University of Connecticut

OpenCommons@UConn
Chemistry Education Materials Department of Chemistry

October 2007

Introductory Mathematics for Quantum

Chemistry
Carl W. David
University of Connecticut, Carl.David@uconn.edu

Follow this and additional works at: https://opencommons.uconn.edu/chem_educ

Recommended Citation
David, Carl W., "Introductory Mathematics for Quantum Chemistry" (2007). Chemistry Education Materials. 50.
https://opencommons.uconn.edu/chem_educ/50
 Introductory Mathematics for Quantum Chemistry
C. W. David
Department of Chemistry
University of Connecticut
Storrs, Connecticut 06269-3060
(Dated: October 9, 2007)

I. SYNOPSIS ture in equations dominates. Dirac introduced a notation

which eliminates the ever-redundant ψ from discourse.
Since much of the mathematics needed for quantum He defined a wave function as a ket.
chemistry is not covered in the first two years of calculus, The ket is defined as ‘|index >≡ ψ’, where ‘index’ is
a short introduction to those methods is presented here. some human chosen index or indicator whose purpose is
Elegance has been eschewed in the Einstein spirit. to label the state (or the function) under discussion. For
a particle in a box, a potential ket is
r
II. ORTHOGONAL FUNCTIONS  nπx  2  nπx 
ψn (x) = |n >≡ Nn sin = sin (2.1)
L L L
A. Dirac Notation
where the domain is 0 ≤ x ≤ L, Nn is a normalization
The notation employed in beginning discussions of constant, and ‘n’ is a quantum number, i.e., an index
quantum chemistry becomes cumbersome when the com- chosen by us.
plexity of problems increase and the need to see struc- For the hydrogen atom, an appropriate ket might be

ψn=3,`=1,m` =−1 (ρ, ϑ, φ) = |3, 1, −1 >= N3,1,1 ρ(4 − ρ)e−ρ sin ϑe−ıφ (2.2)

which is a 3p electron’s state (ρ) is a dimensionless radius.
A harmonic oscillator might have an appropriate ket
of the form
2 2
ψ3 (x) = |3 >= N3 8β 3 x3 − 12βx e−β x /2


(2.3)
i.e., the ket notation allows us to indicate inside the
funny delimiters the essential characteristics of the in-
volved wavefunction without extraneous elements which
might distract attention from the overall viewpoint being
propounded!
B. The Dot Product
Returning to the hydrogen atom, the following ‘vector’
(i.e., the normalization equation, required so that the
probability of finding the electron somewhere is 1, cer-
tain(!)), which itself is shorthand for the complicated
three-dimensional integral, usually display is spherical
polar coördinates.
One sees that constructing the ‘bra’‘ket’ [2] consists of
two separate concepts, making the ‘bra’ from the ‘ket’
by using the complex conjugate, juxtaposing the two in
the proper order, and integrating over the appropriate
domain. The analogy of this process is the dot product
of elementary vector calculus, and it is here where we get
our mental images of the processes described.
In three space, x, y, and z, we have unit vectors î, ĵ, k̂,
so that an arbitrary vector as:
R~1 = X î + Y ĵ + Z k̂

< 3, 1, −1| = N3,1,1 ρ(4 − ρ)e−ρ sin ϑe+ıφ (2.4) could be re-written as

which is the complex conjugate of the original wave func- |R1 >= X|i > +Y |j > +Z|k >
tion [1] , is known as a ‘bra’ vector, and the notation In this space, one doesn’t integrate over all space, one
< 3, 1, −1|3, 1, −1 >= 1 (2.5) adds up over all components, thus

is a shorthand for < R1 |R1 >= R~1 · R~1 = X 2 + Y 2 + Z 2

Z
∗ since |i > is orthogonal (“perpendicular”‘) to |j > and
ψ3,1,−1 ψ3,1,−1 dτ = 1 (2.6)
all space |k >, etc..

Typeset by REVTEX
 2

In the same sense, define a truncated expansion:

r i=m r
X 2  nπx  X 2 iπx
|arb >= cn sin (2.7) Sm = ci sin (3.4)
L L i=0
L L
n
where, contrary to the particle in a box, we have em-
where |arb > is an arbitrary state. This is our normal
ployed the i=0 particle in a box basis function. Now, we
entry point for Fourier Series.
form an error at the point x:
Before we go there, let us re-introduce orthogonality,
i.e., |f unction > −Sm (x) = err(x) (3.5)
Z and notice that if we added up this error for every value
< n|m >= 0 = ψn∗ ψm dτ ; m 6= n (2.8) of x in the domain 0 ≤ x ≤ L we would be approaching
domain
the concept we desire, a measure of the error, the dif-
where the domain is problem specific, as is the volume ference between the function and its approximation, the
element dτ . This orthogonality is the analog of the idea truncated series.
that But, you say, sometimes the error (err(x)) is positive,
and sometimes its negative, depending on circumstances.
î · ĵ = 0 We want a measure which counts either effect properly,
and the answer is to define a new error
as an example in normal vector calculus. 2
(|f unction > −Sm (x)) = ERR(x) ≡ [err(x)]2 (3.6)
where clearly, the l.h.s. is positive definite, and therefore
III. FOURIER SERIES can not give rise to fortuitous cancelation.

A. Particle in a Box vis-a-vis Fourier Series |function>

Sm(x)
The normal introduction to orthogonal functions is via
Fourier Series, but in the context of quantum chemistry,
we can re-phrase that into the Particle in a Box solution
to the Schrödinger Equation, i.e.,

h̄2 ∂ 2 ψ
− + 0ψ = Eψ (3.1)
2m ∂x2 ERR(x)
in the box (domain) 0 ≤ x ≤ L. The solutions are known
to be
r
2 nπx
ψn (x) = sin = |n > (3.2) domain
L L
0 L x
where n is an integer greater than zero, and L remains
the size of the box.
Now, a Fourier Series is an expansion of a function ERR(x)
(periodic) over the same (repeated) domain as above, in
the form
r
X X 2 iπx
|f unction >= ci |i >= ci sin (3.3)
i i
L L
ERR 2(x)
(where we changed index for no particular reason from
n to i, other than to remind you that the index is a
dummy variable). The question is, what is the optimal,
best, nicest, etc., etc., etc., value for each of the ci in this
expansion? FIG. 1: Error involved in approximating a function.

