WEBVTT

00:00:00.890 --> 00:00:05.120
PROFESSOR: Our subject today
then is Fermi's golden rule.

00:00:05.120 --> 00:00:07.800
So that's what we're
going to develop.

00:00:07.800 --> 00:00:11.190
Fermi's golden rule.

00:00:11.190 --> 00:00:17.860
Fermi's golden rule.

00:00:20.670 --> 00:00:24.510
And this has to do with
the study of transitions.

00:00:24.510 --> 00:00:28.860
And typically, the interesting
and sophisticated thing

00:00:28.860 --> 00:00:30.960
about this subject
is that you have

00:00:30.960 --> 00:00:36.060
a transition from
some initial state

00:00:36.060 --> 00:00:40.590
to a state that is part
of a continuum of states.

00:00:40.590 --> 00:00:42.370
That's what makes
it complicated.

00:00:42.370 --> 00:00:44.840
The transition from
one discrete state

00:00:44.840 --> 00:00:48.140
to another discrete
state with a perturbation

00:00:48.140 --> 00:00:51.540
is kind of a simple
matter to do.

00:00:51.540 --> 00:00:54.750
But when you can go
into a continuum,

00:00:54.750 --> 00:00:58.060
you have to integrate over
the set of final states,

00:00:58.060 --> 00:01:00.750
and that makes it a
lot more interesting.

00:01:03.750 --> 00:01:10.770
So we go from a discrete
state into a continuum.

00:01:14.610 --> 00:01:18.910
And that makes it
somewhat challenging.

00:01:18.910 --> 00:01:22.570
So we will consider
this in two forms.

00:01:22.570 --> 00:01:24.390
It's worth considering
the case of what

00:01:24.390 --> 00:01:26.980
we call constant perturbations.

00:01:32.000 --> 00:01:33.770
And you might say,
well, aren't we

00:01:33.770 --> 00:01:36.900
doing time dependent
perturbation theory?

00:01:36.900 --> 00:01:38.370
Yes, we are.

00:01:38.370 --> 00:01:40.350
But this kind of
perturbation, you

00:01:40.350 --> 00:01:45.550
will think of it as saying
that H is H0 plus a V that

00:01:45.550 --> 00:01:49.310
is time independent.

00:01:49.310 --> 00:01:55.970
But the way we think of
it is that here is time,

00:01:55.970 --> 00:02:02.060
here is time 0, and here is V.
V turns on at time equals 0.

00:02:02.060 --> 00:02:04.520
So there's a little
bit of time dependence.

00:02:04.520 --> 00:02:08.539
There was no V
before time equals 0,

00:02:08.539 --> 00:02:13.640
and there is a V
after time equals 0.

00:02:13.640 --> 00:02:17.490
So it's almost like the
Hamiltonian changes.

00:02:17.490 --> 00:02:22.520
And we want to see what
transitions we get.

00:02:22.520 --> 00:02:25.640
Because always the subject is
the subject of transitions.

00:02:25.640 --> 00:02:30.020
And we will see an application
of this later today.

00:02:30.020 --> 00:02:39.680
The other case is what is
called a harmonic perturbation,

00:02:39.680 --> 00:02:48.110
in which H is equal
to H0 plus delta H.

00:02:48.110 --> 00:02:53.120
And delta H is going
to be harmonic.

00:02:53.120 --> 00:02:54.510
2H prime.

00:02:54.510 --> 00:02:59.330
That's conventions
to put a 2 in there.

00:02:59.330 --> 00:03:00.770
It's convenient.

00:03:00.770 --> 00:03:04.520
Times cosine omega t.

00:03:04.520 --> 00:03:10.880
And this will be for
t between 0 and t0.

00:03:10.880 --> 00:03:13.640
Sorry if some people
can't see this.

00:03:13.640 --> 00:03:16.520
Too far to the right.

00:03:16.520 --> 00:03:21.830
2H prime cosine of omega t
when t is between 0 and t0.

00:03:21.830 --> 00:03:23.710
And it's 0 otherwise.

00:03:26.800 --> 00:03:31.950
And H prime, of course, is
time independent as well.

00:03:31.950 --> 00:03:35.590
Time independent.

00:03:35.590 --> 00:03:41.050
So these are the
two main cases when

00:03:41.050 --> 00:03:45.760
we will consider transitions.

00:03:45.760 --> 00:03:51.370
And we'll do the first one
in a lot of detail today.

