WEBVTT
00:00:00.500 --> 00:00:02.370
PROFESSOR: Today,
we're going to continue
00:00:02.370 --> 00:00:04.500
with the adiabatic subject.
00:00:04.500 --> 00:00:09.930
And our main topic is
going to be Berry's Phase.
00:00:09.930 --> 00:00:12.960
It's interesting
part of the phase
00:00:12.960 --> 00:00:15.060
that goes in adiabatic process.
00:00:15.060 --> 00:00:18.150
And we want to
understand what it is
00:00:18.150 --> 00:00:21.550
and why people care about it.
00:00:21.550 --> 00:00:25.140
And then, we'll turn
to another subject
00:00:25.140 --> 00:00:29.620
in which the adiabatic
approximation is of interest.
00:00:29.620 --> 00:00:31.920
And it's a subject of molecules.
00:00:31.920 --> 00:00:35.250
So I don't think I'll manage to
get through all of that today,
00:00:35.250 --> 00:00:37.690
but will we'll make an effort.
00:00:37.690 --> 00:00:44.560
So let me remind you
of what we had so far.
00:00:44.560 --> 00:00:48.150
So we imagine we have
a Hamiltonian that
00:00:48.150 --> 00:00:52.590
depends on time and
maybe had no dependents
00:00:52.590 --> 00:00:55.450
before time equals 0 turns on.
00:00:55.450 --> 00:01:02.160
And it has no further
variation after some time t.
00:01:02.160 --> 00:01:06.760
So the Hamiltonian
changes like that.
00:01:06.760 --> 00:01:13.040
And the adiabatic theorem
states that if you
00:01:13.040 --> 00:01:17.630
have a state at
time equals 0, which
00:01:17.630 --> 00:01:25.760
is a particular instantaneous
eigenstate, that
00:01:25.760 --> 00:01:28.610
is the instantaneous
eigenstate, then
00:01:28.610 --> 00:01:33.200
that's the full wave function at
time equals 0 and it coincides.
00:01:33.200 --> 00:01:38.840
Then, at time at any time in
this process, if it's slow,
00:01:38.840 --> 00:01:42.770
the process, the
state of the system,
00:01:42.770 --> 00:01:48.110
the full wave
function, psi of t,
00:01:48.110 --> 00:01:55.440
will tend to remain in that
instantaneous eigenstate.
00:01:55.440 --> 00:01:57.970
And the way it's
stated precisely
00:01:57.970 --> 00:02:01.480
is that psi of t minus--
00:02:01.480 --> 00:02:04.040
I'll write it like this--
00:02:04.040 --> 00:02:15.170
psi prime n of t, the
norm of this state
00:02:15.170 --> 00:02:25.860
is of order 1/T for any
t in between 0 and T.
00:02:25.860 --> 00:02:31.310
So I'm trying to state the
adiabatic theorem in a way that
00:02:31.310 --> 00:02:33.200
is mathematically precise.
00:02:33.200 --> 00:02:37.280
And let me remind you the
norm of a wave function
00:02:37.280 --> 00:02:41.180
is you integrate the
wave function square
00:02:41.180 --> 00:02:43.010
and take the square root.
00:02:43.010 --> 00:02:46.130
It's sort of the
usual definition
00:02:46.130 --> 00:02:52.430
of the norm of a vector is the
inner product of the vector
00:02:52.430 --> 00:02:54.480
with itself square root.
00:02:54.480 --> 00:02:57.390
So that's the norm
for wave function.
00:02:57.390 --> 00:03:03.560
And here, what this means is
that with some suitable choice
00:03:03.560 --> 00:03:08.510
of phase, the
instantaneous eigenstate
00:03:08.510 --> 00:03:11.600
is very close to the true state.
00:03:11.600 --> 00:03:18.320
And the error is a Fourier
1/T. So if the process is slow,
00:03:18.320 --> 00:03:22.010
it means that the change
occurs over long t,
00:03:22.010 --> 00:03:23.870
this is a small number.
00:03:23.870 --> 00:03:29.370
And there is some
instantaneous eigenstate
00:03:29.370 --> 00:03:31.470
with some peculiar phase--
00:03:31.470 --> 00:03:33.240
that's why I put the prime--
00:03:33.240 --> 00:03:36.720
for which this
difference is very small.
