesn't matter if the particle exists before we measure. It
doesn't make sense to ask questions which have no answers.
- The many worlds position: The wave function represents a
collection of possibilities. Each of these possibilities comes true in a
different universe. When we make a measurement of a probability and
find that the electron has a definite location (arrives in a lump) we
are merely determining which of the many possible universes we and the
electron are actually in. This position is defended by Hugh Everett and
others.
The Bell inequalities (mentioned below) discovered by John Bell in 1964
suggested experiments which would be able to decided which of these
positions is correct. Those experiment have been done and they strongly
suggest that position 2. is correct (they in fact rule out the
possibility of 1.).
The EPR paradox (AKA entanglement).
IN 1935 Einstein, Podolsky and Rosen presented a paradox which
was meant to show that quantum mechanics had to be incomplete
(essentially the realist position). They suggested a hypothetical
experiment which would show that quantum mechanics would have to be
nonlocal if the orthodox position were true. Nonlocality means that the
wave function must react instantly over large distances. This doesn't
conflict with special relativity because no information can be
transmitted faster than the speed of light but it does suggest that
some quantum states are very strange because they can be spread out
over extremely large regions of space (be nonlocal) and yet remain in
communication so that the wave function can instantly collapse when a
measurement is made of the particle like property. When the experiments
were finally done they did not show quantum mechanics to be incomplete
but rather showed that quantum mechanics IS nonlocal.
Here is a version of the experiment. A neutral pion with zero spin
decays into a positron and an electron each traveling in opposite
directions. To maintain conservation of angular momentum we must have
one decay product (positron or electron) with spin and the other with
the opposite spin (so that the total is zero which is what we started
with). But after the decay we do not know which particle (positron or
electron) has the spin up and which has the spin down (quantum
mechanics doesn't tell us which one gets a spin in any particular
direction). So we must describe the two decay particles with a wave
function with a 50% probability of the electron with a spin and the
positron with opposite spin and 50% chance of the other way round.
However if we measure one of the spins and find out what it is we then
know exactly what the other particle spin is (it has to be opposite of
the what we measured and in fact when we measure the other one it is
always the opposite of the first). Suppose the two particles are
produced and move very far apart. Now when we measure one of them we
still know what the other one's spin is, even if it is very far away.
How did the second particle know what the first one's spin was and how
did it get that information instantly (faster than the speed of light)?
A realist would say that the electrons had a particular spin when they
left but the theory didn't tell us, therefore quantum mechanics is
incomplete. The orthodox position says the wave function is nonlocal
(stretches from one particle to the other, even if they are very far
apart) and the measurement of one particle causes the wave function to
collapse so that the other particle instantly 'knows' what spin to
have. A test of Bell's inequalities show that the outcome of a large
number of measurements is slightly different if there are hidden
variables than if there is not. Experiments have shown that the realist
position is wrong, there cannot be hidden variables.
Bell's inequalities: a simple example borrowed from Chapter 4 in Greene.
Suppose two people (A and B) each get a series of numbered boxes. Each
box has a door on it and when the door opens a light flashes either
blue or red, randomly. A and B notice that whenever A gets a red flash,
B also gets a red flash. This is true of all the boxes they open. The
orthodox quantum view is that the light really is random, a random
decision as to which light will flash is made just before the door on
the box is opened, but because of quantum mechanics (quantum
weirdness), A's box always correlates to the color B gets. Einstein
called this 'Spooky action at a distance' because it seems like the
measurement of A affects the measurement of B, even though they may be
far apart. A realist view is that the boxes are actually programed to
give the same colors in some random order with both boxes having the
same program. The program that causes the correlation is the 'hidden
variable'. This is exactly the same as the pion decay mentioned
above.
Now consider a slight modification. Each box has three doors instead of
one but only one door can be opened at a time (once one is opened the
others remain forever locked). Call the top door 1, the side door 2 and
the front door 3. A and B notice that if A opens door 1 and sees blue,
B will also see blue if she opens door 1. This is true of all the boxes
they open, if they open the same door they get the same color. Again,
the realist view is that there is some kind of program inside each pair
of boxes which tells it what color to flash, depending on which door is
opened. Such a program might be 'blue, blue, red' meaning that if door
1 is opened you get blue, door 2 gives blue and door 3 gives red. The
next pair of boxes might have a different program from blue, blue, red
but each has the same program.
