WEBVTT

00:00:01.610 --> 00:00:05.480
PROFESSOR: So I
want to demystify

00:00:05.480 --> 00:00:06.830
a little of this equation.

00:00:06.830 --> 00:00:10.880
We sometimes use
the basis to show

00:00:10.880 --> 00:00:12.510
everything that is happening.

00:00:12.510 --> 00:00:15.000
So we have a good basis.

00:00:15.000 --> 00:00:17.100
So let's look at it.

00:00:17.100 --> 00:00:33.700
So how does it all look
in an orthonormal basis?

00:00:33.700 --> 00:00:34.740
So which basis?

00:00:34.740 --> 00:00:36.080
Well, you have one.

00:00:36.080 --> 00:00:39.100
H 0 was supposed to be known.

00:00:39.100 --> 00:00:43.730
So we'll call these states n.

00:00:43.730 --> 00:00:47.880
These are the
eigenstates of H 0.

00:00:47.880 --> 00:00:49.560
We'll have E n.

00:00:49.560 --> 00:00:50.880
We could put a zero.

00:00:50.880 --> 00:00:53.610
We used to have that zero there.

00:00:53.610 --> 00:00:57.420
But since we're
never going to try

00:00:57.420 --> 00:01:02.820
to solve this equation in a
different way, as time goes on,

00:01:02.820 --> 00:01:05.129
the time-dependent
system, I said to you,

00:01:05.129 --> 00:01:07.230
doesn't have energy eigenstates.

00:01:07.230 --> 00:01:09.780
So perturbation
theory is not going

00:01:09.780 --> 00:01:12.030
to be about perturbing
these energies.

00:01:12.030 --> 00:01:16.050
So I will erase.

00:01:16.050 --> 00:01:18.832
E n, and there will
be no confusion.

00:01:18.832 --> 00:01:20.040
That's never going to change.

00:01:22.670 --> 00:01:24.830
Now I'm going to
write an equation that

00:01:24.830 --> 00:01:28.340
may seem a little strange.

00:01:28.340 --> 00:01:30.790
We want to solve for this psi ~.

00:01:30.790 --> 00:01:38.360
So let's write an
ansatz for psi ~ of t.

00:01:38.360 --> 00:01:51.345
It's going to be the
following, sum over n C n n.

00:01:57.260 --> 00:02:03.650
If C ns are constants, this
is definitely not right.

00:02:03.650 --> 00:02:06.110
This state must have
some time-dependence.

00:02:06.110 --> 00:02:11.030
So at least, I should put
a time-dependence here.

00:02:11.030 --> 00:02:18.500
But even that sounds a
little wrong, a priori.

00:02:18.500 --> 00:02:22.200
You knew that in
time-independent problems,

00:02:22.200 --> 00:02:26.250
you can always write a state as
a superposition of your energy

00:02:26.250 --> 00:02:27.340
eigenstate.

00:02:27.340 --> 00:02:32.240
So now, this is the
interacting picture

00:02:32.240 --> 00:02:40.280
we have not any more energy
eigenstates in any sense.

00:02:40.280 --> 00:02:42.900
Can I write this?

00:02:42.900 --> 00:02:44.540
Does this make sense still?

00:02:47.830 --> 00:02:51.320
Well, the answer
is yes, you can.

00:02:51.320 --> 00:02:53.980
And the reason is
that whatever you

00:02:53.980 --> 00:02:57.820
can say about this
energy eigenstate,

00:02:57.820 --> 00:02:59.650
they apply for early times.

00:02:59.650 --> 00:03:01.240
They apply for late times.

00:03:01.240 --> 00:03:03.170
They anyway exist.

00:03:03.170 --> 00:03:04.990
They're a basis.

00:03:04.990 --> 00:03:08.220
And what I said,
they are a basis.

00:03:08.220 --> 00:03:11.980
So any wave function can be
written in terms of them.

