WEBVTT

00:00:00.500 --> 00:00:05.300
PROFESSOR: Let's do a case
that this mostly solvable

00:00:05.300 --> 00:00:08.390
and illustrates
all these things.

00:00:08.390 --> 00:00:10.960
It's a very entertaining case.

00:00:10.960 --> 00:00:14.740
It's called Landau-Zener
transitions.

00:00:14.740 --> 00:00:17.830
For these two people,
Lev Landau, who you've

00:00:17.830 --> 00:00:20.920
probably heard from
Landau and Lifshitz.

00:00:20.920 --> 00:00:25.090
He's the first person
that tried to do this.

00:00:25.090 --> 00:00:29.060
And Zener did it more carefully.

00:00:29.060 --> 00:00:31.480
In fact, apparently
found that Landau

00:00:31.480 --> 00:00:33.230
made a factor of two error.

00:00:36.260 --> 00:00:41.170
And the paper of Zener,
it's actually quite nice,

00:00:41.170 --> 00:00:47.080
and it's a very nice
example that illustrates

00:00:47.080 --> 00:00:49.060
the physics of this transition.

00:00:49.060 --> 00:00:54.070
So we'll devote the
rest of the lecture

00:00:54.070 --> 00:00:56.710
to that Landau-Zener thing.

00:01:04.230 --> 00:01:06.290
OK.

00:01:06.290 --> 00:01:12.590
So it will give us a
little bit into the spirit

00:01:12.590 --> 00:01:18.380
of the adiabatic approximation
in the language that Berry

00:01:18.380 --> 00:01:19.160
used.

00:01:19.160 --> 00:01:26.960
So Landau-Zener transitions.

00:01:36.120 --> 00:01:42.020
OK, Zener and Landau were
interested in molecules,

00:01:42.020 --> 00:01:47.650
and some way of
thinking of molecules

00:01:47.650 --> 00:01:56.150
is to think of nuclei as fixed,
separated by some distance R,

00:01:56.150 --> 00:01:58.590
and then you assume
they are fixed,

00:01:58.590 --> 00:02:01.160
and they're separated
by some distance R.

00:02:01.160 --> 00:02:09.690
And then you calculate what is
the electronic configuration.

00:02:09.690 --> 00:02:16.470
So Zener imagined that you
would have psi 1, one electronic

00:02:16.470 --> 00:02:17.700
configuration.

00:02:20.280 --> 00:02:22.290
It's a wave function
that depends

00:02:22.290 --> 00:02:29.340
on some x's for the electrons,
but it represents the situation

00:02:29.340 --> 00:02:34.530
where the two protons,
say, for a simple molecule,

00:02:34.530 --> 00:02:39.120
maybe they're more distances,
but in particular, they

00:02:39.120 --> 00:02:42.000
are separated by a
distance, R. So that's

00:02:42.000 --> 00:02:45.500
an electronic configuration.

00:02:45.500 --> 00:02:54.900
Electronic configuration,
protons a distance R away.

00:03:01.740 --> 00:03:09.950
And suppose there's another
configuration psi 2 of R.

00:03:09.950 --> 00:03:13.040
It's another configuration,
so two configurations.

00:03:16.450 --> 00:03:17.800
Two different states.

00:03:17.800 --> 00:03:19.810
Maybe in the first
state, the electrons

00:03:19.810 --> 00:03:21.610
are in some ground state.

00:03:21.610 --> 00:03:25.390
In the second state, they're
in some kind of excited state,

00:03:25.390 --> 00:03:27.910
two different configurations.

00:03:27.910 --> 00:03:33.630
Now, we could plot.

00:03:33.630 --> 00:03:38.260
So we'll have [INAUDIBLE]
here, distance R,

00:03:38.260 --> 00:03:40.770
and here's the
cloud of electrons.

00:03:40.770 --> 00:03:46.410
We could plot a
graph as a function

00:03:46.410 --> 00:03:52.940
of the separation, what are
the values of the energies.

00:03:52.940 --> 00:04:01.120
And here is one
possibility for the states.

00:04:01.120 --> 00:04:02.700
And here's another one.