Now suppose A and B decide to open the doors randomly (they don't both
agree to open door 1 but rather randomly choose which door to open).
Notice that there are 9 possible outcomes; A opens door 1 and B opens
door 1, A opens 1 and B opens 2, etc.: (A, B) = (1,1); (1, 2);
(1, 3); (2, 1); (2, 2); (2, 3); (3, 1); (3, 2); or (3, 3). Now
look at the outcomes if there is a program (hidden variable) 'blue,
blue, red'. In this case there are 5 different ways that A and B could
get the same color: (1,1); (2, 2); (3, 3); (1, 2) or (2, 1) and only
four ways to get different colors. The last two cases of getting the
same color occur because there are only two colors but three doors, so
there has to be two ways to get the same color without opening the same
door. Said another way; we know we always get the same color if we open
the same door and there are two more cases where we will get the same
color IF (!!!!) there is a program. Notice that the same argument would
hold for any other program because there are only two colors but three
doors (try 'red, blue, red' for example). So IF (!!!!) there are hidden
variables (a program) we expect to get the same color on average 5 out
of 9 times if A and B randomly open a door on each new pair of boxes.
So IF (!!!!) there is a program telling the lights how to flash we MUST
get a matching color more than 50% of the time on average. If we get
matching colors less than 50% of the time there can be no program, no
hidden variable. If there really is a quantum correlation between the
boxes (we always get the same color if A and B open the same door) then
we expect to get matches 3 out of 9 times but less than 5 out of 9
times. So 33% = or < number of correlations = or < 55%. This is
one version of Bell's inequality.
A and B correspond to the positron and the electron in the above
pion decay. The three doors correspond to the three coordinate axises
(x, y, z). Opening a door corresponds to measuring the spin in the
direction of one of the axises (one you measure the component along a
certain axis you have disturbed the particle so you cannot go back and
measure the y component). Red corresponds to a clockwise spin, blue to
counterclockwise (in the box example the colors correlate, if A
measures blue, so does B; here the spins anti-correlate, if the
positron is clockwise on the x axis the electron is counterclockwise on
the x axis). Bell's inequalities say that if we randomly look at the
spin along different axis we MUST get anti-correlation 55% of the time
(or more) on average IF (!!!!) there is a hidden variable (a program)
that the electron and positron carry with them, telling them which way
to spin). So quantum mechanics says 33% = or < number of
anti-correlations of spin = or < 55% whereas a hidden variable
theory MUST have the number of anti-correlations of spin = or > 55%.
The experiments show a correlation LESS than 55% of the time! There can be NO hidden variable!
Schrodinger's cat.
Once Bell's inequalities were tested and hidden variables ruled out we
have to take the orthodox position seriously. Wave functions evolve as
described by Schrodinger's equation but only give probabilities we
cannot say the electrons have any position or any particular spin. When
we make a measurement however, the probabilities go away and we see the
particle like behavior of the electron. In some sense, then, when we
make a measurement (find the electron as a lump) and the wave function
collapses we somehow cause the physical state we measure. Beforehand it
was just a wave function, afterwards we have a specific value for the
property we measured.
Example: We put a cat in a box with a Geiger counter and a small amount
of radioactive material and a vial of cyanide. Radioactive decay is
random (a quantum mechanical process), put enough in that within an
hour there is a 50% chance one atom has decayed. If one does decay it
is detected by the Geiger counter which (through some mechanism) breaks
the vial of cyanide, killing the cat. Now how do we describe the state
of the whole system? Since quantum mechanical states are involved we
have to have a wave function which shows a 50% probability of having
the cat alive and a 50% chance of the cat being dead after one hour.
And quantum mechanics says we cannot say the cat has any particular
state before we make a measurement. So if we now peek into the box and
find out if the cat is alive or dead, that measurement is what
collapses the wave function and CAUSES the cat to be either alive or
dead. The cat is alive or dead or neither, depending on our measurement
of its state.
Pretty bizarre, huh?