00:03:11.980 --> 00:03:16.780
So if I look psi ~
at time 1 second,

00:03:16.780 --> 00:03:21.290
I should be able to find
numbers that make this possible.

00:03:21.290 --> 00:03:23.500
And if I look at
it at 2 seconds,

00:03:23.500 --> 00:03:26.810
I will find another set of
numbers and make it possible.

00:03:26.810 --> 00:03:31.780
So by the fact that I've
included here a time-dependence

00:03:31.780 --> 00:03:36.850
for the coefficients, it
is possible to write this.

00:03:36.850 --> 00:03:41.780
These coefficients change
in time in strange ways,

00:03:41.780 --> 00:03:44.680
but that's our unknown.

00:03:44.680 --> 00:03:47.590
If I knew how they
change in time,

00:03:47.590 --> 00:03:49.990
I would have solved the problem.

00:03:49.990 --> 00:03:52.450
I don't know how
the change in time.

00:03:52.450 --> 00:03:59.240
But in general, we will try
to find for this coefficient.

00:03:59.240 --> 00:04:02.950
So if you have psi ~, you
know at the end of the day,

00:04:02.950 --> 00:04:05.530
your goal is psi of t.

00:04:05.530 --> 00:04:07.490
So what is psi of t?

00:04:13.520 --> 00:04:17.940
Psi of t, from this
blackboard, is the action

00:04:17.940 --> 00:04:21.120
of this operator on psi ~.

00:04:21.120 --> 00:04:25.035
That operator moves through the
constants and through the sum.

00:04:29.210 --> 00:04:34.190
And with the end states
eigenstates of H 0,

00:04:34.190 --> 00:04:40.620
this just gives me minus
I E n t over H bar n.

00:04:43.970 --> 00:04:47.180
And this expansion should
make you feel good.

00:04:50.000 --> 00:04:52.460
You say, OK, here it is.

00:04:52.460 --> 00:05:00.350
These are my states in my
time-dependent we function,

00:05:00.350 --> 00:05:01.400
let's say.

00:05:01.400 --> 00:05:04.010
And it's given by this formula.

00:05:04.010 --> 00:05:09.980
If I didn't have a delta H, we
know that without a delta H,

00:05:09.980 --> 00:05:12.710
psi ~ is constant.

00:05:12.710 --> 00:05:17.210
So they C ns would be constant
if there is no delta H.

00:05:17.210 --> 00:05:19.580
And if the C ns are
constant and there's

00:05:19.580 --> 00:05:24.050
no delta H, that's exactly
how energy eigenstates evolve

00:05:24.050 --> 00:05:26.000
in time.

00:05:26.000 --> 00:05:29.000
They are evolving
with exponentials

00:05:29.000 --> 00:05:32.510
of the action of each 0.

00:05:32.510 --> 00:05:35.420
So this is consistent
with all you know.

00:05:35.420 --> 00:05:39.055
If delta H turns on,
the C ns are going

00:05:39.055 --> 00:05:40.730
to acquire time-dependence.

00:05:40.730 --> 00:05:45.440
And we will know what
the state is doing.

00:05:45.440 --> 00:05:47.910
So this is good.

00:05:47.910 --> 00:05:51.470
So let's plug, into
the new Schrodinger

00:05:51.470 --> 00:05:53.360
equation, this expansion.

00:05:56.410 --> 00:05:59.250
And this is our new
Schrodinger equation.

00:05:59.250 --> 00:06:02.630
So what do we get?

00:06:02.630 --> 00:06:10.290
I get I H bar d/dt of this.

00:06:10.290 --> 00:06:13.610
I'll write it here with a sum.

00:06:13.610 --> 00:06:15.560
I'll change to letter m.

00:06:19.128 --> 00:06:21.725
I'll use dots.

00:06:21.725 --> 00:06:23.990
C m of t dot.

00:06:23.990 --> 00:06:26.750
For time derivative,
we'll many time use dots.