00:04:05.800 --> 00:04:11.350
And that's the plot of the style
that Zener drew in his paper.

00:04:11.350 --> 00:04:23.190
And this represents E1
of R and this E2 of R.

00:04:23.190 --> 00:04:26.850
That is the energy of the
first state, the energy

00:04:26.850 --> 00:04:31.990
of the second state
as a function of R.

00:04:31.990 --> 00:04:34.640
So we are having here--

00:04:37.520 --> 00:04:43.250
oops-- two energy eigenstates.

00:04:43.250 --> 00:04:50.170
So we have H of R. The
Hamiltonian depends on the R.

00:04:50.170 --> 00:04:53.380
And basically you're putting
the two protons, the distance R,

00:04:53.380 --> 00:04:56.990
and calculating the
electrons, how they move.

00:04:56.990 --> 00:05:07.390
Psi i of x R equals
E i of R psi i of xR.

00:05:12.350 --> 00:05:16.310
This is for i equal 1 and 2.

00:05:22.650 --> 00:05:26.520
The case that the people
were interested in

00:05:26.520 --> 00:05:31.500
was the case where this molecule
here, for example, in the state

00:05:31.500 --> 00:05:38.700
2, for this value of R,
there is a critical R 0,

00:05:38.700 --> 00:05:42.090
where things, the
levels get very close.

00:05:42.090 --> 00:05:45.510
For some value of R, this
molecule, for example,

00:05:45.510 --> 00:05:47.520
could be a polar molecule.

00:05:50.220 --> 00:05:54.820
A polar molecule is a
permanent dipole moment.

00:05:54.820 --> 00:05:57.540
It has plus charges
and minus charges,

00:05:57.540 --> 00:05:59.680
not evenly distributed.

00:05:59.680 --> 00:06:01.950
So you get a dipole.

00:06:01.950 --> 00:06:05.820
And maybe here, the
molecule is non-polar.

00:06:11.630 --> 00:06:14.540
And here, it's non-polar.

00:06:20.150 --> 00:06:22.330
Here polar.

00:06:22.330 --> 00:06:25.650
So if you would follow one
of the energy eigenstates,

00:06:25.650 --> 00:06:28.420
there's a critical
value of R, where

00:06:28.420 --> 00:06:31.570
the electronic
configuration is such

00:06:31.570 --> 00:06:34.030
that it goes from
non-polar to polar

00:06:34.030 --> 00:06:36.940
and in the other
energy eigenstate,

00:06:36.940 --> 00:06:38.875
it goes from polar to non-polar.

00:06:44.890 --> 00:06:49.540
So the question is
well, OK, what--

00:06:49.540 --> 00:06:52.390
first of all, what
does all this have

00:06:52.390 --> 00:06:57.570
to do with instantaneous
energy eigenstates and time

00:06:57.570 --> 00:06:59.320
dependence?

00:06:59.320 --> 00:07:01.360
Why are we thinking about this?

00:07:05.350 --> 00:07:09.460
The issue is that sometimes,
you can think of this molecules

00:07:09.460 --> 00:07:14.590
as forming or being subjected to
extra interactions in which you

00:07:14.590 --> 00:07:18.940
will have a process or a
reaction in which the radius

00:07:18.940 --> 00:07:21.610
changes in time.

00:07:21.610 --> 00:07:24.760
So it's possible under
some configuration

00:07:24.760 --> 00:07:30.960
that R becomes R of t.

00:07:34.670 --> 00:07:41.530
And then, this Hamiltonian
is a Hamiltonian

00:07:41.530 --> 00:07:45.200
that depends on R of t.

00:07:45.200 --> 00:07:57.730
This wave functions psi i are
R of t E i's become R of t,

00:07:57.730 --> 00:08:03.530
psi i's become x of R of t.

00:08:03.530 --> 00:08:05.370
This is an important point.

00:08:05.370 --> 00:08:06.920
It's simple, but important.

00:08:06.920 --> 00:08:10.020
The most important points
in physics are simple.

00:08:10.020 --> 00:08:12.050
But you have to
stop and recognize

00:08:12.050 --> 00:08:16.880
that something slightly
new is happening here.

