PH-201 Modern Physics 22 November, 2011

Solving the Angular part of the Schrodinger Equation

The time independent Schrodinger Equation is,

~2 2
− ∇ ψ(r, θ, ϕ) + U (r)ψ(r, θ, ϕ) ≡ Eψ(r, θ, ϕ).
2µ

We assume a solution of the form,

ψ(r, θ, ϕ) ≡ R(r) · Θ(θ) · Φ(ϕ).

Using the definition of the Laplacian operator in spherical coordinates (look up the con-
version in Appendix M),

[ ( ) ( ) ]
1 ∂ 2 ∂ 1 ∂ ∂ 1 ∂
∇2 = r + · sin θ + ·
r2 ∂r ∂r r2 sin θ ∂θ ∂θ r2 sin2 θ ∂ϕ
the Schrodinger equation becomes,

[ ( ) ( ) ]
~2 1 ∂ 2 ∂ 1 ∂ ∂ 1 ∂2
− r RΘΦ + sin θ RΘΦ + RΘΦ + U (r)RΘΦ ≡ ERΘΦ
2µ r2 ∂r ∂r r2 sin θ ∂θ ∂θ r2 sin2 θ ∂ϕ2
[ ( ) ( ( )) ( 2 )]
~2 1 ∂ ∂R RΦ ∂ ∂Θ RΘ ∂
− ΘΦ{ 2 r2 }+ 2 sin θ + 2 2 Φ + U (r)RΘΦ ≡ ERΘΦ
2µ r ∂r ∂r r sin θ ∂θ ∂θ r sin θ ∂ϕ2
[ ( ) ( ) ]
~2 1 d 2 dR 1 d dΘ 1 d2 Φ
− r + sin θ + + U (r) ≡ E.
2µ Rr2 dr dr Θr2 sin θ dθ dθ Φr2 sin2 θ dϕ2

−2µr 2 sin2 θ
Multiplying both sides by ~2 , we obtain,

( ) ( ) [ ]
1 d2 Φ sin2 θ d 2 dR sin θ d dΘ 2µr2 sin2 θ
≡ − r − sin θ + U (r) − E ,
Φ dϕ2 R dr dr Θ dθ dθ ~2

where the equation has been adjusted such that the R.H.S is dependent on r and θ, while
L.H.S is dependent on ϕ only. Since both sides are equal, they must equal a common separa-
tion constant, say −ml 2 , yielding,

1 d2 Φ
= −ml 2 (1)
Φ dϕ2
( ) ( ) [ ]
sin2 θ d 2 dR sin θ d dΘ 2µr2 sin2 θ
− r − sin θ + U (r) − E = −ml 2 (2)
R dr dr Θ dθ dθ ~2

Now consider equation (2),

( ) ( ) [ ]
sin2 θ d dR sin θ d dΘ 2µr2 sin2 θ
− r2 − sin θ + U (r) − E = −ml 2
R dr dr Θ dθ dθ ~2

1
Dividing the equation by sin2 θ, we obtain,

( ) ( ) [ ]
1 d 2 dR 1 d dΘ 2µr2 ml 2
− r − sin θ + 2 U (r) − E = − 2
R dr dr Θ sin θ dθ dθ ~ sin θ
( ) ( ) [ ]
ml 2 1 d dΘ 1 d dR 2µr 2
− sin θ = r 2
− 2 U (r) − E .
sin2 θ Θ sin θ dθ dθ R dr dr ~

This time, the L.H.S is dependent on θ only, while R.H.S is dependent on r only. Put both
sides equal to a common separation constant, say l(l + 1). Thus we obtain two additional
equations corresponding to the polar and radial parts,

( )
ml 2 1 d dΘ
− sin θ = l(l + 1) (3)
sin2 θ Θ sin θ dθ dθ
( ) [ ]
1 d dR 2µr2
r2 − 2 U (r) − E = l(l + 1) . (4)
R dr dr ~

(1), (3) and (4) are respectively, the azimuthal, polar and radial equations. Notice that the
d d d
partial derivatives have been replaced by ordinary derivatives dr , dθ , dϕ etc.