So, adding up ERR(x) at each point x in the domain,

B. A Minimum Error Approximation we have
Z L Z L
2
(|f unction > −Sm (x)) dx = ERR(x)dx ≡ ERRORm
To answer this question, we look for a measure of error 0 0
between the function and the expansion. To do this we (3.7)
 3

ERRORm is now a postive definite number which mea- series at m. We re-write this explicitly to show that sum-
sures the error committed in truncating the approximate mation:

m r !2
Z L Z L
X 2 iπx
|f unction > − ci sin dx = ERR(x)dx ≡ ERRORm (3.8)
0 i=0
L L 0

and form the partial holding all other ci constant, where i = 0...j...m, which
is
∂ERRORm
(3.9)
∂cj

(Z m r !r )
L
∂ERRORm X 2 iπx 2 jπ
= −2 |f unction > − ci sin sin dx (3.10)
∂j 0 i=0
L L L L

which we set equal to zero. This means that we are searching for that value of cj which makes the ERROR an
extremum (of course, we want a minimum). This leads to the equation
Z L   Z L( X m r ! )
jπx 2 iπx jπx
|f unction > sin dx = ci sin sin dx (3.11)
0 L 0 i=0
L L L

and, since one can exchange summation and integration, one has
m Z r !
Z L   L
jπx X 2 iπx jπx
|f unction > sin dx = ci sin sin dx (3.12)
0 L i=0 0 L L L

But the right hand side of this equation simplifies because the sines are orthogonal to each other over this domain,
so the r.h.s. becomes
Z L   Z L r !
jπx 2 jπx jπx
|f unction > sin dx = cj sin sin dx (3.13)
0 L 0 L L L

since the only survivor on the r.h.s. is the i = j term. C. Completeness

RL
|f unction > sin jπx


0 L dx What if we wanted to approximate a function of the
cj = R q  (3.14)
L 2 jπx jπx form
0 L sin L sin L dx
7πx
p sin (3.17)
Multiplying top and bottom by 2/L one has L

< f unction|j > and i=7 were omitted from the summations (Equation
cj = (3.15) 3.3), i.e., the summation ran past this particular har-
< j|j > monic. It would look like:
which is fairly cute in its compactness. Of course, if |j > 6πx 8πx
is normalized (as it is in our example) then · · · + c6 sin + c8 sin + ···
L L
cj =< f unction|j > (3.16) Clearly, since each sine is
orthogonal to every other
sine, the seventh sine sin 7πxL would have no expansion
which is even more compact! in this series, i.e., each ci would be zero! We can not
 4

leave out any term in the series, i.e., it must be complete, One has, defining the up-ladder operator M +
before one can assert that one can expand any function
in terms of these sines (and cosines, if need be). Said
another way,   nπx   
(n + 1)πx  πx  nπx  πx 
M + sin → sin = cos sin +sin c
7πx 1πx 2πx L L L L L
sin 6= c0 + c1 sin + c2 sin + ···
L L L
since the one term on the r.h.s which is needed, was, by which looks like
assumption, absent!
 πx   
+ L  πx  ∂|n >
M |n >→ cos |n > + sin = |n+1 >
IV. LADDER OPERATORS L nπ L ∂x

A. Laddering in the Space of the Particle in a Box

L
The nπ takes out the constant generated by the partial
We start by remembering the trigonometric definition derivative.
of the sine of the sum of two angles, i.e., What this is saying is that there is an operator, M + ,
which has a special form involving a multipier function
sin(α + β) = sin α cos β + cos α sin α (4.1) and a partial derivative, whose total effect on an eigen-
function is to “ladder” it from n to n + 1.
We seek an operator, M + which takes a state |n > and
ladders it up to the next state, i.e., |n+1 >. Symbolically,
we define the operator through the statement  πx   
L  πx  ∂
M + = cos + sin
L nπ L ∂x
M + |n >→ |n + 1 > (4.2)

and we will require the inverse, i.e.,

written as an operator.
M − |n >→ |n − 1 > (4.3)

where the down ladder operator is lowering us from n to

n − 1.
One sees immediately that there is a lower bound to
the ladder, i.e., that one can not ladder down from the C. The DownLadder Operator
lowest state, implying

M − |lowest >= 0 (4.4) We need the down operator also, i.e.,

B. Form of the UpLadder Operator 

(n − 1)πx
  πx  nπx  πx  nπx
sin = cos sin −sin cos
L L L L L
What must the form of M + be, in order that it function
properly? Since

(n + 1)πx
  πx  nπx  nπx   πx  making use of the even and odd properties of cosines and
sin = cos sin +cos sin sines, respectively.
L L L L L

 πx   
L  πx  ∂|n >
M − |n >→ cos |n > − sin → |n − 1 >
L nπ L ∂x

 πx   
− L  πx  ∂
D. Ladder Operators in the Operator
M = cos − sin
L nπ L ∂x Representation

Stated another way, in the traditional position-

momentum language, one has
 
+ πx L πx px
M = cos − sin (4.5)
L nπ L ıh̄
 
πx L πx px
M − = cos + sin (4.6)
L nπ L ıh̄
 5

E. Ladder Operators and the Hamiltonian where we want to exchange the order on the l.h.s. so
that we can ascertain what the eigenvalue of the oper-
We need to show that this ladder operator works prop- ator M + |n > i,, i.e., we need the commutator of the
erly with the Hamiltonian, i.e., Hamiltonian with the ladder operator,

p2x
≡ Hop (4.7) [Hop , M + ] ≡ Hop M + − M + Hop
2m
and to do this we assume
Remembering that
Hop |n >= |n >

and ask what happens when we operate from the left with [px , x] = −ıh̄
M + (for example). We obtain

M + Hop |n >= M + |n > (4.8) one forms

p2x
         2 
+ πx L πx ∂ πx L πx ∂ px
[Hop , M ] = cos + sin − cos + sin (4.9)
2m L nπ L ∂x L nπ L ∂x 2m

F. Some Precursor Commutators

 πx   πx  π  πx 
First, one needs [px , cos πx
L ],
px sin = sin px − ıh̄ cos (4.14)
L L L L
πx ∂  πx   πx   ∂

[px , cos ] ≡ −ıh̄ cos − cos −ıh̄ πx


L ∂x L L ∂x Next, one needs [p2x , cos L ] (and its sine equivalent).
(4.10)
which works out to be  πx   πx   πx 
[p2x , cos ] = p2x cos − cos p2x (4.15)
L L L
πx π  πx πx πx
[px , cos ] = ıh̄ sin − cos
= px cospx
L L L L L  πx   πx 
(4.11)
= px px cos − cos p2x
or, in the most useful form for future work: L L
which we re-write, factoring, as
 πx   πx  π  πx   πx 
px cos = cos px + ıh̄ sin (4.12)
 π  πx   πx 
L L L L = px cos px + ıh̄ sin − cos p2x
L L L L
One needs the same commutator involving the appro-
which we re-write as
priate sine also:
 πx   πx  πh  πx i  πx 
π  πx   πx   πx 
− cos p2x
[px , sin ] = −ıh̄ cos = px sin −sin px = px cos L px + ıh̄ L px sin L L
L L L L L
(4.13)

 π  πx   πx   πh  πx i  πx 
= ıh̄ sin + cos px px + ıh̄ px sin − cos p2x
L L L L L L
i.e.,
π  πx   πx  π h  πx  π  πx i  πx 
= ıh̄ sin px + cos p2x + ıh̄ sin px − ıh̄ cos − cos p2x
L L L L L L L L
π  πx  π h  πx  π  πx i
= ıh̄ sin px + ıh̄ sin px − ıh̄ cos
L L L L L L
 6
 πx  π  πx   π 2  πx 
[p2x , cos ] = 2ıh̄ sin px + h̄2 cos
L L L L L

