Lecture 33

IIT Kanpur July 201
20 followers
Oct 22, 2023
Now, let us apply it for hydrogen atom. So, hydrogen atom, so we have operator f of r.
I want to compute the expectation value. All of you, most of you are in physics department,
so f of r. Some people who are not in, from physics, so this is basically the integral,
d r, it is a 3D integral, f r psi r square, mod square. So, this is a wave function.
For ground state, this is called ground state, this is 1 0 0 l n l m. So, ground state is
this. It is a non-dimensionalized ground state, which I think, I did that in several lectures
before. So, we have it, so I need to do the integral. Now, in 3D, d r is, now by the way,
this is a spherically symmetric, spherically symmetric. So, there is no theta dependence.
If there is theta dependence, then you have to do the theta integral first, then go to
r. But I want to just illustrate integral with r, but with theta also it can be done,
no problem, we can just do it. So, I want to compute r expectation.
So, if you recall, so let me just do it in the side, d r will be 4 pi r square d r. I
hope this, people who are not from physics, you have done this in your math course. This
is d r, 3D volume integral, f r is r and mod psi square, I need to integrate, basically
mod square is a real function, so mod is not a problem, minus 2 r pi pi. So, pi has
been pulled out, 4 pi r square here and another r, but I make r prime. So, this is, for non-dimensional
stuff, I just make everything r prime, so d r prime, d r r prime.
So, what do I get? So, this is a pi pi will cancel, but this 4 is coming out, but there
is a problem, there is a argument of integral is, exponential is minus 2 r prime, but inside
is, I mean the integral is d r prime. So, you make a simple change of variable, 2 r
prime equal to, you call it s prime, r s. So, we just rewrite as 2 r prime is your variable
to integrate. So, how many of them are here? There is a 2 cube and 2, so there are 16,
so 16 is, I have to multiply and divide by 16.
So, by change of variable, I am getting 1 by 16 outside, so I get 1 by 4, this is 1
by 4 and this part, we can do by Gauss quadrature, because this is basically integral e to power
minus s prime s cube d s prime. Now, how many points do we need? We can just do it 2 points,
s cube with n equal to 2, I am not doing the algebra here, but we can just, I mean
I am not trying to do real mathematical integral and computational integral, but you can just
invoke Gauss quadrature and you will get the whole answer to be 3 by 2.
So, this is the power of Gauss quadrature. So, in fact, any function, you can make any
r prime here, it will work. We just use more points if r is, I mean this function is higher
power of n, so 3 by 2. Now, we can also compute 1 by r prime. So, 1 by r prime, we can just,
well I mean instead of r cube, it is going to be r, because 1 r will get cancelled and
this is 1. Now, we can see one interesting thing is that 1 by r expectation value or
average is not 1 by average r. This is, r is not a number, r is an operator. So, is expect,
I mean this is, we get that. We can also compute the potential energy.
So, what is the potential energy? So, u is, in CGS minus u square by r. So, this,
now if a non-dimensional line, this is basically 1 by r. But in any case, you can put this
r and is r prime r naught. r naught is a Bohr radius. So, it is minus e square by r naught,
which is 13.6, well it is not 13.6, so this is r naught, 1 by r prime is 1. So, this is u.
Kinetic energy, I am not, well it is not, we cannot use this cancellation, but you can use
virial theorem. So, kinetic energy is half of that, which is going to be plus is positive,
2 r naught. So, total energy is kinetic plus potential. So, you add these two,
is going to be minus 2 by 2 r naught. So, I hope I did not make algebraic error,
but this is 13.6 electron volts. Anyway, so whatever algebra you did in quantum mechanics,
you can also do in Gauss-Cortes. Let us end this topic with the third polynomial,
Gauss-Hermit. So, Gauss-Hermit gives you integral for quantum oscillator. So, these are first 3,
first 4 of this Hermit polynomial. I am going to tell you orthogonal property in the next slide.
So, 4 of them and we can plot them. So, these are again functions we are growing very fast with x,
see it reaches 10 power 5 with when x is minus 10. That is why we need to suppress it by e to
the power minus x square. These polynomials are growing very fast when with, for large x.
So, let us look at the functional form is minus infinity to plus infinity. So,
a is minus infinity, b is plus infinity and I do the integral and f x is that function.
So, this is the weight function, which comes from the orthogonal property
that is e to the power minus x square h i h j. These are the Hermit polynomials dx.
It is j factorial square root 2 pi. So, this is gamma, gamma j and delta i j. Weight and f c
is a weight, is a root of h j and w j will be this object. No, sorry, weight of roots h n.
So, this is again, Python will give you, you do not need to compute by yourself.
So, let us compute this function. This is infinity. So, I need 3 points,
whether this is default, this is the weight function. So, we need to worry about this
polynomial and we need the 3 points. With n equal to 3, the answer is this and which is exact.
Of course, in computer, it will compute 1 root pi. So, you will not exactly get 3 by 4 root pi,
but you get the value of square root pi and check, you will find the answer to be matching
16 digits to precision. So, computer will have error, of course, with round off errors.
That we cannot avoid. I cannot get square root pi with a numerical computer.
But let me also make a remark that this function, this is a module for SimPy. I am not covering in
this course. So, SimPy is a symbolic Python, which is like Mathematica. So, with Mathematica,
we are treating. So, there you can get this kind of thing. So, it will do it analytically.
It is symbolic Python. So, x is a, this will be a symbol. So, it can be done.
You can read up, it is very useful module. But SimPy, I would not cover.
Quantum oscillator, we can again do exactly the thing like what I did for hydrogen atom.
So, I am not doing it right now, but you can just reproduce yourself.
You can compute expectation value of x squared, you will get as n plus half by 2. Of course,
h bar omega is a dimension, which I am not writing here, but non-dimensionalized h bar omega is
already taken care of by scaling. Scanning energy is same as that and
total energy is sum of the two, n plus half h bar omega.