00:08:16.880 --> 00:08:25.670
If you have solved this
equation for all values of R,

00:08:25.670 --> 00:08:28.760
if you know those
energy eigenstates

00:08:28.760 --> 00:08:33.490
for all separations
of the molecule,

00:08:33.490 --> 00:08:39.210
you now have found
instantaneous energy eigenstates

00:08:39.210 --> 00:08:43.559
if it so happens that
R is a function of t,

00:08:43.559 --> 00:08:49.620
because if this is true for any
value of this [INAUDIBLE] R,

00:08:49.620 --> 00:08:53.970
well, then this is
true for all times.

00:08:53.970 --> 00:08:58.320
Because for any specific
time, this is the R,

00:08:58.320 --> 00:09:00.960
the same R is here,
the same R is here,

00:09:00.960 --> 00:09:02.280
and the same R is here.

00:09:02.280 --> 00:09:06.480
And that equation
holds for all R.

00:09:06.480 --> 00:09:13.950
So if this can be solved for
all R, this holds for all times.

00:09:13.950 --> 00:09:18.480
And you have your instantaneous
energy eigenstates.

00:09:18.480 --> 00:09:20.850
You have found those
instantaneous energy

00:09:20.850 --> 00:09:22.500
eigenstates.

00:09:22.500 --> 00:09:26.040
And therefore, the
instantaneous energy eigenstate

00:09:26.040 --> 00:09:29.650
are these ones.

00:09:29.650 --> 00:09:34.090
And the instantaneous
energies are this ones.

00:09:34.090 --> 00:09:38.140
So many times in quantum
mechanics, you do that.

00:09:38.140 --> 00:09:43.210
You solve for the energy
eigenstates for a whole range

00:09:43.210 --> 00:09:46.110
of some parameters.

00:09:46.110 --> 00:09:48.990
And then it so happens
that those parameters

00:09:48.990 --> 00:09:50.590
may change in time.

00:09:50.590 --> 00:09:54.210
But then you have found
the instantaneous energy

00:09:54.210 --> 00:09:58.300
eigenstates for all times.

00:09:58.300 --> 00:10:08.680
So in that picture, we have
the following situation,

00:10:08.680 --> 00:10:12.100
in which the energies
now could be thought

00:10:12.100 --> 00:10:20.960
if R is some
function alpha of t,

00:10:20.960 --> 00:10:27.610
then the same picture would
basically hold true for time

00:10:27.610 --> 00:10:31.420
here and the energies
as a function of time,

00:10:31.420 --> 00:10:38.470
because as time changes, R
changes, and as R changes,

00:10:38.470 --> 00:10:40.990
you already know how
the figure looks.

00:10:40.990 --> 00:10:43.780
So this is a figure
of the energy

00:10:43.780 --> 00:10:45.940
levels as a function of time.

00:10:50.590 --> 00:11:00.220
And now the physical question is
do we get a transition or not?

00:11:00.220 --> 00:11:03.430
So the adiabatic
theorem would say, OK,

00:11:03.430 --> 00:11:07.080
you should state in your
instantaneous energy

00:11:07.080 --> 00:11:11.940
eigenstate, but we're going to
get precisely to this situation

00:11:11.940 --> 00:11:16.920
where these things could
be so small, so little,

00:11:16.920 --> 00:11:21.960
that there is a possibility
of a non adiabatic transition,

00:11:21.960 --> 00:11:27.060
in which you jump to the other
one, because the gap is small.

00:11:27.060 --> 00:11:31.290
So this goes to the real physics
of the adiabatic theorem,

00:11:31.290 --> 00:11:36.030
can we get an estimate or
a calculation that tells us

00:11:36.030 --> 00:11:40.020
how much probability you
have of jumping the gap

00:11:40.020 --> 00:11:42.360
and going to the other branch?

00:11:42.360 --> 00:11:45.570
That's what we're
going to try to do.

00:11:45.570 --> 00:11:51.630
So for that, we'll do
a particular example.

00:11:55.230 --> 00:11:58.590
So let's do that.

00:11:58.590 --> 00:12:03.240
It's an easy one to begin with.

00:12:03.240 --> 00:12:06.100
I'll erase this.