00:03:51.370 --> 00:03:54.430
The subject is trying to get
the transitions understood

00:03:54.430 --> 00:03:55.900
for this case.

00:03:55.900 --> 00:03:59.470
And after that,
next time, we will

00:03:59.470 --> 00:04:02.770
do the case of the
harmonic perturbation that

00:04:02.770 --> 00:04:07.120
brings in a few new issues.

00:04:07.120 --> 00:04:10.400
But then, all of the
rest is the same.

00:04:10.400 --> 00:04:14.680
So once you've understood how
to do the constant transition,

00:04:14.680 --> 00:04:18.970
the harmonic perturbation
is going to be easier.

00:04:18.970 --> 00:04:22.089
So constant
perturbations will be

00:04:22.089 --> 00:04:25.210
useful to understand, for
example, the phenomenon

00:04:25.210 --> 00:04:28.270
of auto-ionization.

00:04:28.270 --> 00:04:33.130
Some atoms sometimes
ionize spontaneously,

00:04:33.130 --> 00:04:36.970
and this has to do
with this subject.

00:04:36.970 --> 00:04:40.240
Harmonic perturbations has to
do with even a more popular

00:04:40.240 --> 00:04:44.330
subject, which is atoms
interacting with radiation.

00:04:44.330 --> 00:04:46.930
You send in an
electromagnetic field.

00:04:46.930 --> 00:04:51.340
That is a harmonically
varying perturbation,

00:04:51.340 --> 00:04:54.220
and that's going to allow
you to calculate transitions.

00:04:54.220 --> 00:04:56.680
And by the time
we're done with this,

00:04:56.680 --> 00:05:03.580
you will be experts in
calculating atomic transitions.

00:05:03.580 --> 00:05:06.640
So one subject we
need since we're

00:05:06.640 --> 00:05:11.070
going to be doing transitions
between a discrete state

00:05:11.070 --> 00:05:16.540
or continuum is to
describe the continuum.

00:05:16.540 --> 00:05:19.810
And for that, for
the continuum, we

00:05:19.810 --> 00:05:30.352
use the concept of a density
of states in the continuum.

00:05:37.280 --> 00:05:43.330
Now, we will be considering
momentum eigenstates.

00:05:43.330 --> 00:05:45.520
And the momentum,
you know very well

00:05:45.520 --> 00:05:48.280
the momentum of a free
particle is a continuum.

00:05:48.280 --> 00:05:51.580
Takes absolutely
continuous values.

00:05:51.580 --> 00:05:54.190
So there is no way
you can count them

00:05:54.190 --> 00:06:00.850
or you can tell how many
there are per momentum range.

00:06:00.850 --> 00:06:05.770
It's like saying how many
numbers are there from 0 to 1.

00:06:05.770 --> 00:06:08.230
The way you have
to do that always

00:06:08.230 --> 00:06:13.090
is by adding an
extra parameter you

00:06:13.090 --> 00:06:15.070
wish you didn't have to add.

00:06:18.120 --> 00:06:20.400
And that's kind of
things that we have

00:06:20.400 --> 00:06:22.140
to do in physics sometimes.

00:06:22.140 --> 00:06:23.460
We have to add parameters.

00:06:23.460 --> 00:06:27.390
You consider when we discussed
last time the delta function

00:06:27.390 --> 00:06:32.160
perturbation, we broadened it,
and we were able to calculate.

00:06:32.160 --> 00:06:35.580
And then we saw that
the broadening size

00:06:35.580 --> 00:06:40.170
didn't matter for the
result. Therefore, good.

00:06:40.170 --> 00:06:42.400
So here, we'll do the same.

00:06:42.400 --> 00:06:44.670
You know that the
momentum eigenstates

00:06:44.670 --> 00:06:50.010
of a particle in an interval
or in a circle are quantized.

00:06:50.010 --> 00:06:52.890
And then you can count
quantized states.

00:06:52.890 --> 00:06:56.070
So we will put the
whole world in a box.

00:06:56.070 --> 00:06:57.720
A big box.

00:06:57.720 --> 00:07:02.130
Size of the galaxy, size of the
Earth, size of the laboratory

00:07:02.130 --> 00:07:04.320
maybe is big enough.

00:07:04.320 --> 00:07:09.540
And we'll put this parameter l
there for the size of the box.

00:07:09.540 --> 00:07:11.520
And we will get the
density of states

00:07:11.520 --> 00:07:14.800
because now the
states can be counted.