00:03:36.720 --> 00:03:40.480
And we calculated
this phase, and we
00:03:40.480 --> 00:03:47.190
found the state, psi
of t is roughly equal
00:03:47.190 --> 00:03:56.640
to e to theta n of t, e to the
I gamma n of t, psi n of t.
00:03:59.350 --> 00:04:03.310
And in this statement,
this is what I would
00:04:03.310 --> 00:04:08.310
call the psi n prime of t.
00:04:08.310 --> 00:04:11.950
And that's why the real
state is just approximately
00:04:11.950 --> 00:04:14.590
equal to that one.
00:04:14.590 --> 00:04:24.550
And we have these phases in
which theta of t is minus 1
00:04:24.550 --> 00:04:33.250
over h bar integral from 0
to t E n of t prime dt prime.
00:04:33.250 --> 00:04:35.800
That is kind of
a familiar phase.
00:04:35.800 --> 00:04:41.440
If you had a normal energy
eigenstate, time independent 1,
00:04:41.440 --> 00:04:44.170
this would be 1 to the--
00:04:44.170 --> 00:04:49.570
well, would be minus e
times t over h bar with an I
00:04:49.570 --> 00:04:51.370
would be the familiar
phase that you
00:04:51.370 --> 00:04:53.620
put to an energy eigenstate.
00:04:53.620 --> 00:04:58.830
Then it comes the
gamma n of t, which
00:04:58.830 --> 00:05:07.330
is an integral from 0 to t of
some new n of t prime dt prime.
00:05:07.330 --> 00:05:19.848
And this new n of t is I
psi n of t psi n dot of t.
00:05:23.264 --> 00:05:25.830
I think I have it right.
00:05:25.830 --> 00:05:30.270
So the second part of
the phase is the integral
00:05:30.270 --> 00:05:33.570
of this new function.
00:05:33.570 --> 00:05:39.240
And this new function is
real, because this part
00:05:39.240 --> 00:05:44.910
we showed before is imaginary,
where with an I, this is real.
00:05:44.910 --> 00:05:48.420
And this second
part, this gamma n,
00:05:48.420 --> 00:05:51.390
is called the geometric phase.
00:05:51.390 --> 00:05:56.080
This is the phase that has
to do with Berry's phase.
00:05:56.080 --> 00:06:00.120
And it's a phase that
we want to understand.
00:06:00.120 --> 00:06:03.120
And it's geometrical
because of one reason
00:06:03.120 --> 00:06:04.950
that we're going
to show that makes
00:06:04.950 --> 00:06:12.600
it quite surprising and quite
different from the phase theta.
00:06:12.600 --> 00:06:17.620
The phase theta is a
little like a clock,
00:06:17.620 --> 00:06:20.230
because it runs with time.
00:06:20.230 --> 00:06:22.690
The more time you
wait on an energy
00:06:22.690 --> 00:06:25.975
eigenstate, the more
this phase changes.
00:06:28.500 --> 00:06:34.590
What will happen with
this geometric phase
00:06:34.590 --> 00:06:40.580
is that somehow
properly viewed is
00:06:40.580 --> 00:06:48.030
independent of the time it takes
the adiabatic process to occur.
00:06:48.030 --> 00:06:52.460
So whether it takes us
small time or a long time
00:06:52.460 --> 00:06:55.430
to produce this
change of the system,
00:06:55.430 --> 00:07:00.410
the geometric phase will
be essentially the same.
00:07:00.410 --> 00:07:04.140
That's very, very unusual.
00:07:04.140 --> 00:07:06.530
So that's the main thing
we want to understand
00:07:06.530 --> 00:07:10.700
about this geometric
phase, that it depends only
00:07:10.700 --> 00:07:15.960
on the evolution of the state
in that configuration space--
00:07:15.960 --> 00:07:19.430
we'll make that
clear, what it means--
00:07:19.430 --> 00:07:25.340
and not the time it takes
this evolution to occur.
00:07:25.340 --> 00:07:29.460
It's a little more subtle, this
phase, than the other phase.
00:07:29.460 --> 00:07:34.920
So I want to introduce this
idea of a configuration space.