00:12:06.100 --> 00:12:09.795
So baby example, toy example.

00:12:16.060 --> 00:12:22.710
So take a Hamiltonian,
H of t, which

00:12:22.710 --> 00:12:28.410
is going to be of this
form, time dependent one,

00:12:28.410 --> 00:12:32.100
but relatively simple.

00:12:32.100 --> 00:12:36.690
Elements just
along the diagonal.

00:12:36.690 --> 00:12:40.590
OK, that's your Hamiltonian,
two by two matrix,

00:12:40.590 --> 00:12:44.580
elements on the
diagonal, but just

00:12:44.580 --> 00:12:48.340
simple things, the same thing.

00:12:48.340 --> 00:12:53.520
So let's calculate the
instantaneous energy

00:12:53.520 --> 00:12:55.590
eigenstates.

00:12:55.590 --> 00:13:00.150
OK, sounds like a task.

00:13:00.150 --> 00:13:01.810
It's actually pretty simple.

00:13:01.810 --> 00:13:13.500
The instantaneous energy
eigenstates are 1, 0 and 0, 1.

00:13:13.500 --> 00:13:16.440
They don't depend on
time, because essentially

00:13:16.440 --> 00:13:25.320
this Hamiltonian is just alpha
t over 2, 1 minus 1, 0, 0 is--

00:13:25.320 --> 00:13:32.820
it's a constant matrix
times an overall factor.

00:13:32.820 --> 00:13:39.060
The eigenstates of this
matrix are 1, 0 and 0, 1.

00:13:39.060 --> 00:13:44.160
And they are the eigenstates
of this matrix for any time,

00:13:44.160 --> 00:13:46.770
because the time goes in front.

00:13:46.770 --> 00:13:50.460
The matrix doesn't
quite change shape.

00:13:50.460 --> 00:13:54.300
So these are the instantaneous
energy eigenstates.

00:13:54.300 --> 00:13:55.425
They are good forever.

00:13:58.230 --> 00:14:04.020
To plot this, I will assume from
now on that alpha is positive.

00:14:04.020 --> 00:14:10.490
The energy of the
first state is--

00:14:10.490 --> 00:14:13.070
well, what do you get when
you add with the Hamiltonian

00:14:13.070 --> 00:14:14.480
on this state?

00:14:14.480 --> 00:14:22.270
The matrix [INAUDIBLE] and
this is just alpha t over 2.

00:14:22.270 --> 00:14:24.710
And the energy of
the second state

00:14:24.710 --> 00:14:27.410
is going to be minus
alpha t over 2.

00:14:30.430 --> 00:14:35.530
We can plot those
energies, and here

00:14:35.530 --> 00:14:45.490
is the energy of the first state
is alpha t over 2 with alpha

00:14:45.490 --> 00:14:46.600
positive.

00:14:46.600 --> 00:14:51.860
This is like this,
thick output of here.

00:14:51.860 --> 00:14:57.530
This is the state 1, 0
is here, 1, 0 is here,

00:14:57.530 --> 00:15:03.290
alpha t over [? 6. ?] Here
is the energy E1 of t.

00:15:03.290 --> 00:15:06.140
The energy time dependent.

00:15:06.140 --> 00:15:08.560
Here is time.

00:15:08.560 --> 00:15:10.160
And here are energies.

00:15:10.160 --> 00:15:12.200
This is E1.

00:15:12.200 --> 00:15:17.480
And then we have the E2 is the
other one that goes like this.

00:15:17.480 --> 00:15:20.060
It's state 0, 1.

00:15:20.060 --> 00:15:23.750
That's the state 2, 0, 1.

00:15:23.750 --> 00:15:31.310
And the energy is E2 of t,
which is minus alpha t over 2.

00:15:31.310 --> 00:15:35.000
So it's negative for
large positive time

00:15:35.000 --> 00:15:38.730
and positive for the other one.

00:15:38.730 --> 00:15:43.130
So these are your instantaneous
energy eigenstates.

00:15:43.130 --> 00:15:49.200
OK, and this is not quite
what we wanted here.

00:15:49.200 --> 00:15:53.080
We wanted things to
avoid themselves.