00:07:14.800 --> 00:07:17.100
Afterwards, by
the time we're all

00:07:17.100 --> 00:07:21.330
done with the transition rates
of the Fermi's golden rule,

00:07:21.330 --> 00:07:25.510
we hope that length
is going to disappear.

00:07:25.510 --> 00:07:28.080
So that's something
we will see that it

00:07:28.080 --> 00:07:29.850
happens in calculations.

00:07:29.850 --> 00:07:33.190
And there will be good
reasons why it happens.

00:07:33.190 --> 00:07:35.400
So we'll put the
world in a big box.

00:07:40.060 --> 00:07:42.150
It's not that big
the way I draw it,

00:07:42.150 --> 00:07:45.810
but it's supposed
to be very big.

00:07:45.810 --> 00:07:47.370
Length l.

00:07:47.370 --> 00:07:50.760
And you think of the
quantization of momentum

00:07:50.760 --> 00:08:02.694
by considering a wave function
that would be normalized

00:08:02.694 --> 00:08:03.910
and has--

00:08:03.910 --> 00:08:06.090
it's a momentum eigenstate.

00:08:10.536 --> 00:08:13.640
It has momentum in the x
direction, y direction,

00:08:13.640 --> 00:08:14.460
and z direction.

00:08:14.460 --> 00:08:18.770
And that wave function,
if all the sides

00:08:18.770 --> 00:08:24.680
of this world, this cubical
world with length l,

00:08:24.680 --> 00:08:29.220
that wave function is
properly normalized.

00:08:29.220 --> 00:08:34.470
You square it, and the norm is
equal to 1 times this factor

00:08:34.470 --> 00:08:34.970
squared.

00:08:34.970 --> 00:08:36.020
You integrate it.

00:08:36.020 --> 00:08:36.890
It's all good.

00:08:36.890 --> 00:08:41.179
So this is a nice
wave function in which

00:08:41.179 --> 00:08:47.900
the integral psi squared V
cubed x over the box is 1.

00:08:47.900 --> 00:08:49.700
And then what else do we do?

00:08:49.700 --> 00:08:53.840
We decide that we have to
consider periodic boundary

00:08:53.840 --> 00:08:54.650
conditions.

00:08:54.650 --> 00:08:59.300
Well, think of it as a
torus, properly speaking,

00:08:59.300 --> 00:09:02.300
in which each
direction is a circle.

00:09:02.300 --> 00:09:06.110
So the wave function
repeats itself

00:09:06.110 --> 00:09:11.750
after x increases by l,
after y increases by l,

00:09:11.750 --> 00:09:15.140
or after z increases by l.

00:09:15.140 --> 00:09:19.760
You could have chosen a box with
a finite big wall at the end.

00:09:19.760 --> 00:09:22.610
It would make no
difference for the counting

00:09:22.610 --> 00:09:24.630
at the end of the day.

00:09:24.630 --> 00:09:31.670
So the conditions are
that kx multiplied by l

00:09:31.670 --> 00:09:40.830
should be a multiple
of nx 2 pi times nx,

00:09:40.830 --> 00:09:52.380
ky times l, 2 pi times
ny, kz times l, 2 pi nz.

00:09:52.380 --> 00:09:59.730
And therefore, the
total number of states

00:09:59.730 --> 00:10:04.620
can be calculated by taking
a little differential

00:10:04.620 --> 00:10:05.400
of this thing.

00:10:05.400 --> 00:10:09.660
We say, well, if we let
kx vary by a little bit,

00:10:09.660 --> 00:10:21.440
dkx from this equation times
l is 2 pi dnx dky times l,

00:10:21.440 --> 00:10:28.920
2 pi times dny, and dkz
times l is 2 pi times dnz.

00:10:32.100 --> 00:10:37.140
So if I were to consider a
little interval of momentum

00:10:37.140 --> 00:10:43.950
defined by dkx, dky, dkz, a
little cube in momentum space,

00:10:43.950 --> 00:10:47.130
the number of
states in this cube

00:10:47.130 --> 00:10:58.170
would be dn, which
is dnx, dny, dnz.

00:10:58.170 --> 00:11:10.470
And this is equal to l over 2 pi
cubed d cubed k, which is dkx,

00:11:10.470 --> 00:11:12.720
dky, dkz.

00:11:12.720 --> 00:11:19.460
I say, if in this momentum range
the quantum number nx contains

00:11:19.460 --> 00:11:25.880
this number of values,
dnx, the quantum number ny

00:11:25.880 --> 00:11:29.600
can take a set of values, and
the same for the quantum number

00:11:29.600 --> 00:11:31.290
in the z direction.