00:06:26.750 --> 00:06:30.350
So I'm taking the time
derivative of psi ~

00:06:30.350 --> 00:06:33.230
as in the Schrodinger equation.

00:06:33.230 --> 00:06:34.668
I go here, C m m.

00:06:37.360 --> 00:06:40.930
And this is supposed
to be equal to delta H

00:06:40.930 --> 00:06:44.780
~ times this same state.

00:06:44.780 --> 00:06:56.940
So it's going to be sum
over n C n of t delta H ~ n.

00:06:59.580 --> 00:07:03.780
So this is the Schrodinger
equation for this state.

00:07:03.780 --> 00:07:06.090
We're now looking at
the Schrodinger equation

00:07:06.090 --> 00:07:09.750
in a basis, because that may
be a good way to solve it.

00:07:15.740 --> 00:07:20.635
Well, one way to
solve it is to do,

00:07:20.635 --> 00:07:24.080
in the right-hand
side, a complete set,

00:07:24.080 --> 00:07:26.540
introduce a complete
set of states.

00:07:26.540 --> 00:07:36.280
So I'll put a sum over
m m m and this whole sum

00:07:36.280 --> 00:07:45.260
over n C n of t delta H n.

00:07:45.260 --> 00:07:50.060
So this will be
equal to sum over m.

00:07:50.060 --> 00:07:52.850
The m is here.

00:07:52.850 --> 00:07:59.090
And sum over n, the bra
can go all the way in.

00:07:59.090 --> 00:08:07.520
And you write C n
of t m delta H n.

00:08:15.180 --> 00:08:18.420
So this is our Schrodinger
equation still.

00:08:18.420 --> 00:08:23.790
And now it's kind of in a
nice way in which we have

00:08:23.790 --> 00:08:28.815
similar letters on both sides.

00:08:35.220 --> 00:08:36.496
A little bit of notation.

00:08:42.010 --> 00:08:51.280
I'll call this delta H ~ m n.

00:08:51.280 --> 00:08:57.740
We've done that in perturbation
theory many, many times.

00:08:57.740 --> 00:09:01.850
So what is our equation?

00:09:01.850 --> 00:09:07.900
Well, compare terms with
equal value of the function

00:09:07.900 --> 00:09:10.930
in front of the state m.

00:09:10.930 --> 00:09:18.100
So we gave I H bar
C m dot of t is

00:09:18.100 --> 00:09:33.450
equal to the sum over n C
n or delta H m n C n of t.

00:09:40.000 --> 00:09:42.850
That's it.

00:09:42.850 --> 00:09:45.830
It's a nice looking equation.

00:09:45.830 --> 00:09:49.570
There's a couple sets of
differential equations

00:09:49.570 --> 00:09:56.320
for an infinite set of functions
in which the derivatives are

00:09:56.320 --> 00:10:00.310
obtained in terms of the
Hamiltonian matrix elements

00:10:00.310 --> 00:10:04.150
for the transition Hamiltonian
for the perturbation

00:10:04.150 --> 00:10:05.560
times these functions there.

00:10:09.340 --> 00:10:13.840
A little more notation
here in the sense

00:10:13.840 --> 00:10:19.926
of just understanding the
structure of that matrix

00:10:19.926 --> 00:10:20.425
element.

00:10:23.200 --> 00:10:27.640
That's useful, because in
practice, all the tilde things,

00:10:27.640 --> 00:10:31.840
at the end of the day,
we don't want them.

00:10:31.840 --> 00:10:38.050
We want the original ones,
things without tilde.

00:10:38.050 --> 00:10:40.190
And we always look for them.

00:10:40.190 --> 00:10:55.120
So what is this matrix element
delta H m n ~ is m delta H ~ n.

00:10:55.120 --> 00:10:59.530
But remember what delta H ~ was.

00:10:59.530 --> 00:11:02.630
You have it here.

00:11:02.630 --> 00:11:12.805
So we have m E to the I
H 0 t over H bar delta H

00:11:12.805 --> 00:11:19.630
E to the minus I H
0 t over H bar n.