 πx  π  πx   π 2  πx   πx 
p2x cos = 2ıh̄ sin px + h̄2 cos + cos p2x
L L L L L L

Now one needs [p2x , sin πx

L ] which we re-write as
 πx   πx   πx 
[p2x , sin ] = p2x sin − sin p2x (4.16)
L L L
  πx  πh  πx i  πx 
  πx   πx  = px sin px − ıh̄ px cos − sin p2x
= px px sin − sin p2x L L L L
L L

  πx  π  πx   πx 
= px sin px − ıh̄ cos − sin p2x
L L L L

  πx  π  πx  πh  πx i  πx 
= sin px − ıh̄ cos px − ıh̄ px cos − sin p2x
L L L L L L

π  πx   h π  πx i
= −ıh̄ cos px − ıh̄ px cos
L L L L
i.e.,
π  πx   h  π   πx π  πx i
= −ıh̄ cos px − ıh̄ cos px + ıh̄ sin
L L L L L L

 πx  π  πx   π 2  πx 
[p2x , sin ] = −2ıh̄ cos px + h̄2 sin
L L L L L

 πx  π  πx   π 2  πx   πx 
p2x sin = −2ıh̄ cos px + h̄2 sin + sin p2x
L L L L L L

We had, in Equation 4.8 the following, which we now wish to reverse. We need the commutator
of the Hamiltonian with the UpLadder Operator [3].
M + Hop |n >= M + |n >

p2
  πx  
L n  πx o px
[Hop , M ] = x
+
cos − sin − M + Ho p
2m L nπ L ıh̄

which becomes
p2x  πx  p2 L n  πx o px
[Hop , M + ] = cos − x sin − M + Ho p (4.17)
2m L 2m nπ L ıh̄

expanding, and using the previously obtained commutators,

1 2  πx 
1 L
[Hop , M + ] = px cos −
n 2m  πx o L px
2m nπ
2
px sin − M + Ho p
L ıh̄
 7

  π 2  πx  
1 π  πx   πx 
[Hop , M + ] = 2ıh̄ sin px + h̄2 cos + cos p2x
2m L L L L L
  π 2  πx   p
1 L π  πx   πx 
x
− −2ıh̄ cos px − h̄2 sin + sin p2x − M + Ho p (4.18)
2m nπ L L L L L ıh̄

which expands, at first, to

  π 2  πx 
1 π  πx 
[Hop , M + ] = 2ıh̄ sin px + h̄2 cos
2m L L L L
  π 2  πx  p
1 L π  πx 
x
− −2ıh̄ cos px − h̄2 sin (4.19)
2m nπ L L L L ıh̄

  2  πx 
+ 1 π  πx 
2 π
[Hop , M ] = 2ıh̄ sin px + h̄ cos
2m L L L L
 
1 1  πx  h̄ π
   πx 
+ −2 cos px − sin px (4.20)
2m n L ı L L

π  πx   π 2  πx  2  πx  1 h̄  π   πx 
2m[Hop , M + ] = 2ıh̄ sin px + h̄2 cos − cos p2x − sin px (4.21)
L L L L n L nı L L

   π 2
1 h̄ π  πx   πx  2  πx 
2m[Hop , M ] = +
2ıh̄ + sin px + h̄2 cos + cos p2x (4.22)
nı L L L L n L

   π 2
1 h̄ π n  πx  px o  πx  2  πx 
+
2m[Hop , M ] = ıh̄ 2ıh̄ + sin + h̄2 cos + cos p2x (4.23)
nı L L ıh̄ L L n L
px
and substituting for sin πx


L ıh̄ one has
 π 2 n  πx o  π 2  πx  2  πx 
2m[Hop , M + ] = h̄2 (1 + 2n) M + − cos + h̄2 cos + cos p2x (4.24)
L L L L n L

(1 + 2n)  π 2 n +  πx o 1  π 2  πx  2  πx  p2
[Hop , M + ] = h̄2 M − cos + h̄2 cos + cos x
(4.25)
2m L L 2m L L n L 2m

(1 + 2n)  π 2  + (1 + 2n)  π 2 n  πx o 1  π 2  πx  2  πx  p2
[Hop , M + ] = h̄2 M − h̄2 cos + h̄2 cos + cos x
2m L 2m L L 2m L L n L 2m
(4.26)

(1 + 2n)  π 2  + h̄2  π 2  πx  2  πx 
[Hop , M + ] = h̄2 M + (2n) cos + cos Hop (4.27)
2m L 2m L L n L

1. Alternative Formulation

∂2
  πx  
L  πx  ∂
cos + sin f=
∂x2 L nπ L ∂x
 8

∂
  πx  ∂ π  πx  L  πx  ∂ 2 1  πx  ∂ 
cos − sin + sin + cos f=
∂x L ∂x L L nπ L ∂x2 n L ∂x

π  πx  ∂f  πx  ∂ 2 f
− sin + cos
L L ∂x L ∂x2
 π 2  πx  π  πx  ∂f
− cos f − sin
L L L L ∂x
1  πx  ∂ 2 f L  πx  ∂ 3 f
+ cos + sin
n L ∂x2 nπ L ∂x3
π  πx ∂f
 1  πx  ∂ 2 f
− sin + cos (4.28)
nL L ∂x n L ∂x2
So the commutator would be:
∂2  πx  ∂  ∂ 2 f
  πx   πx  ∂  
L  πx  L
cos + sin f − cos + sin =
∂x2 L nπ L ∂x L nπ L ∂x ∂x2
2π  πx  ∂f
− sin
L L ∂x
 π 2  πx 
− cos f
L L
2
2  πx ∂ f

+ cos
n L ∂x2
π  πx  ∂f
− sin (4.29)
nL L ∂x
or

2π π
  πx  ∂  π 2  πx  2  πx  ∂ 2
= − − sin − cos + cos (4.30)
L nL L ∂x L L n L ∂x2


2π π

nπ  +  πx   π 2  πx  2  πx  ∂ 2
= − − M − cos − cos + cos (4.31)
L nL L L L L n L ∂x2

2m  π 2   πx   π 2  πx  2  πx  ∂ 2
+ +
− [H, M ] = − (2n − 1) M − cos − cos + cos (4.32)
h̄2 L L L L n L ∂x2

2m  π 2  π 2  πx  2  πx  ∂ 2
+
− 2 [H, M ] = − L (2n + 1) M + + (2n) cos + cos (4.33)
h̄ L L n L ∂x2

h̄2  πx  ∂ 2 
  
π 2  π 2  πx  2
[H, M + ] = − − (2n + 1) M + + (2n) cos + cos (4.34)
2m L L L n L ∂x2

h̄2  πx  h̄2 ∂ 2
    πx  2
+ π 2 +
 π 2
[H, M ] = − − (2n + 1) M + 2n cos − cos (4.35)
2m L L L n L 2m ∂x2