00:11:31.290 --> 00:11:35.630
The total set of quantum
numbers is the product of them.

00:11:35.630 --> 00:11:41.510
So there is that many
states in this little cube

00:11:41.510 --> 00:11:42.950
of momentum space--

00:11:42.950 --> 00:11:45.740
that is, momentum space
that is between some

00:11:45.740 --> 00:11:48.995
k and a little more, called dk.

00:11:54.680 --> 00:11:57.140
Well, that's a famous formula.

00:11:57.140 --> 00:12:08.660
And the n is equal to l
over 2 pi cubed d cubed k.

00:12:08.660 --> 00:12:10.940
Some people know this by heart.

00:12:10.940 --> 00:12:15.740
But now we want to
write this in the terms

00:12:15.740 --> 00:12:19.565
of a density of states
as a function of energy.

00:12:25.860 --> 00:12:29.720
So if this is the
number of states

00:12:29.720 --> 00:12:34.310
in some interval
in momentum space,

00:12:34.310 --> 00:12:37.820
I can try to convert
this into saying,

00:12:37.820 --> 00:12:42.890
well, all these states,
because the momentum is varying

00:12:42.890 --> 00:12:45.710
within some little
bounds, the energy

00:12:45.710 --> 00:12:47.840
is varying within
some little bounds.

00:12:47.840 --> 00:12:51.560
So let's figure out
how much the dE is.

00:12:51.560 --> 00:12:54.950
And then the total
number of states

00:12:54.950 --> 00:12:59.060
will be given in this
range by the number

00:12:59.060 --> 00:13:02.960
of states per unit energy
multiplied by this thing.

00:13:02.960 --> 00:13:10.660
So this is the density
of states, which

00:13:10.660 --> 00:13:16.760
is states per unit energy.

00:13:21.730 --> 00:13:25.600
And many times, this is the
quantity we really want.

00:13:30.280 --> 00:13:34.640
We have here an
energy E as well.

00:13:34.640 --> 00:13:39.280
So we think of this V
cubed k as a little cube

00:13:39.280 --> 00:13:40.660
in momentum space.

00:13:40.660 --> 00:13:43.720
If momentum space
has an origin here,

00:13:43.720 --> 00:13:48.390
you can imagine states with
some momentum, and then

00:13:48.390 --> 00:13:51.740
a little cube in here saying
how much the momentum varies.

00:13:51.740 --> 00:13:54.110
So all these states
have some momentum

00:13:54.110 --> 00:13:56.450
and vary a little
bit in the momentum.

00:13:56.450 --> 00:13:59.510
Therefore, they all
basically have some energy

00:13:59.510 --> 00:14:02.940
up to some little variation.

00:14:02.940 --> 00:14:06.500
So how do we connect
these two things?

00:14:06.500 --> 00:14:16.740
Well, you remember, E is equal
to h squared k squared over 2m.

00:14:19.820 --> 00:14:21.515
And how about d cubed k?

00:14:24.120 --> 00:14:31.550
We'll try to think of d cubed
k physically as all the states

00:14:31.550 --> 00:14:34.670
that have momentum k.

00:14:34.670 --> 00:14:37.775
And now, this is a
space diagram now.

00:14:40.400 --> 00:14:42.210
This is x, y, and z.

00:14:45.050 --> 00:14:48.070
So here are the states
with momentum k.

00:14:48.070 --> 00:14:51.540
They point in this direction.

00:14:51.540 --> 00:14:54.710
So this is the direction
the states are pointing.

00:14:54.710 --> 00:15:02.330
And now we think of them as
having some possible angle

00:15:02.330 --> 00:15:03.560
here.

00:15:03.560 --> 00:15:04.340
d omega.

00:15:15.690 --> 00:15:18.060
So this is the direction.

00:15:18.060 --> 00:15:20.610
So let me not say
they're x, y, and z.

00:15:20.610 --> 00:15:22.340
Let me use angles here.

00:15:22.340 --> 00:15:33.450
Theta and phi, representing
the azimuthal and polar angle

00:15:33.450 --> 00:15:34.600
of this direction.

00:15:34.600 --> 00:15:41.550
So we have a little range
here and a little range

00:15:41.550 --> 00:15:44.530
of magnitude of momentum.

00:15:44.530 --> 00:15:47.030
So let me see if I can
draw it kind of nicer.