00:15:53.080 --> 00:15:57.720
But this is going to
illustrate an important effect.

00:15:57.720 --> 00:16:04.380
I claim, actually, that the true
solutions of the Schrodinger

00:16:04.380 --> 00:16:09.360
equation are in
this case dressed up

00:16:09.360 --> 00:16:14.730
versions of the instantaneous
energy eigenstates.

00:16:14.730 --> 00:16:17.160
So what I claim is
the kind of-- you

00:16:17.160 --> 00:16:20.230
do the adiabatic state
corresponding to this,

00:16:20.230 --> 00:16:22.860
the adiabatic state
corresponding to that,

00:16:22.860 --> 00:16:25.870
and those are exact solutions.

00:16:25.870 --> 00:16:30.960
So there is no coupling
between the states, 1 and 2.

00:16:30.960 --> 00:16:32.780
So this is plausible.

00:16:32.780 --> 00:16:36.990
So let's write those solutions.

00:16:36.990 --> 00:16:42.600
I claim here is
psi 1 of t I claim

00:16:42.600 --> 00:16:47.670
is the exponential
of minus i over h bar

00:16:47.670 --> 00:16:57.640
integral up to t of E1
of t prime dt prime times

00:16:57.640 --> 00:16:58.675
the state 1.

00:17:01.690 --> 00:17:04.210
I claim this solves the
Schrodinger equation.

00:17:04.210 --> 00:17:13.150
i h bar dt of this psi
should be equal to H psi.

00:17:13.150 --> 00:17:14.980
Is it clear?

00:17:14.980 --> 00:17:16.930
Yes, I think it's clear.

00:17:16.930 --> 00:17:18.700
It solves it,
because if you take

00:17:18.700 --> 00:17:29.330
the time derivative of this
thing, it multiplies by E1.

00:17:29.330 --> 00:17:34.330
The i h bar cancels that
factor of minus i over h bar.

00:17:34.330 --> 00:17:38.860
The time derivatives
brings out an E1 of t.

00:17:38.860 --> 00:17:44.380
But this state, despite the
phase when h [INAUDIBLE],,

00:17:44.380 --> 00:17:47.740
it goes through the
phase, hits the state 1,

00:17:47.740 --> 00:17:51.650
and produces the E1 energy.

00:17:51.650 --> 00:17:56.140
So this is solved
by that equation,

00:17:56.140 --> 00:17:58.300
and you can do the integral.

00:17:58.300 --> 00:17:59.530
It looks OK.

00:17:59.530 --> 00:18:09.700
It's exponential of minus i
alpha t squared over 4 h bar 1,

00:18:09.700 --> 00:18:21.940
and the state psi 2 of t is the
same exponential with E2 with 2

00:18:21.940 --> 00:18:30.620
here, and it's the
exponential of plus i alpha

00:18:30.620 --> 00:18:34.700
t squared over 4 h bar 2.

00:18:37.890 --> 00:18:39.500
OK.

00:18:39.500 --> 00:18:41.690
Let's appreciate
the lesson again.

00:18:41.690 --> 00:18:48.870
We got a very simple system,
two levels, crossing--

00:18:48.870 --> 00:18:50.180
they cross.

00:18:50.180 --> 00:18:52.415
The energy levels cross.

00:18:52.415 --> 00:18:54.450
That generally doesn't happen.

00:18:54.450 --> 00:18:58.010
You have to have a very special
Hamiltonian for the energy

00:18:58.010 --> 00:18:59.570
levels to cross.

00:18:59.570 --> 00:19:03.770
We found the instantaneous
energy eigenstate,

00:19:03.770 --> 00:19:10.320
and we found two exact solutions
of this Schrodinger equation,

00:19:10.320 --> 00:19:14.190
two perfect complete exact
solutions of the Schrodinger

00:19:14.190 --> 00:19:17.970
equation that represent
the system doing just zoom,

00:19:17.970 --> 00:19:22.050
like that, or doing like that.

00:19:22.050 --> 00:19:26.740
Totally oblivious that there's
a state they're crossing,

00:19:26.740 --> 00:19:31.100
the Schrodinger equation doesn't
couple them in this case.