Solving the Azimuthal equation (1) :

Equation (1) can be written as,

d2 Φ
+ ml 2 Φ = 0.
dϕ2

To solve this equation we write the auxiliary equation,

D 2 + ml 2 = 0
D = ±iml ,

Hence,

Φ(Φ) = eiml ϕ

Actually there are two solutions, e+iml ϕ and e−iml ϕ , but we’ll cover the latter by allowing
ml to run negative. Here we consider the constraint that the eigenfunctions must be single
valued. Since the azimuthal angles ϕ = 0 and ϕ = 2π are the same angles (because when ϕ
advances by 2π, we return to the same point in space), it is natural to require that,

Φ(ϕ) = Φ(ϕ + 2π)

eiml ϕ = eiml (ϕ+2π)
= eiml ϕ · ei2ml π
1 = ei2ml π
1 = cos(2ml π) + i sin(2ml π)

2
From the last expression it follows that ml must be an integer, i.e.,

ml = 0, ±1, ±2, ±3, ±4, . . . .

Solving the Polar equation (3) :

( )
ml 2 1 d dΘ
− + · sin θ = −l(l + 1)
sin2 θ sin θ · Θ dθ dθ

Multiplying throughout by Θ,

( )
ml 2 Θ 1 d dΘ
− + · sin θ = −l(l + 1)Θ
sin2 θ sin θ dθ dθ
( ) [ ]
1 d dΘ ml 2
· sin θ + l(l + 1) − Θ = 0.
sin θ dθ dθ sin2 θ

Now, multiplying by sin2 θ, we obtain,

2
[ ]
dΘ 2 d Θ
sin θ · cos θ + sin θ 2 + sin θ · l(l + 1) − ml Θ = 0.
2 2
(5)
dθ dθ

We perform the substitution,

cos θ = z.
⇒ sin θ 2
= 1 − z2
√
sin θ = 1 − z2.
dz
= − sin θ.
dθ

Now if Θ(θ) = P (z), P being a function of z, we can use the chain rule to compute and
express Θ and θ in terms of P and z,

dΘ dΘ dz
= ·
dθ dz dθ
dP
= − sin θ
dz
dP √
= − 1 − z2
dz
dP 1 dΘ
⇒ = − · .
dz sin θ dθ

Furthermore,

3
 ( )
d2 Θ d dΘ
=
dθ2 dθ dθ
( )
d dP
= − sin θ
dθ dz
( )
d dP dz
= − sin θ ·
dz dz dθ
( )
d dP
= − sin θ · (− sin θ)
dz dz
( 2 )
d P dP d
= sin θ sin θ + · sin θ
dz 2 dz dz
d2 P dP d
= sin2 θ 2 + sin θ · · sin θ
dz dz dz

Since,
√
sin θ = 1 − z2
( )
d d √
⇒ sin θ = 1 − z2
dz dz
z
= −√ .
1 − z2

d2 Θ
Thus dθ2 becomes,

d2 Θ d2 P √ dP z
= (1 − z 2 ) + 1 − z2 · · (− √ )
dθ2 dz 2 dz 1 − z2
d2 P dP
= (1 − z 2 ) 2 − z .
dz dz

As a result the polar equation (5) assumes the form,

√ ( ) [ ] [ ]
dP √ 2
2 d P dP
1−z ·z −
2 1 − z + (1 − z ) (1 − z ) 2 − z ·
2 2
+ (1 − z )l(l + 1) − ml P (z) = 0
2 2
dz dz dz
[ ]
dP d2 P dP
−(1 − z 2 ) · z + (1 − z 2 )2 2 − (1 − z 2 )z · + (1 − z 2 )l(l + 1) − ml 2 P (z) = 0
dz dz dz
2
[ ]
d P dP
(1 − z 2 )2 2 − 2(1 − z 2 ) · z + (1 − z 2 )l(l + 1) − ml 2 P (z) = 0.
dz dz

Divide by (1 − z 2 ) to obtain,

[ ]
d2 P dP m2l
(1 − z 2 ) − 2z + l(l + 1) − P = 0.
dz 2 dz (1 − z 2 )

This is the historically famous Legendre equation. This equation may also be written in the
compact form,

[ ] [ ]
d dP ml 2
(1 − z 2 ) · + l(l + 1) − P (z) = 0 . (6)
dz dz (1 − z 2 )

4
This ordinary differential equation is frequently encountered in physics and other technical
fields. In particular it occurs when solving Laplace’s equation and related partial differential
equations in spherical coordinates. The Legendre equation may be solved using the standard
power series method. Let us perform (yet another) substitution,

P (z) = (1 − z 2 )|ml |/2 G(z),

dP |ml | |ml | |ml | ′
⇒ = (1 − z 2 )( 2 −1) (−2z)G(z) + (1 − z 2 ) 2 G (z)
dz 2
|ml | |ml | ′
= −|ml |z(1 − z 2 )( 2 −1) G(z) + (1 − z 2 ) 2 G (z).