00:15:49.950 --> 00:15:53.630
You can imagine a
little cone as it grows,

00:15:53.630 --> 00:15:59.320
and the last part is a little
thick piece of the cone here

00:15:59.320 --> 00:16:00.730
of thickness dk.

00:16:04.950 --> 00:16:20.680
So d cubed k is the volume
of that little pillbox here.

00:16:20.680 --> 00:16:28.020
And it's k square
d omega times dk.

00:16:34.360 --> 00:16:36.370
So that's d cubed k.

00:16:36.370 --> 00:16:38.380
And now to relate
it to the energy

00:16:38.380 --> 00:16:40.490
we have here, this equation.

00:16:40.490 --> 00:16:42.520
So we take a differential.

00:16:42.520 --> 00:16:50.695
And it's d energy is
h squared kdk over m.

00:16:58.070 --> 00:17:08.540
So d cubed k here is k
dk, k d omega times k vk,

00:17:08.540 --> 00:17:11.700
and this is the same k dk here.

00:17:11.700 --> 00:17:16.859
So we can write it as k d omega.

00:17:16.859 --> 00:17:26.240
And for k dk we have m
over h squared d energy.

00:17:26.240 --> 00:17:28.970
So we're almost done
with our computation.

00:17:35.760 --> 00:17:41.370
We go back to this equation.

00:17:41.370 --> 00:17:48.940
And we have rho of E dE,
or dn, is equal to l over

00:17:48.940 --> 00:17:56.030
2 pi cubed d cubed k.

00:17:56.030 --> 00:17:58.910
But now we have
what d cubed k is.

00:17:58.910 --> 00:18:06.570
It's m over h
squared k d omega dE.

00:18:11.900 --> 00:18:15.830
So this is our formula.

00:18:15.830 --> 00:18:18.350
And this is rho of E dE.

00:18:21.470 --> 00:18:29.650
And here, for example,
you have rho of E.

00:18:29.650 --> 00:18:36.760
Actually, if it were rho of E,
I should have here k expressed

00:18:36.760 --> 00:18:42.850
in terms of E. Just makes the
formula a little more messy.

00:18:42.850 --> 00:18:50.860
But if you think of it as
a function of the energy,

00:18:50.860 --> 00:18:52.840
you have to say
the words properly.

00:18:52.840 --> 00:18:55.540
What is rho of E?

00:18:55.540 --> 00:19:00.490
Rho of E is the density
of states per unit energy,

00:19:00.490 --> 00:19:03.700
but when you're
only counting states

00:19:03.700 --> 00:19:08.530
that have a momentum
within an angle d omega.

00:19:08.530 --> 00:19:16.490
So this is the density of
states per unit energy,

00:19:16.490 --> 00:19:17.930
because this gives you states.

00:19:17.930 --> 00:19:20.720
So this must be states
per unit energy.

00:19:20.720 --> 00:19:23.390
But you're only looking
at the states that

00:19:23.390 --> 00:19:25.910
are within an angle d omega.

00:19:25.910 --> 00:19:29.510
If you were to look at the
density of all states that

00:19:29.510 --> 00:19:34.100
have energy E, you would
have to integrate over omega.

00:19:34.100 --> 00:19:38.210
But many times, we want
to make a transition.

00:19:38.210 --> 00:19:40.030
You have an atom
and you're sending

00:19:40.030 --> 00:19:41.350
an electromagnetic wave.

00:19:41.350 --> 00:19:42.890
And you want to
make a transition,

00:19:42.890 --> 00:19:46.700
and you want to see how
many electrons, for example,

00:19:46.700 --> 00:19:49.190
are kicked out in
some direction.

00:19:49.190 --> 00:19:54.830
So you need the density of
states within some solid angle.

00:19:54.830 --> 00:19:58.130
So this is a
relatively useful thing

00:19:58.130 --> 00:20:02.600
to have, not to integrate,
so that you can keep control

00:20:02.600 --> 00:20:06.230
over your states and orient
them at a given angle.

00:20:09.230 --> 00:20:16.310
So we've done a little
bit of basic preparation.

00:20:16.310 --> 00:20:20.570
We've said what the Fermi
golden rule aims to do,

00:20:20.570 --> 00:20:24.740
and it requires transition
to the continuum where we'll

00:20:24.740 --> 00:20:28.070
have to use density of states.

00:20:28.070 --> 00:20:30.590
And here is an example
of how you calculate

00:20:30.590 --> 00:20:32.830
the density of states.