PHY 314: Introduction To Quantum Mechanics, Varsha 2014

Download as pdf or txt
Download as pdf or txt
You are on page 1of 3

PHY 314: Introduction to Quantum Mechanics, Varsha 2014

Lecture 7
Anil Shaji
School of Physics, IISER - TVM
(Dated: 27-08-2014)
The Schr
odinger equation. The Hamiltonian and the idea of generators. Finite time
evolution and unitary transformations. Properties of unitary transformations.

The centerpiece of classical mechanics is Newtons second law,


F = ma,
which is a cause and effect relation that lets us compute what a particle will do at all times given the
causes, namely forces, that change its state of motion. The description of a particle (kinematics)
in classical physics is in terms of its position, x. The dynamical law relates the second derivative
of the position to the forces acting on the particle allowing us to find the position of the particle,
x(t), as a function of time.
The description of a quantum state is in terms of its state vector and we need a dynamical
law that governs the time evolution of the state vector. This will let us compute what happens to
the state of the quantum system in the future. The dynamical law in quantum mechanics is the
Schrodinger equation:

i~

|(t)i
= H(t)|(t)i
dt

We have already seen that operators acting on state vectors can make changes to the state. Here
we encounter a special operator, H(t) which is called the Hamiltonian operator. The Hamiltonian
operator - or just Hamiltonian for brevity - is identified as the generator of translations in time.
As part of trying to understand what the Hamiltonian is and what it does, let us see how we can
get to the Schr
odinger equation starting from things we know.

I.

FINITE TIME EVOLUTION AND UNITARY OPERATORS

If we know the state, |(t)i of a system at time t, what we want to do is to find the state of the
system, |(t0 )i at a later time, t0 . From what we know already about the Hilbert space of states
of a quantum system and the operators that act on the states, we expect that we will be able to
represent this transformation of the state as the action of an operator,
|(t0 )i = U (t0 , t)|(t)i.
The operator, U (t, t0 ) is called the time evolution operator for obvious reasons.

2
The most important property that we require of this operator is that it preserve the norm of
the state that it is acting on,
h(t)|(t)i = h(t0 )|(t0 )i.
In other words, if we write both |(t)i and |(t0 )i in some basis as
|(t)i =

ci (t)|i i,

|(t0 )i =

ci (t0 )|i i,

then we require that


X
i

|ci (t)|2 =

|ci (t0 )|2 .

This is just the statement that the sum of probabilities for all the possibilities allowed for a
quantum system add up to unity at all times. The total probability for something happening
cannot be decreased as a result of time evolution.
The condition that the normalization of the states be preserved under time evolution translates
to the following condition on the time evolution operator
U (t0 , t)U (t0 , t) = 11,
where 11 is the unit operator. Operators that satisfy this condition are called unitary operators.
Note that the time evolution operator need not be Hermitian. It just has to be unitary.
Another property that we require of the time evolution operator is the composition property:
U (t00 , t0 )U (t0 , t) = U (t00 , t)
This condition says that we can always break up the time evolution from t to t00 into two (or more)
steps. The first step taking the state from t to t0 and the second step from t0 to t00 .
We can now imagine an infinitesimal time evolution operator that takes the state forward in
time through a tiny amount dt,
|(t + dt)i = U (t + dt, t)|(t)i.
We can show that the form
U (t + dt, t) = 11 idt,
satisfies the two requirements we placed on the time evolution operators.
U (t + dt, t)U (t + dt, t) = (11 + it)(1 it) = 11,

assuming (dt)2 ' 0

and
U (t + dt + dt0 , t) = 11 idt idt0 ' (1 idt0 )(1 idt) = U (t + dt0 , t + dt)U (t + dt, t)

3
The action of the infinitesimal time evolution operator is given by
|(t + dt)i = U (t + dt, t)|(t)i = (11 idt)|(t)i.
so we have
lim

dt0

|(t + dt)i |(t)i


= i|(t)i.
dt

Identifying /~ = H we get the Schr


odinger equation,
i~

|(t)i
= H(t)|(t)i.
dt

Note that in the expression for U (t + dt, t) that we wrote down, t should be dimensionless.
So has dimensions of a frequency (1/s) and since ~ has units of action (Joule-second), H = ~
has units of energy.
Using the composition rule for the time evolution operator, one can write down an equation of
motion for the time evolution operator itself;
U (t0 + dt, t) = U (t0 + dt, t0 )U (t0 , t) = (1 iH(t) dt/~)U (t0 t)
and so

U (t0 + dt, t) U (t0 , t)


= i~ U (t0 , t) = H(t)U (t0 t).
dt0
dt
t

i~ lim

Let us now come back to the finite time evolution operator. If the Hamiltonian does not depend
on time then we can solve the above equation as
0

U (t0 t) = eiH(t t)/~


If H is dependent on time, then the formal solution is


i
U (t , t) = exp
~
0

Z
t

t0


dt H(t ) .
00

00

You might also like