Multiplying both sides by (1 − z 2 ), we get,

dP |ml | |ml | ′
(1 − z 2 ) = −|ml |z(1 − z 2 ) 2 G(z) + (1 − z 2 ) 2 +1 G (z).
dz

Now differentiate both sides with respect to z to obtain,

[ ] [ ]
d dP d |ml | |ml | ′
(1 − z 2 ) = −|ml |z(1 − z 2 ) 2 G(z) + (1 − z 2 ) 2 +1 G (z)
dz dz dz
|ml | |ml | |ml | |ml |
= −|ml |(1 − z 2 ) 2 G(z) − |ml |z(1 − z 2 ) 2 −1 (−2z)G(z) − |ml |z(1 − z 2 ) 2 G′ (z)
( ) 2
|ml | |ml | ′ |ml | ′′
+ + 1 (1 − z ) 2 (−2z)G (z) + (1 − z ) 2 +1 G (z)
2 2
2
|ml | |ml | |ml |
= −|ml |(1 − z 2 ) 2 G(z) − |ml |2 z 2 (1 − z 2 ) 2 −1 G(z) − |ml |z(1 − z 2 ) 2 G′ (z)
( )
|ml | |ml | ′ |ml | ′′
− 2z + 1 (1 − z 2 ) 2 G (z) + (1 − z 2 ) 2 +1 G (z).
2
[ ]
Inserting the value of P (z) and d
dz (1 − z 2 ) dP
dz back into equation (6) we obtain,

[
|ml | |ml | |ml |
−|ml |(1 − z 2 ) 2 G(z) − |ml |2 z 2 (1 − z 2 ) 2 −1 G(z) − |ml |z(1 − z 2 ) 2 G′ (z)
( ) ] [ 2 ]
|ml | |ml | ′ |ml |
+1 ′′ |ml |
−2z + 1 (1 − z )
2 2 G (z) + (1 − z ) 2 2 G (z) + l(l + 1) − (1 − z 2 )|ml |/2 G(z) = 0
2 (1 − z 2 )
[ ( ) ]
|ml |
+1 ′′
|ml | |ml | |ml |
(1 − z ) 2 2 G (z) + −|ml |z(1 − z )2 2 − 2z + 1 (1 − z ) 2 2 G′ (z)
2
[ ( 2 ) ]
|ml | |ml | |ml |
+ −|ml |(1 − z 2 ) 2 + |ml |2 z 2 (1 − z 2 ) 2 −1 + l(l + 1) − (1 − z 2 |ml |/2
) G(z) = 0
(1 − z 2 )

ml
Dividing both sides by (1 − z 2 ) 2 ,

[ ( )] [ { 2 }]
′′ |ml | |ml |2 z 2 |ml |
(1 − z 2 )G (z) + −|ml |z − 2z + 1 G′ (z) + −|ml | + + l(l + 1) − G(z) = 0
2 (1 − z 2 ) (1 − z 2 )
[ ] [ 2 ]
′′
′ |ml |
(1 − z )G (z) + −2|ml |z − 2z G (z) + l(l + 1) − |ml | −
2
(z − 1) G(z) =
2
0
(1 − z 2 )

5
 [ ]
′′
(1 − z 2 )G (z) − 2z(|ml | + 1)G′ (z) + l(l + 1) − |ml | − |ml |2 G(z) = 0. (7)

This is the equation we set to solve. We have to find a G(z) that solves (7) and back sub-
stitute it to find P (z) and ultimately, Θ(θ). Suppose G(z) is a power series (a polynomial in z),

∞
∑
G(z) = an z n
n=0
∑∞
′
G (z) = nan z n−1
n=0
∑∞
′′
G (z) = n(n − 1)an z n−2 .
n=0

Substitution of these values in Equation 7 yields,

∞
∑ ∞
∑ ∞
∑
(1 − z 2 ) n(n − 1)an z n−2 − 2z(|ml | + 1) nan z n−1 + [l(l + 1) − |ml | − |ml |2 ] an z n = 0
n=0 n=0 n=0
∞
∑ ∑∞ ∑∞
(1 − z 2 ) n(n − 1)an z n−2 − 2(|ml | + 1) nan z n + [l(l + 1) − |ml | − |ml |2 ] an z n = 0
n=0 n=0 n=0
(1 − z 2 )(2a2 + 3 · 2 · a3 z + 4 · 3 · a4 z 2 + 5 · 4 · a5 z 3 + . . .) − 2(|ml | + 1)(a1 z + 2a2 z 2 + . . .)
+[l(l + 1) − |ml | − |ml |2 ](ao + a1 z + a2 z 2 + . . .) = 0