H|n >= n |n >

and, operating from the left with M + one has

M + H|n >= n M + |n >= n |n + 1 >

 9

and since [H, M + ] = HM + − M + H, one has

HM + |n > −[H, M + ]|n >= n |n + 1 >

H|n + 1 > −[H, M + ]|n >= n |n + 1 >

h̄2  πx  h̄2 ∂ 2
    πx 
π 2 +
 π 2 2
H|n + 1 > + − (2n + 1) M + 2n cos |n > + cos |n >= n |n + 1 >
2m L L L n L 2m ∂x2

h̄2
    πx 
π 2  π 2 2  πx 
H|n + 1 > + − (2n + 1) M + + 2n cos |n > − cos n |n >= n |n + 1 >
2m L L L n L

h̄2
        πx 
π 2 + π 2 4m
H|n + 1 > + − (2n + 1)M + 2n − n cos |n >= n |n + 1 >
2m L L nh̄2 L

h̄2
       
π 2 π 2 4m  πx 
H|n + 1 > + − (2n + 1)M + |n > + 2n − 2 n cos |n >= n |n + 1 >
2m L L nh̄ L

h̄2
       
π 2 π 2 4m  πx 
H|n + 1 > + − (2n + 1) |n + 1 > + 2n − n cos |n >= n |n + 1 >
2m L L nh̄2 L

h̄2  π 2
     
π 2 4m  πx 
H|n + 1 > + 2n − n cos |n >= n |n + 1 > + (2n + 1) |n + 1 >
L nh̄2 L 2m L

From this we conclude that, if the elements in square V. HERMITE POLYNOMIALS

brackets cancelled perfectly, M + |n + 1 > would be an
eigenfunction, i.e., if The relevant Schrödinger Equation is
h̄2 ∂ 2 k
   
π 2 4m − ψ + z 2 ψ = Eψ (5.1)
+ 2n − n = 0 (4.36) 2µ ∂z 2 2
L nh̄2
where k is the force constant (dynes/cm) and µ is the
i.e., reduced mass (grams). Cross multiplying, one has

n2 h̄2 π 2 ∂2 kµ 2µ
n = ψ − 2 z 2 ψ = − 2 Eψ (5.2)
2mL2 ∂z 2 h̄ h̄
which would be simplified if the constants could be sup-
then pressed. To do this we change variable, from z to some-
thing else, say x, where z = αx. Then
h̄2
   
π 2
H|n + 1 >= n + (2n + 1) |n + 1 > ∂ ∂x ∂ 1 ∂
2m L = =
∂z ∂z ∂x α ∂x
so
Sines and cosines form a complete orthonormal set, suit-
∂2
 
able for Fourier Series expansion, where ‘ortho’ means 1 kµ 2µ
ψ − 2 α2 x2 ψ = − 2 Eψ (5.3)
that each member is orthogonal to every other member, α2 ∂x2
h̄ h̄
and ‘normal’ means that one can normalize them (if one
wishes). There are other such sets, and each of them and
shows up in standard Quantum Chemistry, so each needs ∂2 kµ 2µ
to be addressed separately. 2
ψ − 2 α4 x2 ψ = −α2 2 Eψ (5.4)
∂x h̄ h̄
 10

which demands that we treat which leads to  = 1 and

kµ 4 s
1= α h̄ k
h̄2 E=
2 µ
!1/4 1/4
h̄2

1
α= kµ
=
kµ
h̄2
B. Laddering up on the Guessed Solution
With this choice, the differential equation becomes

∂2ψ Given
− x2 ψ = −ψ (5.5)
∂x2
x2

where ψ0 = |0 >= e− 2

q
h̄2
2 kµ µE 2E µk with  = 1, it is possible to generate the next solution by
p
2
2α µE
= = = using
h̄2 h̄ 2 h̄

∂
A. A Guessed Solution N+ = − +x (5.6)
∂x

The easiest solution to this differential equation is as an operator, which ladders up from the ground (n=0)
2
state to the next one (n=1) To see this we apply N + to
− x2
e ψ0 obtaining

 
∂ x2 2
+
N ψ0 = N |0 >= +
− + x e− 2 = − (−x) ψ0 + xψ0 = 2xe−x /2 = ψ1 = |1 > (5.7)
∂x

Doing this operation again, one has where H(x) is going to become a Hermite polynomial.
  One then has
+ + ∂ − x2
2
2 −x2 /2
N ψ1 = N |1 >= − + x 2xe = (−2+4x )e
∂x
(5.8)
etc., etc., etc..
dψ 2 2 dH(x)
= −xe−x /2 H(x) + e−x /2
dx dx
C. Generating Hermite’s differential equation

Assuming
2
ψ = e−x /2
H(x) and

d2 ψ −x2 /2 2 −x2 /2 −x2 /2 dH(x)

2
−x2 /2 d H(x)
= −e H(x) + x e H(x) − 2xe + e
dx2 dx dx2
From Equation 5.5 one has,
∂2 2 2 dH(x) 2 d2 H(x) 2

2
ψ − x2 ψ = −e−x /2 H(x) − 2xe−x /2 + e−x /2 = −e−x /2 H(x) (5.9)
∂x dx dx2

or which we re-write in normal lexicographical order

2
dH(x) d H(x) d2 H(x) dH(x)
−H(x) − 2x + = −H(x) (5.10) − 2x − (1 − )H(x) = 0 (5.11)
dx dx2 dx 2 dx
 11

D. An alternative solution scheme so, integrating each side separately, one has

Starting with `ny = −x2 + `nC

dy or, inverting the logarithm,

+ 2xy = 0 (5.12)
dx
2
one has y = Ce−x
dy
= −2xdx We now differentiate Equation 5.12, obtaining
y

d2 y d(xy) d2 y dy dx d2 y dy
2
+ 2 = 2
+ 2x + 2y = 2
+ 2x + 2y = 0 ; n = 0 (5.13)
dx dx dx dx dx dx dx
Doing this again, i.e., differentiating this (second) equation (Equation 5.13), one has
   
dy dy
d2 y
d dx 2 d2x d(y) dy d2 dx d dx  
dy
dx
+ +2 = 2
+ 2x +4 = 0; n = 1
dx dx dx dx dx dx

which is the same equation, (but with a 4 multiplier of d2 f (x) df (x)

2
+ 2x + 6f (x) = 0 ; n = 2
the last term) applied to the first derivative of y. Take dx dx
the derivative again:
2
 2 dy
d dy
d dx dy
 f (x) has the form g(x)e−x where g(x) is a polynomial
d dxdx 2 + 2x dx + 4 dx in x.
=0
dx
2 2

i.e., d2 g(x)e−x dg(x)e−x 2

2
+ 2x + 2(n + 1)g(x)e−x = 0
    dx dx
d2 y d2 y
d2 dx2 d dx2

d2 y

+ 2x +6 =0 i.e.,
dx2 dx dx2

2
g 00 (x) − 4xg 0 (x) − 2g(x) + 4x2 g(x) + 2xg 0 (x) − 4x2 g(x) + 2(n + 1)g(x) e−x = 0

or i.e.,

g 00 (x) − 2xg 0 (x) + 2ng(x) = 0

s
k
E = h̄(n + 1/2)
µ
and we had (see Equation 5.11)

H 00 (x) − 2xH 0 (x) − (1 − )H(x) = 0

E. Frobenius Method
which leads to
The most straight forward technique for handling the
2n = −1 +  Hermite differential equation is the method of Frobenius.
We assume a power series Ansatz (ignoring the indicial
i.e., equation argument here), i.e.,
p X
2E µ/k ψ= ai xi
 = 1 + 2n =
h̄ i=0
 12

and substitute this into Equation 5.11, obtaining Consider Laplace’s Equation:

∂2ψ X
= i(i − 1)ai xi−2 ∂2χ ∂2χ ∂2χ
∂x2 i=2 ∇2 χ = + + 2 =0 (6.1)
∂x2 ∂y 2 ∂z
∂ψ X
−2x = −2 iai xi
∂x i=1
X which is a partial differential equation for χ(x, y, z). For
( − 1)ψ = ( − 1) ai xi = 0 our purposes, the solutions can be obtained by guessing,
i starting with χ = 1, and proceeding to χ = x, χ = y and
χ = z. It takes only a little more imagination to obtain
i.e., χ = xy and its associates χ = yz and χ = xz. It takes
just a little more imagination to guess χ = x2 − y 2 , but
∂2ψ
= 2(1)a2 + (3)(2)a3 x + (4)(3)a4 x2 + · · · once done, one immediately guesses its two companions
∂x2 χ = x2 − z 2 and χ = y 2 − z 2 .
∂ψ
−2x = −2a1 x1 − 2a2 x2 − 2a3 x3 − · · · Reviewing, there was one simple (order 0) solution,
∂x
2 three not so simple, but not particularly difficult solu-
( − 1)ψ = ( − 1)a0 + ( − 1)a1 x + ( − 1)a2 x − · · · = 0
tions (of order 1) and six slightly more complicated solu-
which leads to tions of order 2. Note that the first third-order solution
one might guess would be χ = xyz!
(3)(2)a3 + ( − 1)a1 − 2a1 = 0 (odd) If one looks at these solutions, knowing that they are
quantum mechanically important, the 1, x, y, and z so-
lutions, accompanied by the xy, xz, yz, and x2 − y 2 so-
(4)(3)a4 − 2a2 + ( − 1)a2 = 0 (even)
lutions suggest something having to do with wave func-
tions, where the first function (1) is associated in some
(5)(4)a5 − 2a3 + ( − 1)a3 = 0 (odd) way with s-orbitals, the next three (x, y, and z) are as-
sociated with p-orbitals, and the next (of order 2) are
etc., which shows a clear distinction between the even associated with d-orbitals. If all this is true, the per-
and the odd powers of x. We can solve these equations ceptive student will wonder where is the dz2 orbital, the
sequentially. fifth one, and how come there are six functions of order
We obtain two listed, when, if memory serves correctly, there are 5
d-orbitals! Ah.
2+1−
a3 = a1
(3)(2)

y 2 − z 2 + x2 − y 2
  


2+1− 2+1− 1−
a4 = a2 =
(4)(3) (4)(3) (2)(1)
can be rewritten as
i.e.,
 
(3 − )(1 − ) y 2 − z 2 + x2 − y 2 + z 2 − z 2




a4 =
(4)(3)(2)(1)
etc.. which is
This set of even (or odd) coefficients leads to a series
which itself converges unto a function which grows to
positive infinity as x varies, leading one to require that x2 + y 2 + z 2 − 3z 2
the series be terminated, becoming a polynomial.
We leave the rest to you and your textbook.
or as

VI. LEGENDRE POLYNOMIALS r2 − 3z 2

A. From Laplace’s Equation

which combines the last two (linearly dependent) solu-
There are so many different ways to introduce Leg- tions into one, the one of choice, which has been employed
endre Polynomials that one searchs for a path into this for more than 50 years as the dz2 orbital.
subject most suitable for chemists. In Spherical Polar Coördinates, these solutions become
 13

1⇔1
x ⇔ r sin ϑ cos φ
y ⇔ r sin ϑ sin φ
z ⇔ r cos ϑ
xy ⇔ r sin ϑ cos φr sin ϑ sin φ = r2 sin2 ϑ cos φ sin φ
xz ⇔ r sin ϑ cos φr cos ϑ = r2 sin ϑ cos ϑ cos φ
yz ⇔ r sin ϑ sin φr cos ϑ = r2 sin ϑ cos ϑ sin φ
x2 − y 2 ⇔ r2 sin2 ϑ cos2 φ − r2 sin2 ϑ sin2 = r2 sin2 ϑ cos2 φ − sin2 φ
 

r2 − 3z 2 ⇔ r2 − 3r2 cos2 ϑ = r2 1 − 3 cos2 ϑ

..
. (6.2)

Now, Laplace’s equation in spherical polar coördinates which, by the nature of partial differentiation, becomes
is
" #
∂χ
2 1 ∂r2 ∂r 1 ∂ sin ϑ ∂χ
∂ϑ ∂2χ
∇ χ= 2 + 2 2 sin ϑ + =0
r ∂r r sin ϑ ∂ϑ ∂φ2
`
(6.3) 1 ∂r2 ∂r
Next, we notice that our earlier solutions were all of the ∇2 r` Y`,m` = Y`,m` ∂r
r2 ∂r
form " ∂Y`,m`
#
` 1 ∂ sin ϑ ∂ϑ ∂ 2 Y`,m`
+r 2 2 sin ϑ + = 0 (6.5)
χ = r` Y`,m` (r, ϑ, φ) (6.4) r sin ϑ ∂ϑ ∂φ2

where ` specifies the order, 0, 1, 2, etc., and the func-

tion of angles is dependent on this order, and on another
quantum number as yet unspecified. Substituting this
form into the spherical polar form of Laplace’s equation and the first term in Equation 6.5 is
results in a new equation for the angular parts alone,
which we know from our previous results. We substitute
Equation 6.4 into Equation 6.3 to see what happens, ob-
taining
`
∂r ` Y`,m 1 ∂r2 ∂r 1
1 ∂r2 ∂r ` ∂r
= 2 `(` + 1)r` (6.6)
∇2 r` Y`,m` = 2 r2 ∂r r
 r ∂r 
∂r ` Y`,m` 2 `
1 ∂ sin ϑ ∂ϑ ∂ r Y`,m` 
+ 2 2 sin ϑ +
r sin ϑ ∂ϑ ∂φ2
=0 so, substituting into Equation 6.5 we obtain

∂Y`,m`
" #
2 ` `−2 `−2 1 ∂ sin ϑ ∂ϑ ∂ 2 Y`,m`
∇ r Y`,m` = `(` + 1)r Y`,m` + r sin ϑ + =0 (6.7)
sin2 ϑ ∂ϑ ∂φ2

which, of course, upon cancelling common terms, gives Laplace’s Equation in Spherical Polar Coördinates on the
" ∂Y`,m`
# unit sphere, i.e. Legendre’s Equation!
1 ∂ sin ϑ ∂ϑ ∂ 2 Y`,m`
`(`+1)Y`,m` + 2 sin ϑ + =0
sin ϑ ∂ϑ ∂φ2
(6.8)
 14

What then are these angular solutions we found? and if you check back concerning their origin, you will
see where the Hydrogenic Orbitals naming pattern comes
` = 0; 1 from for s, p and d orbitals.
` = 1; sin ϑ cos φ
` = 1; sin ϑ sin φ
` = 1; cos ϑ
` = 2; sin2 ϑ cos φ sin φ
` = 2; sin ϑ cos ϑ cos φ
` = 2; sin ϑ cos ϑ sin φ
` = 2; sin2 ϑ cos2 φ − sin2 φ
 

