WEBVTT

00:00:00.500 --> 00:00:03.480
PROFESSOR: There is an
effective potential,

00:00:03.480 --> 00:00:05.160
as it's described here.

00:00:05.160 --> 00:00:08.630
Yes, there is a potential,
the central force,

00:00:08.630 --> 00:00:11.250
but if there is
angular momentum,

00:00:11.250 --> 00:00:14.500
it's like a centrifugal barrier.

00:00:14.500 --> 00:00:17.850
With angular momentum
it becomes very

00:00:17.850 --> 00:00:20.535
difficult to reach the origin.

00:00:20.535 --> 00:00:22.485
Because if you want
to reach the origin,

00:00:22.485 --> 00:00:24.495
you have to spin
faster and faster.

00:00:24.495 --> 00:00:27.800
So it becomes very hard
to reach the origin.

00:00:27.800 --> 00:00:31.347
And this is like a
repulsive potential here.

00:00:31.347 --> 00:00:35.670
So you have an
effective potential.

00:00:35.670 --> 00:00:37.870
You started with
just a potential,

00:00:37.870 --> 00:00:40.222
the central potential,
but that will

00:00:40.222 --> 00:00:41.885
become an effective potential.

00:00:45.950 --> 00:00:49.260
This whole thing in
brackets over there

00:00:49.260 --> 00:00:53.360
is sometimes called the
effective radial potential.

00:00:53.360 --> 00:01:00.340
And it's this V of r plus h
squared over 2mr squared l

00:01:00.340 --> 00:01:03.510
times l plus 1.

00:01:03.510 --> 00:01:09.090
So if you have a
Coulomb potential,

00:01:09.090 --> 00:01:15.050
that would try to
make the electrons go

00:01:15.050 --> 00:01:16.440
all the way to the proton.

00:01:16.440 --> 00:01:18.980
This is attractive.

00:01:18.980 --> 00:01:23.630
Well, even that is not
problematic for l equals 0.

00:01:23.630 --> 00:01:27.760
But for l different from
0, you have to add here

00:01:27.760 --> 00:01:29.830
a potential that diverges.

00:01:29.830 --> 00:01:35.990
And it's positive, say,
for some l positive

00:01:35.990 --> 00:01:40.684
and falls off very fast because
forms of like r squared, this,

00:01:40.684 --> 00:01:44.560
the Coulomb potential
falls off at 1/r.

00:01:44.560 --> 00:01:49.770
So when you combine the
two, so this is just

00:01:49.770 --> 00:01:53.480
the centrifugal barrier,
but you combine the two,

00:01:53.480 --> 00:01:57.200
this diverges faster as
well, so the total potential

00:01:57.200 --> 00:02:00.473
is something like
that in between.

00:02:03.300 --> 00:02:06.470
And then you can
have bound states.

00:02:06.470 --> 00:02:10.820
And all the theorems we've
learned about bound states

00:02:10.820 --> 00:02:16.250
and eigenstates of
one-dimensional potential, now

00:02:16.250 --> 00:02:20.630
you discover that they're very
useful in three dimensions.

00:02:20.630 --> 00:02:24.030
Many things just carry through.

00:02:24.030 --> 00:02:26.630
So this is effective potential.

00:02:26.630 --> 00:02:31.320
And it's going to
recap and remember

00:02:31.320 --> 00:02:36.010
that the solution that
you've written is Re.

00:02:36.010 --> 00:02:49.674
But Re is u of r over r
Y lm of theta and phi.

00:02:52.440 --> 00:03:02.810
And one more thing that
I want you to realize,

00:03:02.810 --> 00:03:09.770
does the function u depend on l?

00:03:09.770 --> 00:03:11.060
Yes.

00:03:11.060 --> 00:03:13.460
The differential
equation depends on l.

00:03:13.460 --> 00:03:15.942
Does the function u depend on m?

00:03:19.180 --> 00:03:22.480
Yes or no?

00:03:22.480 --> 00:03:25.370
STUDENT: [INAUDIBLE]

00:03:25.370 --> 00:03:26.420
PROFESSOR: No.

00:03:26.420 --> 00:03:28.990
There's no m in the
differential equation,

00:03:28.990 --> 00:03:32.970
so that should be good enough.