00:07:34.920 --> 00:07:39.040
So basically, we have that--
00:07:39.040 --> 00:07:42.250
let me forget about time
dependence for one second
00:07:42.250 --> 00:07:48.330
and think of the
Hamiltonian as a function
00:07:48.330 --> 00:07:52.500
of a set of coordinates,
or parameters.
00:07:52.500 --> 00:07:58.650
So the Rs are some coordinates.
00:07:58.650 --> 00:08:05.340
R1, R2, maybe up to R capital
N are some coordinates
00:08:05.340 --> 00:08:09.450
inside some vector space RN.
00:08:12.370 --> 00:08:13.900
So its N components.
00:08:13.900 --> 00:08:15.260
And what does that mean?
00:08:15.260 --> 00:08:17.620
It means maybe that
your Hamiltonian
00:08:17.620 --> 00:08:21.180
has capital N parameters.
00:08:21.180 --> 00:08:23.180
And those are these things.
00:08:23.180 --> 00:08:26.260
So you buy this Hamiltonian.
00:08:26.260 --> 00:08:27.880
It comes with some parameters.
00:08:27.880 --> 00:08:28.750
You buy another one.
00:08:28.750 --> 00:08:31.030
It comes with another
set of parameters.
00:08:31.030 --> 00:08:32.799
Those parameters can be changed.
00:08:32.799 --> 00:08:36.370
Or you construct them in
the lab, your Hamiltonians
00:08:36.370 --> 00:08:38.020
with different parameters.
00:08:38.020 --> 00:08:40.870
Those are the parameters
of the Hamiltonian.
00:08:51.840 --> 00:08:54.690
And suppose you have
learned to solve
00:08:54.690 --> 00:08:58.740
this Hamiltonian for all
values of the parameters.
00:08:58.740 --> 00:09:03.270
That is whatever the
Rs are you know how
00:09:03.270 --> 00:09:06.030
to find the energy eigenstates.
00:09:06.030 --> 00:09:11.820
So H of R times--
00:09:11.820 --> 00:09:24.390
there are some eigenstates, psi
n of R with energies En of R,
00:09:24.390 --> 00:09:33.500
psi n of R. And n
maybe is 1, 2, 3.
00:09:33.500 --> 00:09:41.310
And these are
orthonormal states,
00:09:41.310 --> 00:09:43.615
those energy eigenstates.
00:09:46.250 --> 00:09:49.160
So this equation
says that you have
00:09:49.160 --> 00:09:52.880
been able to solve this
Hamiltonian whatever
00:09:52.880 --> 00:09:55.130
the values of the
parameters are.
00:09:55.130 --> 00:09:58.140
And you have found all
the states of the system,
00:09:58.140 --> 00:10:00.950
n equal 1, 2 3, 4, 5, 6.
00:10:00.950 --> 00:10:03.980
All of them are in now.
00:10:03.980 --> 00:10:08.180
So this is a general situation.
00:10:08.180 --> 00:10:12.180
And now, we imagine
that for some reason,
00:10:12.180 --> 00:10:17.940
these parameters start to
begin to depend on time.
00:10:17.940 --> 00:10:22.230
So they become time
dependent parameters--
00:10:26.330 --> 00:10:33.030
can become time dependent.
00:10:33.030 --> 00:10:36.650
So that you now
have R of t vector.
00:10:36.650 --> 00:10:42.650
These are 1 of t
up to our Rn of t.
00:10:48.040 --> 00:10:50.155
So how do we represent this?
00:10:58.350 --> 00:11:03.750
Well, this is a Cartesian
space of parameters.
00:11:03.750 --> 00:11:06.540
This is not our normal space.
00:11:06.540 --> 00:11:10.110
This is a space where one axis
could be the magnetic field.
00:11:10.110 --> 00:11:12.690
Another axis could be
the electric field.
00:11:12.690 --> 00:11:16.980
Another axis could be
the spring constant.
00:11:16.980 --> 00:11:20.880
Those are abstract axis
of configuration space.
00:11:20.880 --> 00:11:29.190
Or this could be R1,
R2, the axis, R3.
00:11:29.190 --> 00:11:32.200
And those are your axes.
00:11:32.200 --> 00:11:39.770
And now, how do you represent
in this configuration space
00:11:39.770 --> 00:11:43.200
the evolution of the system?