` = 2; 1 − 3 cos2 ϑ


These functions have been written in “real” form, and
.. they have “complex” (not complicated) forms which al-
. (6.9) low a different simplification:

1s ⇔ ` = 0; 1
eıφ + e−ıφ
2px ⇔ ` = 1; sin ϑ
2
e − e−ıφ
 ıφ 
2py ⇔ ` = 1; sin ϑ
2ı
2pz ⇔ ` = 1; cos ϑ
e + e−ıφ e − e−ıφ
 ıφ   ıφ 
3dxy ⇔ ` = 2; sin2 ϑ
2 2ı
eıφ + e−ıφ
3dxz ⇔ ` = 2; sin ϑ cos ϑ
2
e − e−ıφ
 ıφ 
3dyz ⇔ ` = 2; sin ϑ cos ϑ
2ı
" 2  ıφ 2 #!
2 e + e−ıφ
ıφ
e − e−ıφ
3dx2 −y2 ⇔ ` = 2; sin ϑ −
2 2ı
2


3dz2 ⇔ ` = 2; 1 − 3 cos ϑ
..
. (6.10)

Once they have been written in imaginary form, we can We have a trio of functions
create intelligent linear combinations of them which illu-
minate their underlying structure. Consider 2px + ı2py
2pm` =−1 = sin ϑe−ıφ
which becomes
2pm` =0 = cos ϑ
2pm` =+1 = sin ϑe+ıφ (6.11)
ıφ (6.12)
sin ϑe

which correspond to the m` values expected of p-orbitals.

(employing DeMoivre’s theorem) aside from irrelevant
constants. Next consider 2px − ı2py which becomes
sin ϑe−ıφ
You will see that the same “trick” can be applied to
dxz and dyz .
Finally, the same trick, almost, can be applied to dxy
and dx2 −y2 with the m` = 0 value reserved for dz2 , all by
itself.
 15

B. From the Hydrogen Atom the sense that

Perhaps the next best place to introduce Spherical Har- |n, `, m` >= Rn,` (r)Y`,m` (ϑ, φ)
monics and Legendre Polynomials is the Hydrogen Atom,
since its eigenfunctions have angular components which Here, Rn,` is the radial wave function, and Y`,m` is the
are known to be Legendre Polynomials. angular wave function (a function of two variables), Since
The Schrödinger Equation for the H-atom is
2
" #
h̄2 2 Ze2 2 1 ∂ r∂r∂ 1 ∂
∂ sin ϑ ∂ϑ ∂2
− ∇ |n, `, m` > − |n, `, m` >= En |n, `, m` > ∇ = 2 + 2 2 sin ϑ +
2me r r ∂r r sin ϑ ∂ϑ ∂φ2
(6.13)
We know that this equation is variable separable, in

" #
2 2 Y ∂r2 ∂R
∂r R ∂ sin ϑ ∂Y
∂ϑ ∂2Y
∇ |n, `, m` >= ∇ Rn,` (r)Y`,m` (ϑ, φ) = 2 + 2 2 sin ϑ +
r ∂r r sin ϑ ∂ϑ ∂φ2

which means that the Schrödinger Equation has the form

( " #)
h̄2 Y ∂r2 ∂R
∂r R ∂ sin ϑ ∂Y
∂ϑ ∂ 2
Y Ze2
− + 2 sin ϑ + − RY = En RY
2me r2 ∂r r2 sin ϑ ∂ϑ ∂φ2 r

Dividing through by Rn,` Y`,m` (abreviated as RY) we have

( " #)
h̄2 1 ∂r2 ∂R
∂r 1 ∂ sin ϑ ∂Y
∂ϑ ∂2Y Ze2
− 2
+ 2 2 sin ϑ + 2
− = En
2me Rr ∂r r Y sin ϑ ∂ϑ ∂φ r

Multiplying through by r2 shows that the expected variable separation has resulted in a proper segregation of the
angles from the radius.
" #
1 ∂r2 ∂R 1 ∂ sin ϑ ∂Y
∂ 2
Y 2mZe2 r 2m
∂r
+ 2 sin ϑ ∂ϑ
+ 2
+ 2 = − 2 En r 2
R ∂r Y sin ϑ ∂ϑ ∂φ h̄ h̄
| {z }

We can see this by noting that the underbracketed part nating. We write
of this last equation is a pure function of angles, with no
radius explicitly evident. µ = cos ϑ
It is now standard to obtain (recover?) the equation
(6.8)
" # and
1 ∂ sin ϑ ∂Y
∂ϑ ∂2Y
sin ϑ + = −`(` + 1) (6.14) ∂µ p p
Y sin2 ϑ ∂ϑ ∂φ2 = − sin ϑ = − 1 − cos2 ϑ = − 1 − µ2
∂ϑ
where the minus sign (which is essentially arbitrary) is
demanded by convention. so
Cross multipilying by Y, one has  
∂ ∂µ ∂
=
" #
1 ∂ sin ϑ ∂Y
∂ϑ ∂ 2
Y ∂ϑ ∂ϑ ∂µ
sin ϑ + = −`(` + 1)Y (6.15)
sin2 ϑ ∂ϑ ∂φ2
which is, substituting into Equation 6.15
which is written in traditional eigenfunction/eigenvalue
form.
∂  p  ∂
C. Another Form for the Legendre Differential = − 1 − µ2
Equation ∂ϑ ∂µ

A change of variables can throw this equation (Equa-

tion 6.15) into a special form, which is sometimes illumi-
 16

" p p
∂ 1 − µ2 1 − µ2 ∂Y
#
1 p
2
p
2 ∂µ ∂2Y
1−µ 1−µ + = −`(` + 1)Y (6.16)
(1 − µ2 ) ∂µ ∂φ2

where the `(`+1) form is a different version of a constant, To proceed, we need to generate a function |2 > which
looking ahead to future results! is not only normalizeable but orthogonal to |0 > and
|1 >.
    Z −1 r
∂ (1 − µ2 ) ∂Y ∂ 2
Y
1
∂µ
 = −`(` + 1)Y < 2|0 >= ψ2 dµ = 0
 + (6.17) +1 2
∂µ ∂φ2
and
where Z −1
r
2
< 2|1 >= ψ2 µdµ = 0
Y`,m` (ϑ, φ) = e ±ım` φ
S`,m` (ϑ) +1 3

For m` = 0 we have where |2 > is a polynomial of order 2 in µ, i.e.,

  ∂S
 ψ2 = |2 >= aµ2 + b
∂ (1 − µ2 ) ∂µ`,0
  = −`(` + 1)S`,0 (6.18) Substituting, we have
∂µ
−1
Z r
2 1
< 2|0 >= (aµ + b) dµ = 0
i.e., Legendre’s equation. +1 2
and
−1
Z r
D. Schmidt Orthogonalization 2 3
< 2|1 >= (aµ + b) µdµ = 0
+1 2
We assume polynomials in µn and seek orthonormal We obtain two equations in two unknowns, a and b,
combinations which can be generated starting with a con- r 
stant (n=0) term. Then the normalization integral