00:03:32.970 --> 00:03:35.850
So what's happened
is the m dependence

00:03:35.850 --> 00:03:42.440
is kind of very simple,
always very simple.

00:03:42.440 --> 00:03:46.885
It is the e to the im
phi and nothing else.

00:03:46.885 --> 00:03:49.962
The u doesn't know about it.

00:03:49.962 --> 00:03:53.230
On the other hand,
the u depends on l

00:03:53.230 --> 00:03:55.440
because it shows in the
differential equation.

00:03:55.440 --> 00:04:04.240
So I could write here of e
and l because it depends on l,

00:04:04.240 --> 00:04:08.830
and it is the solutions of
the Schrodinger equation

00:04:08.830 --> 00:04:10.030
in one potential.

00:04:10.030 --> 00:04:12.690
So there will be
quantization of energy,

00:04:12.690 --> 00:04:14.740
or there might be
stationary states

00:04:14.740 --> 00:04:17.529
that depend on the energy.

00:04:17.529 --> 00:04:23.850
So this is the function
that knows about l,

00:04:23.850 --> 00:04:25.935
knows about the energy.

00:04:25.935 --> 00:04:28.790
And we've been
totally successful

00:04:28.790 --> 00:04:31.130
with the angular dependence.

00:04:31.130 --> 00:04:31.790
Yes.

00:04:31.790 --> 00:04:34.320
STUDENT: That should
be theta and phi, dy?

00:04:34.320 --> 00:04:36.830
PROFESSOR: Yes, theta and phi.

00:04:36.830 --> 00:04:37.330
Thank you.

00:04:40.461 --> 00:04:40.960
Good.

00:04:43.470 --> 00:04:49.030
Normalization, the last thing
that has to work out nicely.

00:04:49.030 --> 00:04:53.640
Let's try to see what
does the normalization say

00:04:53.640 --> 00:04:55.440
about this function.

00:04:55.440 --> 00:05:00.030
Well, we should find
that the integral of psi

00:05:00.030 --> 00:05:06.840
squared d cubed x is equal to 1.

00:05:06.840 --> 00:05:11.190
But that integral,
as you now know,

00:05:11.190 --> 00:05:19.730
it's the integral of r
squared dr from 0 to infinity.

00:05:19.730 --> 00:05:21.860
And how do you
integrate over volume?

00:05:21.860 --> 00:05:24.340
You integrate over r.

00:05:24.340 --> 00:05:26.710
That has the right units.

00:05:26.710 --> 00:05:31.730
And you integrate over
solid angle, d omega.

00:05:34.590 --> 00:05:40.340
Of psi squared, so let's
do the arithmetic here.

00:05:40.340 --> 00:05:53.450
We have a Y lm star of theta and
phi, a Y lm of theta and phi.

00:05:53.450 --> 00:05:59.810
We have a u squared
and then r squared.

00:06:03.740 --> 00:06:06.560
Poor graph.

00:06:06.560 --> 00:06:10.240
But anyway, you can read it.

00:06:10.240 --> 00:06:16.270
Now, what happens
is just good stuff.

00:06:16.270 --> 00:06:19.404
r squared cancels.

00:06:19.404 --> 00:06:23.290
And this solid
angle angle integral

00:06:23.290 --> 00:06:26.680
is a perfect integral
for our normalization.

00:06:26.680 --> 00:06:29.540
So this gives you 1.

00:06:29.540 --> 00:06:32.530
And therefore the end
result for all this integral

00:06:32.530 --> 00:06:40.380
is just the integral from 0 to
infinity of dr u of r squared.

00:06:43.170 --> 00:06:46.520
And that must be equal to 1.

00:06:46.520 --> 00:06:53.210
So not only is little
u a nice variable

00:06:53.210 --> 00:06:57.650
that satisfies a one-dimensional
Schrodinger equation,

00:06:57.650 --> 00:07:02.930
but you can remember that
your more complicated wave

00:07:02.930 --> 00:07:08.320
function will be normalized
if u is normalized

00:07:08.320 --> 00:07:12.260
in the one-dimensional sense.

00:07:12.260 --> 00:07:17.330
If u squared integrated
over x is equal to 1, yes,

00:07:17.330 --> 00:07:21.170
you're normalized.