00:11:19.630 --> 00:11:21.440
That's no problem.

00:11:21.440 --> 00:11:25.060
Everything is good here,
because those are eigenstates.

00:11:25.060 --> 00:11:30.040
So we know how much
you get by letting

00:11:30.040 --> 00:11:34.720
this act on that state and
that other exponential act

00:11:34.720 --> 00:11:38.030
on this state on the right.

00:11:38.030 --> 00:11:39.340
So what do we get?

00:11:39.340 --> 00:11:46.370
E to the I e m t over H bar.

00:11:46.370 --> 00:11:54.570
And from the other one, E
n t over H bar like this.

00:11:54.570 --> 00:11:57.280
The two exponential
give you that.

00:11:57.280 --> 00:12:02.350
And then you have just delta
H, your original perturbation,

00:12:02.350 --> 00:12:04.930
between states of
the Hamiltonian.

00:12:12.570 --> 00:12:14.040
I'll write it this way.

00:12:14.040 --> 00:12:23.680
Delta H mn, just without the
tilde, which is m delta H n.

00:12:30.760 --> 00:12:38.710
Most people call this
the frequency m n.

00:12:38.710 --> 00:12:42.560
E over H bar is a frequency.

00:12:42.560 --> 00:12:46.870
So the harmonic
oscillator reminds you

00:12:46.870 --> 00:12:50.680
of that, E equal H bar omega.

00:12:50.680 --> 00:12:57.150
And for easier writing,
you write an omega mn.

00:12:57.150 --> 00:13:08.250
So this becomes E to the I
omega mn t times delta H mn.

00:13:08.250 --> 00:13:11.265
So this is kind of
a nice notation.

00:13:13.940 --> 00:13:17.810
So this matrix elements there--

00:13:17.810 --> 00:13:23.950
well, I'll write again the
equation I H bar C m dot of t

00:13:23.950 --> 00:13:37.472
would be the sum over n of E to
the I omega mn t delta H mn C

00:13:37.472 --> 00:13:38.770
n of t.

00:13:45.500 --> 00:13:47.900
That same equation
has been rewritten

00:13:47.900 --> 00:13:55.070
in the viewpoint of this,
where we simplified the tilde.

00:13:55.070 --> 00:13:58.010
And we can refer everything
to our original basis.

00:14:00.820 --> 00:14:06.060
So in a sense, I think this
should demystify things.

00:14:06.060 --> 00:14:08.850
What's the situation?

00:14:08.850 --> 00:14:13.980
What has happened is
that if you have a basis,

00:14:13.980 --> 00:14:20.130
you write an ansatz for the
wave function of this form.

00:14:20.130 --> 00:14:23.820
And the coefficients are
solutions of those differential

00:14:23.820 --> 00:14:24.700
equations.

00:14:24.700 --> 00:14:32.850
So we've translated the
problem to something doable.

00:14:32.850 --> 00:14:37.210
If you have lots of
resources, a computer,

00:14:37.210 --> 00:14:40.830
and you solve coupled
time-dependent differential

00:14:40.830 --> 00:14:42.250
equations first order.

00:14:42.250 --> 00:14:44.970
There this are not
partricularly--

00:14:44.970 --> 00:14:48.210
well, it all depends
how difficult

00:14:48.210 --> 00:14:49.680
is the time-dependence here.

00:14:49.680 --> 00:14:52.290
The exponential is not bad.

00:14:52.290 --> 00:14:53.610
This?

00:14:53.610 --> 00:14:55.630
Well, it all depends.

00:14:55.630 --> 00:14:57.300
But this can be
solved numerically.

00:14:57.300 --> 00:14:59.610
It can be solved
with many methods.

00:14:59.610 --> 00:15:02.430
This has made the
problem concrete.

00:15:02.430 --> 00:15:04.920
And we're going to try to
understand how to solve it

00:15:04.920 --> 00:15:07.280
in cases of interest.