(2a2 + 3 · 2 · a3 z + 4 · 3 · a4 z 2 + 5 · 4 · a5 z 3 + . . .) − (2a2 z 2 + 3 · 2 · a3 z 3 + 4 · 3 · a4 z 4 + 5 · 4 · a5 z 5 + . . .)
−2(|ml | + 1)(a1 z + 2a2 z 2 + . . .) + [l(l + 1) − |ml | − |ml |2 ](ao + a1 z + a2 z 2 + . . .) = 0. (8)
We notice that,

∞
∑
(2a2 + 3 · 2 · a3 z + 4 · 3 · a4 z + 5 · 4 · a5 z + . . .)
2 3
= (n + 2)(n + 1)an+2 z n
n=0
∑∞
(2a2 z 2 + 3 · 2 · a3 z 3 + 4 · 3 · a4 z 4 + 5 · 4 · a5 z 5 + . . .) = n(n − 1)an z n
n=0
∑∞
(a1 z + 2a2 z 2 + . . .) = nan z n
n=0
∑∞
(ao + a1 z + a2 z 2 + . . .) = an z n .
n=0

Therefore, Equation (8) can be written compactly as,

∞
∑ ∞
∑ ∞
∑ [ ]∑
∞
(n + 2)(n + 1)an+2 z n − n(n − 1)an z n − 2(|ml | + 1) nan z n + l(l + 1) − |ml | − |ml |2 an z n = 0.
n=0 n=0 n=0 n=0
n
The coefficients of each z must satisfy the algebraic equation,

[ ]
(n + 2)(n + 1)an+2 − n(n − 1)an − 2(|ml | + 1)nan + l(l + 1) − |ml | − |ml | an = 0, 2

6
resulting in the recursive formula,

[ ]
n(n − 1) + 2n(|ml | + 1) − [l(l + 1) − |ml | − |ml |2 ]
an+2 = an
(n + 2)(n + 1)
[ 2 ]
n − n + 2n|ml | + 2n − l(l + 1) + |ml | + |ml |2
= an
(n + 2)(n + 1)
[ 2 ]
n + n + 2n|ml | − l(l + 1) + |ml | + |ml |2
= an
(n + 2)(n + 1)
[ 2 ]
n + n + 2n|ml | − l(l + 1) + |ml | + |ml |2
= an
(n + 2)(n + 1)
[ ]
(n + |ml |)(n + |ml | + 1) − l(l + 1)
= an
(n + 2)(n + 1)

∑∞
Where n, ml are integers. Since G(z) = n=0 an z n is an infinite series, and z = cos θ, the
entire range of z is (−1 ≤ z ≤ 1) including z ± 1. When z assumes the value ±1, the series
diverges and G(z) becomes infinite. This is not physically possible, because then the wave
functions Θ(θ) blows up. In other words, we cannot let an ’s go on up to a∞ , otherwise the
series G will diverge. A value for n = 0, 1, 2, 3, . . . reaches, (let’s call it nmax ) for which,

(nmax + |ml |)(nmax + |ml | + 1) = l(l + 1)

Since n and ml are integers, so l is also an integer. There always exists an nmax , such that,
l = nmax + |ml |. The possible values of l are (|ml |), (|ml | + 1), (|ml | + 2), ...., (|ml | + nmax ),
ensuring that the series converges. The constraint l = nmax +|ml | also indicates that |ml | ≤ l.

Let us summarize our main results so far.

l = nmax + |ml | (9)

Don’t confuse nmax with the principal quantum number n.

n∑
max

G(z) = an z n (10)
n=0

P (z) = (1 − z 2 )|ml |/2 G(z) (11)

Θ(θ) = P (z) (12)
[ ]
(n + |ml |)(n + |ml | + 1) − l(l + 1)
an+2 = an (13)
(n + 2)(n + 1)

Example:

Case I: When l = 0, ml = 0

If l = 0 and ml = 0, then from equation (9), nmax = 0.