00:07:21.170 --> 00:07:28.850
So the set-up to convert the
three-dimensional differential

00:07:28.850 --> 00:07:31.730
equation into a one-dimensional
differential equation

00:07:31.730 --> 00:07:34.520
has been very successful.

00:07:34.520 --> 00:07:36.905
We've reduced it to a
one-dimensional problem.

00:07:36.905 --> 00:07:39.980
We have to solve those.

00:07:39.980 --> 00:07:43.490
Each time somebody gives
you a spherical potential,

00:07:43.490 --> 00:07:46.930
you look at that equation,
the radial equation,

00:07:46.930 --> 00:07:49.430
and try to solve for u's.

00:07:49.430 --> 00:07:51.417
If you solve for
u's, then you can

00:07:51.417 --> 00:07:54.800
append the angular
dependents that

00:07:54.800 --> 00:07:59.180
correspond to angular
momentum eigenstates.

00:07:59.180 --> 00:08:03.130
We call this angular
moment eigenstates.

00:08:03.130 --> 00:08:07.520
They are the most you can
ask from angular momentum.

00:08:07.520 --> 00:08:11.160
And then you have
many solutions.

00:08:11.160 --> 00:08:15.280
So I want to conclude
that part of the analysis

00:08:15.280 --> 00:08:19.610
by mentioning something
about solutions

00:08:19.610 --> 00:08:26.870
with the appropriate
boundary conditions.

00:08:26.870 --> 00:08:30.680
We have x, in the
one-dimensional problem

00:08:30.680 --> 00:08:31.410
we call it x.

00:08:31.410 --> 00:08:35.340
And you'd run from minus
infinity to infinity.

00:08:35.340 --> 00:08:38.600
The one difference here
is that you have r,

00:08:38.600 --> 00:08:41.909
and r runs from 0 to infinity.

00:08:41.909 --> 00:08:48.260
So you may wonder if you have
some issues with r going to 0.

00:08:48.260 --> 00:08:52.180
What should the wave
function do when r goes to 0?

00:08:54.910 --> 00:08:55.662
OK.

00:08:55.662 --> 00:09:00.590
Then the way we think about
it is not completely general

00:09:00.590 --> 00:09:03.550
but is good enough.

00:09:03.550 --> 00:09:06.780
We think of the
differential equation

00:09:06.780 --> 00:09:11.525
as we have here and
imagine a potential

00:09:11.525 --> 00:09:17.630
that when r goes to 0, the
centrifugal barrier dominates.

00:09:17.630 --> 00:09:22.470
So our potential of the
form 1 over r to the fourth

00:09:22.470 --> 00:09:25.300
would be even more
singular than the barrier.

00:09:25.300 --> 00:09:27.860
And I don't know what
happens in that case.

00:09:27.860 --> 00:09:30.659
Most likely it's no
good, not interesting.

00:09:30.659 --> 00:09:33.900
You cannot find solutions.

00:09:33.900 --> 00:09:37.230
But if the potential, like
the Coulomb potential,

00:09:37.230 --> 00:09:43.240
is weaker than the centrifugal
barrier as r goes to 0,

00:09:43.240 --> 00:09:47.940
the centrifugal barrier
dominates when r goes to 0.

00:09:47.940 --> 00:09:52.673
And this differential
equation, as r goes to 0,

00:09:52.673 --> 00:09:59.220
has a potential infinite
term which corresponds

00:09:59.220 --> 00:10:02.110
to the centrifugal barrier.

00:10:02.110 --> 00:10:12.370
So we think of this
as r goes to 0,

00:10:12.370 --> 00:10:14.650
the differential
equation roughly

00:10:14.650 --> 00:10:24.620
becomes minus h squared over
2m d second u dr squared

00:10:24.620 --> 00:10:35.740
plus h squared l times l plus 1
over r squared of u, roughly 0.

00:10:35.740 --> 00:10:40.130
At least the leading
behavior of these things

00:10:40.130 --> 00:10:43.570
should work out correctly.

00:10:43.570 --> 00:10:47.600
So the h squareds--

00:10:47.600 --> 00:10:50.150
here is a 2m, as
well, I'm sorry.