1 2a
+ 2b = 0
Z −1 2 3
< 0|0 >= ψ02 dµ = 1 and the other integral vanishes automatically. i.e.,
+1
a
[4] implies that +b=0
3
i.e., a = −3b so
r
1
|0 >=
2 |2 >= −3bµ2 + b
We now seek a function |1 > orthogonal to |0 > which in unnormalized form. Normalizing gives
itself is normalizeable. We have Z −1
Z −1 r < 2|2 >= (−3bµ2 + b)2 dµ = 1
1 +1
0 =< 1|0 >= ψ1 dµ
+1 2 i.e.,
which has solution ψ1 = N1 x. Normalizing, we have Z −1

Z −1 Z −1 < 2|2 >= b2 (−3µ2 + 1)2 dµ = 1

+1
< 1|1 >= ψ12 dµ = 1 = N12 µ2 dµ
+1 +1 which gives
s
which yields 1
b= R −1
2 +1
(−3µ2 + 1)2 dµ
1 = N12
3 where, of course, we recognize the sign ambiguity in this
i.e., result.
r One can continue forever with this Schmidt Orthog-
3 onalization, but the idea is clear, and there are better
|1 >= x ways, so why continue?
2
 17

E. Frobenius Solution to Legendre’s Equation which leads to

   
We start a Frobenius solution without worrying about ∂ ∂y(µ)
∂ µ2 ∂y(µ)
∂µ ∂µ
the technical details of the indicial equation, and just − + `(` + 1)y(µ) = 0 (6.21)
assert that the proposed solution Ansatz will be ∂µ ∂µ

∞
X
S`,0 (µ) = y(µ) = cn µn = c0 +c1 µ+c2 µ2 +· · · (6.19)
n=0 ∂ 2 y(µ) 2
2 ∂ y(µ) ∂y(µ)
2
− µ 2
− 2µ + `(` + 1)y(µ) = 0 (6.22)
which we substitute into the differential equation: ∂µ ∂µ ∂µ
 
∂ (1 − µ2 ) ∂y(µ)
∂µ
so that when we feed the Ansatz into this differential
= −`(` + 1)y(µ) (6.20) equation we obtain
∂µ

(2)(1)c2 + (3)(2)c3 µ + (4)(3)c4 µ2 + (5)(4)c5 µ3 · · ·

−µ2 (2)(1)c2 + (3)(2)c3 µ + (4)(3)c4 µ2 + (5)(4)c5 µ3 · · ·

−2µ c1 + (2)c2 µ + (3)c3 µ2 + (4)c4 µ3 + · · ·

+`(` + 1) c0 + c1 µ + c2 µ2 + c3 µ3 + · · · = 0


(6.23)

(2)(1)c2 + (3)(2)c3 µ + (4)(3)c4 µ2 + (5)(4)c5 µ3 · · ·

−(2)(1)c2 µ2 − (3)(2)c3 µ3 − (4)(3)c4 µ4 − (5)(4)c5 µ5 − · · ·
−2c1 µ − 2(2)c2 µ2 − 2(3)c3 µ3 − 2(4)c4 µ4 − · · ·
+`(` + 1)c0 + `(` + 1)c1 µ + `(` + 1)c2 µ2 + `(` + 1)c3 µ3 + · · · = 0 (6.24)

which, in standard Frobenius form, we separately equate `(` + 1)

c2 = − c0
to zero (power by power) to achieve the appropriate re- 2!
cursion relationships. Note that there is an even and an (2` + (` + 1))
c3 = − c1
odd set, based on starting the c0 or c1 , which correspond (3)(2)
to the two arbitrary constants associated with a second (−(2)(1) − 2(2) + `(` + 1))
order differential equation. We obtain c4 = − c2
(4)(3)
(2)(1)c2 + `(` + 1)c0 = 0 ((3)(2) − 2(3) + `(` + 1))
+(3)(2)c3 µ − 2c1 µ + `(` + 1)c1 µ = 0 c5 = − c3
(5)(4)
+(4)(3)c4 µ − (2)(1)c2 µ2 − 2(2)c2 µ2 + `(` + 1)c2 µ2 = 0
2
etc. (6.27)
(5)(4)c5 µ3 − (3)(2)c3 µ3 − 2(3)c3 µ3 + `(` + 1)c3 µ3 = 0
which we re-write in terms of c0 and c1 only, i.e.,
−(4)(3)c4 µ4 − 2(4)c4 µ5 + · · · etc = 0
`(` + 1)
c2 = − c0
(2)(1)c2 + `(` + 1)c0 = 0 2!
+(3)(2)c3 − 2c1 + `(` + 1)c1 = 0 (2 + `(` + 1))
c3 = − c1 +
+(4)(3)c4 − (2)(1)c2 − 2(2)c2 + `(` + 1)c2 = 0 (3)(2)
  
(5)(4)c5 − (3)(2)c3 − 2(3)c3 + `(` + 1)c3 = 0 ((2)(1) − 2(2) + `(` + 1)) `(` + 1)
c4 = − (− c0
−(4)(3)c4 µ4 − 2(4)c4 µ5 etc + · · · = 0 (6.25) (4)(3) 2!
  
((3)(2) − 2(3) + `(` + 1)) (2 + `(` + 1))
(2)(1)c2 = −`(` + 1)c0 c5 = − − c1
(5)(4) (3)(2)
+(3)(2)c3 = (2 − `(` + 1))c1 etc.
(6.28)
+(4)(3)c4 = ((2)(1) + 2(2) − `(` + 1))c2
so S`,0 (µ) = f1 c0 + f2 c1 where f1 and f2 are power series
(5)(4)c5 ((3)(2) + 2(3) − `(` + 1))c3 based on the above set of coëfficients. For an even series,
−(4)(3)c4 µ4 − 2(4)c4 µ5 etc + · · · = 0 (6.26) declare c1 = 0 and choose an ` value which truncates the
 18

power series into a polynomial. Do the opposite for an To show this, we start by defining
odd solution.
g` ≡ (µ2 − 1)`

F. Rodrique’s Formula and find that

dg`
Rodrique’s formula is = 2µ`(µ2 − 1)`−1
dµ
1 d` (µ2 − 1)`
S`,0 → P` (µ) = and
2` `! dµ`
d2 g`
where Legendre’s Equation is = 2`(µ2 − 1)`−1 + 4µ2 `(` − 1)(µ2 − 1)`−2
dµ2
P` (µ) dP`
(1 − µ2 ) 2
− 2µ + `(` + 1)P` (µ) = 0 We now form (construct)
dµ dµ

d2 g`
(1 − µ2 ) = −2`(µ2 − 1)` − 4µ2 `(` − 1)(µ2 − 1)`−1
dµ2
dg`
2(` − 1)µ = 4µ2 `(` − 1)(µ2 − 1)`−1
dµ
+2`g` = 2`(µ2 − 1)` (6.29)