Thus G(z) will be,

G(z) = ao
P (z) = (1 − z 2 )0 ao = a0
Θ(θ) = P (z) ∝ 1

7
Case II: When l = 1, |ml | = 0, 1, ⇒ ml = 0, ±1

(a) Consider l = 1, and ml = 0, then from equation (9), nmax = 1.

G(z) = ao + a1 z

From equation (9) it is clear that, a2 = 0, because nmax = 1 and ao = 0, thus

G(z) = a1 z
P (z) = (1 − z 2 )0 a1 z = a1 z
Θ(θ) ∝ z = cos θ

(b) Consider l = 1, and ml = 1, then from equation (9), nmax = 0.

G(z) = ao
P (z) = (1 − z 2 )1/2 ao = ao (1 − z 2 )1/2
Θ(θ) ∝ (1 − z 2 )1/2 = sin θ

(c) Consider l = 1, and ml = −1, then from equation (9), nmax = 2.

G(z) = ao + a1 z + a2 z 2

From equation (9), a2 = −ao ,

G(z) = ao + a1 z − ao z 2
= ao (1 − z 2 ) + a1 z

Since nmax = 2, so a3 = 0, ⇒ a1 = 0.

G(z) = ao (1 − z 2 )
P (z) = (1 − z 2 )−1/2 ao (1 − z 2 )
= a0 (1 − z 2 )1/2
Θ(θ) ∝ (1 − z 2 )1/2 = sin θ

Case III: When l = 2, |ml | = 0, 1, 2, ⇒ ml = 0, ±1, ±2

(a) Consider l = 2, and ml = 0, then from equation (9), nmax = 2.

G(z) = ao + a1 z + a2 z 2

Where a2 = −3ao . Also a3 = 0, ⇒ a1 = 0.

G(z) = ao − 3ao z 2
= ao (1 − 3z 2 )
P (z) = (1 − z 2 )0/2 ao (1 − 3z 2 )
= ao (1 − 3z 2 )
Θ(θ) ∝ (1 − 3z 2 ) = (1 − 3 cos2 θ)

(b) Consider l = 2, and ml = 1, then from equation (9), nmax = 1.

G(z) = ao + a1 z

8
Where a2 = −2ao . Since due to nmax = 1, we have a2 = 0, ⇒ ao = 0.

G(z) = a1 z
P (z) = a1 z(1 − z 2 )1/2
Θ(θ) ∝ z(1 − z 2 )1/2 = sin θ cos θ.

(c) Consider l = 2, and ml = 2, then from equation (9), nmax = 0.

G(z) = ao
P (z) = ao (1 − z 2 )
Θ(θ) ∝ (1 − z 2 ) = sin2 θ.

Case IV: When l = 3, |ml | = 0, 1, 2, 3, ⇒ ml = 0, ±1, ±2, ±3

(a) Consider l = 3, and ml = 0, then from equation (9), nmax = 3.

G(z) = ao + a1 z + a2 z 2 + a3 z 3

Where a2 = −6ao and a3 = − 53 a1 . Since ao = 0 otherwise series diverges, hence a2 = 0.

Therefore,
5
G(z) = a1 z − a1 z 3
3
= a1 (3z − 5z 3 )
P (z) = a1 (3z − 5z 3 )
Θ(θ) ∝ (3z − 5z 3 ) = z(3 − 5z 2 ) = cos θ(3 − 5 cos2 θ)

(b) Consider l = 3, and ml = 1, then from equation (9), nmax = 2.

G(z) = ao + a1 z + a2 z 2

Where a2 = −5ao and a3 = −a1 . Since nmax = 2, this gives a3 = 0 ⇒ a1 = 0. Hence,

G(z) = a0 − 5ao z 2
= ao (1 − 5z 2 )
P (z) = ao (1 − 5z 2 )(1 − z 2 )1/2
Θ(θ) ∝ (1 − 5z 2 )(1 − z 2 )1/2 = (1 − 5 cos2 θ) sin θ

(c) Consider l = 3, and ml = 2, ⇒ nmax = 1.

G(z) = ao + a1 z

Where a2 = −2ao . Also due to nmax = 1, a2 = 0, which implies ao = 0.

G(z) = a1 z
P (z) = a1 z(1 − z 2 )
Θ(θ) ∝ z(1 − z 2 ) = cos θ sin2 θ

(d) Consider l = 3, and ml = 3, ⇒ nmax = 0.

G(z) = ao
P (z) = ao (1 − z 2 )3/2
Θ(θ) ∝ (1 − z 2 )3/2 = sin3 θ