00:10:53.670 --> 00:10:57.740
The h squared over 2m's
cancel, and you get

00:10:57.740 --> 00:11:02.520
d second u dr squared plus--

00:11:02.520 --> 00:11:11.010
no, is equal to l times l
plus 1 over r squared u.

00:11:11.010 --> 00:11:16.460
And we're only interested
for this as r goes to 0.

00:11:16.460 --> 00:11:19.450
And it's not an exact statement.

00:11:19.450 --> 00:11:21.580
It's a discovery.

00:11:21.580 --> 00:11:24.440
We're trying to discover
what's happening with the wave

00:11:24.440 --> 00:11:26.260
function near r equals 0.

00:11:26.260 --> 00:11:29.410
Well, this has two
kinds of solutions.

00:11:29.410 --> 00:11:32.110
You can try a polynomial.

00:11:32.110 --> 00:11:38.590
So this could go like
r to the l plus 1.

00:11:38.590 --> 00:11:42.430
If you take two
derivatives, that works out.

00:11:42.430 --> 00:11:46.355
You get the l plus 1 and l.

00:11:46.355 --> 00:11:48.610
And it solves the equation.

00:11:48.610 --> 00:11:57.580
Or r to the minus l also
solves the equation.

00:11:57.580 --> 00:11:59.930
As you can imagine,
this is going

00:11:59.930 --> 00:12:01.900
to be problematic in general.

00:12:01.900 --> 00:12:03.595
It's too divergent.

00:12:03.595 --> 00:12:08.120
It will not be possible to
normalize it, in general,

00:12:08.120 --> 00:12:10.480
for arbitrary values of l.

00:12:10.480 --> 00:12:16.310
So this is a very
brief analysis.

00:12:16.310 --> 00:12:18.450
I'm not going into
all the detail

00:12:18.450 --> 00:12:22.370
that probably this
deserves at this moment.

00:12:22.370 --> 00:12:26.076
But this is ruled out,
and this is ruled in.

00:12:30.740 --> 00:12:35.480
The only thing-- so it's true
that this one is ruled out,

00:12:35.480 --> 00:12:39.015
and it has problems
for normalization.

00:12:39.015 --> 00:12:41.780
It is too divergent as l.

00:12:41.780 --> 00:12:48.320
But for l equals 0,
it's not divergent.

00:12:48.320 --> 00:12:51.800
But for l equals 0,
there's another reason

00:12:51.800 --> 00:12:53.555
why this is not good.

00:12:53.555 --> 00:12:58.470
It turns out that for l
equals 0 this doesn't quite

00:12:58.470 --> 00:13:02.990
solve the Schrodinger equation,
the exact Schrodinger equation.

00:13:02.990 --> 00:13:06.740
So the bottom line
of this analysis

00:13:06.740 --> 00:13:13.863
is that we will have u of r
behave like r to the l plus 1

00:13:13.863 --> 00:13:16.810
as r goes to 0.

00:13:16.810 --> 00:13:23.330
And in particular, when
l is equal to 0, u of r

00:13:23.330 --> 00:13:25.190
will behave like r.

00:13:25.190 --> 00:13:28.602
So it will vanish
as r goes to 0.

00:13:33.270 --> 00:13:36.690
So the only question is
how fast it vanishes.

00:13:36.690 --> 00:13:41.960
It vanishes as r goes
to 0 for l equals 0.

00:13:41.960 --> 00:13:46.030
It vanishes even
faster for higher l.

00:13:46.030 --> 00:13:50.040
So it always vanishes, the
wave function at r equals 0.

00:13:50.040 --> 00:13:53.320
And that's why we can
usually think of it

00:13:53.320 --> 00:13:56.410
as having an infinite barrier.

00:13:56.410 --> 00:14:00.260
The wave function could not
exist for r less than 0.

00:14:00.260 --> 00:14:03.005
That physically doesn't exist.

00:14:03.005 --> 00:14:10.050
And the boundary conditions are
such that that cannot happen.

00:14:10.050 --> 00:14:12.590
All right.

00:14:12.590 --> 00:14:20.430
We've finished the discussion
of the radial equation.

00:14:20.430 --> 00:14:24.270
Now, our main interest with
the radial equation, of course,

00:14:24.270 --> 00:14:26.530
is, at this moment,
the hydrogen atom.

00:14:26.530 --> 00:14:29.270
So we're going to turn to that.