Consider a spinless particle of mass m in a central potential V(r).

  The potential energy

of the particle depends only on its distance from the origin.

For a classical particle in a central potential the force  is always directed

towards the origin, the torque t=r´F  is zero, and the angular momentum L=r´p is a
constant of motion.  The particle's trajectory lies in a plane passing through the origin. 
Its velocity v may be decomposed into a component vr along and a component v^
perpendicular to in this plane.  The energy of the particle is given by

Using p=mv and L=mrv^ we can write

, 

with

To make the transition to quantum mechanics we write down the quantum-mechanical

Hamiltonian H of the particle by replacing in the expression for E the classical physical
quantities expressed in terms of r and p by their corresponding quantum mechanical
operators, symmetrizing suitably, if necessary.  In coordinate representations we have
already found expressions for the operators px, py, pz, p2, and L2, but we have not yet
found an expression for pr2.

In coordinate representation we write

                                          

See notes!

.
The expression for pr2 in coordinate representation therefore is

To find the wave functions y(r,q,f) of the eigenstates of H we have to solve the
eigenvalue equation Hy(r)=Ey(r).  For a particle in a central potential the entire
angular dependence of H is contained in the L2 term.  L2 commutes with all
components Li of L.  All other terms of H depend only on r, not on q and f, and
commute with Li.  We have [H,Li]=0, [H,L2]=0.  The angular momentum L of the
particle is therefore a constant of motion.  We can find a common eigenbasis of H, L2
and Lz.  We denote the states of this basis by |k,l,m> and the corresponding
eigenfunctions by yklm(r,q,f).  We have

We already know the general form of yklm(r,q,f).  yklm(r,q,f)=Rkl(r)Ylm(q,f) is a product

of a radial function Rkl(r) and the spherical harmonic Ylm(q,f).  (See notes!)

Substituting this form of yklm(r,q,f) into the eigenvalue equation yields the differential
equation

for Rkl(r).

Note: Since H commutes with L± , E cannot depend on m.  E is therefore at least (2l+1)
fold degenerate.  Accidental degeneracies may add to this essential degeneracy.

Writing we find a differential equation for ukl(r), which has a simpler

form than the differential equation for Rkl(r).

This differential equation has the same form as the one-dimensional Schroedinger
equation for a particle moving in an effective potential

(Here is one reason why we have spend so much time on 1-d problems.)

But r is not x.  The variable r can take on only non-negative real values.  R(r) has to
stay finite at r=0, therefore u(r) has to go to zero at r=0.  Otherwise y(r,q,f) is not an
acceptable wave function.  Choosing Veff(r)=¥ for r<0 lets us treat the radial equation
as an ordinary one-dimensional equation while guaranteeing the proper behavior at the
origin.

The asymptotic behavior of Rkl(r)

Assume Rkl(r) is sufficiently regular that it may be expanded in powers of r near the
origin.

where s is the lowest power with a non-zero coefficient.

Therefore

Substituting this expression for ukl(r) into

yields

+ (terms containing powers of r greater than s-1) = 0, if |V(r)|£ C'r-2 as r®0.

The coefficients of successive powers of r have to vanish independently, since this

equation holds for all r.  Letting the coefficient of rs-1 vanish we obtain -s(s+1)+l(l+1)=0
if |V(r)|< C'r-2 as r®0, i.e. if |V(r)| does not blow up as fast as r-2 as r®0.  This is a
quadratic equation with two solutions.  It implies that s=l, or that s=-(l+1) if |V(r)|< r-2
as r®0.  As r®0 the lowest power of r dominates and ukl(r)®Crl+1.  The solution
ukl(r)®Crl+1 goes to zero at r=0, but the solution ukl(r)®Cr-l blows up at r=0 and is
unacceptable.

Near the origin the radial behavior of an acceptable wave function of a particle
moving in a central potential is proportional to rl, (if |V(r)|<C'r-2). Note: R(r)=u(r)/r.

Two interacting particles

Consider an isolated system of two interacting spinless particles of masses m1 and m2

at positions r1 and r2, respectively.  The potential energy of the particles depends only
on their relative position r1-r2.  In classical mechanics the Lagrangian of such a particle
can be written as
 ,

where M=m1+m2 is the total mass of the system, 

is the reduced mass of the system, 

are the coordinates of the center of mass of the system, 

and r=r1-r2 are the relative coordinates.  

Using  we define  .  The classical equations of motion are

obtained from

They are

If we take P, R and p, r as the new canonical variables of the system then the
Hamiltonian of the system is

The first term represents the kinetic energy associated with the motion of the center
of mass, which is in uniform rectilinear motion ( ).  The other terms represent
the energy associated with the relative motion.  If we choose an inertial frame in
which the center of mass is at rest, then  

is the Hamiltonian of a fictitious particle with mass m, whose position is given by the
relative coordinate r, moving in an external potential V(r).  The relative motion of
two interacting spinless particles reduces to the motion of a single fictitious particle
in an external potential.

To make the transition to quantum mechanics we again replace in the expression for H
the classical physical quantities expressed in terms of the canonical variables by their
corresponding quantum mechanical operators and symmetrize suitably, if necessary.

where H, P, p, and r are now operators.  Expressing these operators in terms of p1, p2,
r1, and r2 and using we can show that
and that for all i and j.

We therefore have

There exists a common eigenbasis of H, HCM, and Hr.  We may consider the state space
E to be the tensor product of ECM and Er, E=ECMÄEr.  P and R operate only in ECM and p
and r operate only in Er.  The motion of the center of mass and the relative motion are
completely independent of each other.

Let {|c>} be an eigenbasis of HCM and {|w>} be an eigenbasis of Hr.  Then

{|f>=|c>Ä|w>}is an eigenbasis of H in E.

In coordinate representation we write

We can solve for c(R).

The fictitious particle associated with the motion of the center of mass behaves like a
free particle and is represented by a plane wave.

We are often only interested in the relative motion, i.e. in the behavior of the two
interacting particles in their center of mass frame.  If their mutual interaction depends
only on the distance between them, then V(r)=V(r) and the eigenvalue equation
becomes

Then

.
 .

is the radial function of a particle of mass m moving in a central potential.

Problem:

Assume that the potential energy of the deuteron is given by V(r)=-V0, r<r0; V(r)=0
r³r0.

(a) Show that the ground state of the deuteron possesses zero orbital angular
momentum (l=0).  Since this is true for any central potential, you may not need the
detailed nature of the square well potential.

(b) Assume that l=0 and estimate the value of V0 under the additional condition that
the value of the binding energy is much smaller than V0.

Solution:

(a)  The Hamiltonian of the relative motion of the two particles is

Let yl be the lowest energy eigenfunction of H with orbital angular momentum

quantum number l.

 .

Let yl+1 be the lowest energy eigenfunction of H with orbital angular momentum
quantum number l+1.

since yl is assumed to be the lowest energy eigenfunction of the Hamiltonian

.

Therefore and the ground state possesses zero orbital angular

momentum.

(b)  The relative motion of the two particles is described in the same way as the
motion of a fictitious particle of reduced mass m in a central potential. 
Therefore we have

The problem is reduced to a one dimensional "square well problem" with E<0.

Let

Let r<r0 define region 1 and r>r0 define region 2.  The coordinate r is never
negative.  Therefore

The boundary conditions are

in regions where .

If we plot we find solutions at the intersections of the

two curves in regions where , i.e. .

For only one solution to exist we need .

(See QM1, homework 2, problem 3, solutions.)

If |E|<<|V0| then .  Then .

(Rule of thumb: To estimate the range of a force, divide by the mass of the
particle that carries it times c, ).

See notes!