00:11:43.200 --> 00:11:46.790
What is the evolution of the
system in this configuration
00:11:46.790 --> 00:11:47.290
space?
00:11:50.960 --> 00:11:51.760
How does it look?
00:11:51.760 --> 00:11:53.240
Is it a point?
00:11:53.240 --> 00:11:54.020
A line?
00:11:54.020 --> 00:11:55.603
A surface?
00:11:55.603 --> 00:11:56.790
What is it?
00:12:02.646 --> 00:12:03.622
Sorry?
00:12:03.622 --> 00:12:05.100
STUDENT: A path.
00:12:05.100 --> 00:12:06.180
PROFESSOR: It's a path.
00:12:06.180 --> 00:12:07.190
It's a line.
00:12:07.190 --> 00:12:10.130
Indeed, you look at your clock.
00:12:10.130 --> 00:12:13.410
And at time equals 0,
well, it takes some values.
00:12:13.410 --> 00:12:17.490
And you're fine, OK,
here it at time equals 0.
00:12:17.490 --> 00:12:20.760
At time equal 1,
the values change.
00:12:20.760 --> 00:12:23.130
There's one parameter,
which is time.
00:12:23.130 --> 00:12:26.760
So this traces a path.
00:12:26.760 --> 00:12:31.860
As time goes by, the core in
this changing in time and this
00:12:31.860 --> 00:12:37.090
is a line
parameterized by time--
00:12:37.090 --> 00:12:51.060
so a path gamma
parameterized by time.
00:12:51.060 --> 00:12:56.250
And that represents the
evolution of your system.
00:12:56.250 --> 00:13:02.370
At time equals 0, this point
could be R at t equals 0.
00:13:02.370 --> 00:13:09.210
And maybe this point
is R a t equal T final.
00:13:09.210 --> 00:13:12.510
And this system is
going like that.
00:13:12.510 --> 00:13:17.130
You should imagine the system as
traveling in that configuration
00:13:17.130 --> 00:13:18.330
space.
00:13:18.330 --> 00:13:19.510
That's what it does.
00:13:19.510 --> 00:13:22.780
That's why we put the
configuration space.
00:13:22.780 --> 00:13:26.040
And we now have a set--
00:13:26.040 --> 00:13:29.190
not the set-- a time
dependent Hamiltonian,
00:13:29.190 --> 00:13:35.640
because while H was a function
of R from the beginning, now
00:13:35.640 --> 00:13:37.750
R is a function of time.
00:13:37.750 --> 00:13:41.400
So this is your new Hamiltonian.
00:13:41.400 --> 00:13:45.010
And this is time dependent--
00:13:45.010 --> 00:13:46.920
dependent Hamiltonian.
00:13:51.450 --> 00:13:54.450
But now, the
interesting thing is
00:13:54.450 --> 00:13:59.610
that the work you did
before in finding the energy
00:13:59.610 --> 00:14:05.820
eigenstates for any position
in this configuration space
00:14:05.820 --> 00:14:10.920
is giving you the instantaneous
energy eigenstates,
00:14:10.920 --> 00:14:19.740
because if this equation here
holds for any value of R,
00:14:19.740 --> 00:14:24.090
it certainly holds for the
values of R corresponding
00:14:24.090 --> 00:14:25.870
to some particular time.
00:14:28.830 --> 00:14:43.770
So psi n of R of t is equal to
En of R of t psi n of R of t.
00:14:43.770 --> 00:14:47.780
So it's an interesting
interplay in which
00:14:47.780 --> 00:14:52.890
the act that you know your
energy eigenstates everywhere
00:14:52.890 --> 00:14:55.980
in your configuration
space allows
00:14:55.980 --> 00:15:01.980
you to find the
time evolved states,
00:15:01.980 --> 00:15:07.800
the time dependent energy
eigenstates, the instantaneous
00:15:07.800 --> 00:15:12.060
energy eigenstates
are found here.
00:15:12.060 --> 00:15:17.780
So what we want to do now
is evaluate in this language
00:15:17.780 --> 00:15:22.830
the geometric phase, this phase.
00:15:22.830 --> 00:15:25.380
I want to understand
what this phase is
00:15:25.380 --> 00:15:28.130
in this geometric language.