The r.h.s. of this equation set adds up to zero, and one obtains on the left:

d2 g` dg`
A(µ) = (1 − µ2 ) + 2(` − 1)µ + 2`g` = 0 (6.30)
dµ2 dµ

Defining the l.h.s of this equation as A(µ), we form

d3 d2 d2
 
dA(µ) d d
= (1 − µ2 ) 3 − 2µ 2 + 2(` − 1)µ 2 + 2(` − 1) + 2` g`
dµ dµ dµ dµ dµ dµ

d2
 
d dg` Continuing, one notices that the changing coëfficients are
= (1 − µ2 ) 2 − (2` − 4)µ + (4` − 2)µ
dµ dµ dµ regular in their appearance, so that the following table,
which summarizes the pattern of coëfficients,

d2 A(µ) 2
  2
2 d d d g`
= (1 − µ ) 2 − (2` − 6)µ + (6` − 6)µ
dµ2 dµ dµ dµ2

d3 A(µ) 2
  3
2 d d d g`
3
= (1 − µ ) 2 − (2` − 8)µ + (8` − 12)µ
dµ dµ dµ dµ3

κ = 1 2` − 4 = 2` − 2 ∗ (κ + 1) 4` − 2 = (3 + κ)` − 2 ∗ κ =6 if ` = 1
κ = 2 2` − 6 = 2` − 2 ∗ (κ + 1) 6` − 6 = (3 + κ)` − 2 ∗ κ = 6 if ` = 2
κ = 3 2` − 8 = 2` − 2 ∗ (κ + 1) 8` − 12 = (3 + κ)` − 2 ∗ κ = 12 if ` = 3
.. .. .. ..
. . . .
κ=` 2` − 2 ∗ (` + 1) = 2 8` − 12 = (3 + `)` − 2 ∗ ` = `(` + 1)

leads to generalization by which one finally obtains so, if

∂ ` A(x) 2
 
2 ∂ ∂ g` (µ) d` g`
= (1 − µ ) − 2µ + `(` + 1) ≡ KP` (µ)
∂µ` ∂µ2 ∂µ dµ` dµ`
(6.31)
 19

with K constant, then Here, we expand the denominator using the binomial the-
orem,
1 ∂d` g` 1 d` (µ2 − 1)`
P` (µ) = =
K dµ` K dµ`
Here,

2` `!
1 m(m + 1) 2 m(m + 1)(m + 2) 3
is chosen for K’s value to make the normalization auto- = 1−my+ y − y +· · ·
matic. (1 + y)m 2! 3!

G. Generating Function for Legendre Polynomials

The technically correct generating function for Legen-

dre polynomials is via the equation
where m = 12 and the series converges when y < 1. Notice
1 X ∞ that it is an alternating series. Identifying y = u2 − 2xu
√ = Pn (x)un (6.32) we have
1 − 2xu + u 2
0

1 (1/2)((1/2) + 1) 2 (1/2)((1/2) + 1)((1/2) + 2) 2

2 (
= 1−(1/2)(u2 −2xu)+ (u −2xu)2 − (u −2xu)3 +· · ·
(1 − 2xu + u ) 1/2) 2! 3!
which we now re-arrange in powers of u (in the mode required by Equation 6.32), obtaining
1 u2 (3/4) 4 (1/2)(3/2)(5/2) 2
= 1 − + xu + (u − 4xu3 + 4x2 u2 ) − (u − 2xu)3 + · · ·
(1 − 2xu + u2 )1/2 2 2! 3!

1. Alternative Generating Function Method where r1 and r2 are the distances from the nucleus to
electrons 1 and 2 respectively. ϑ is the angle between
Another method of introducing Legendre Polynomials the vectors from the nucleus to electron 1 and electron
is through the generating function. Since this method 2. It is required that we do this is two domains, one in
is very important in Quantum Mechanical computations which r1 > r2 and one in which r2 > r1 . This is done for
concerning poly-electronic atoms and molecules, it is convergence reasons (vide infra). For the first case, we
worth our attention. When one considers the Hamilto- define
nian of the Helium Atom’s electrons, one has r2
ζ=
Ze2 Ze2 e2 r1
− − + (6.33)
r1 r2 r12
so that ζ < 1, and Equation 6.34 becomes
where r12 is the distance between electron 1 and elec-
tron 2, i.e., it is the electron-electron repulsion term. We 1 1 1
= p (6.35)
examine this term in this discussion. We can write this r12 r1 1 + ζ 2 − 2ζ cos ϑ
electron-electron repulsion term as
1 1 which we now expand in a power series in cos ϑ (which
=p 2 (6.34) will converge while zeta < 1). We have
r12 2
r1 + r2 − 2r1 r2 cos ϑ

     
√ 21 2 √ 21
 
d  d 
1 1 1 1 + 1
1+ζ −2ζ cos ϑ 1 1+ζ −2ζ cos ϑ
cos2 ϑ + · · · 
  
=  cos ϑ + 
d cos ϑ2
p
r1 1 + ζ 2 − 2ζ cos ϑ r1  1! d cos ϑ 
 2! 


 
cos ϑ=0 cos ϑ=0
(6.36)
 20

It is customary to change notation from cos ϑ to µ, so which we now expand in a power series in µ (which will
converge while ζ < 1). We have
1 1 1
= p (6.37)
r12 r1 1 + ζ 2 − 2ζµ

     
√ 1  2 √ 1 
d  d 
1 1 1 1 + 1
1+ζ 2 −2ζµ 1 1+ζ 2 −2ζµ
µ2 + · · · 
  
=  µ+ 
+ (6.38)
dµ2
p
r1 1 + ζ 2 − 2ζµ r1  1! dµ 
 2! 

 
µ=0 µ=0

which we leave in this form, ready to evaluate and equate the “bond length” is “a”, and therefore the “dipole mo-
to earlier versions of this expansion. ment” is “qa”.
At some point P (x, y, z), located (also) at r,ϑ,φ, we
have that the potential energy due to these two point
H. The Expansion of a Finite Dipole in Legendre charges is
Polynomials
−q q
U (x, y, z, a) = p +p
There is yet another way to see Legendre Polynomials x2 + y 2 + (z − a/2)2 x2 + y 2 + (z + a/2)2
in action, through the expansion of the potential energy
of point dipoles. To start, we assume that we have a which is just Coulomb’s law.
dipole at the origin, with its positive charge (q) at (0,0,- If we expand this potential energy as a function of “a”,
a/2) and its negative charge (−q) at (0,0,+a/2), so that the “bond distance”, we have

1 d2 U  1 d3 U 
  
1 dU  2
U (x, y, z, a) = U (x, y, z, 0) + a+ a + a3 + · · ·
1! da a=0 2! da2 a=0 3! da3 a=0

All we need do, now, is evaluate these derivatives. We have, for the first
         
1 dU  1 2(z − a/2)(−1/2) 1 2(z + a/2)(1/2)
= q − − + −
1! da a=0 2 (x2 + y 2 + (z − a/2)2 ))3/2 2 (x2 + y 2 + (z + a/2)2 ))3/2

which is, in the limit a → 0, i.e., to the dipolar form.

 z
q − 3
r
so, we have, so far,
z 
cos ϑ
 For the second derivative, we take the derivative of the
U (x, y, z, a) = 0 − qa + · · · = −qa + ··· first derivative:
r3 r2

  h (z−a/2)
i h
(z+a/2)
i 
1 d dU
 1  d (x2 +y 2 +(z−a/2)2 ))3/2
+ (x2 +y 2 +(z+a/2)2 ))3/2
da
= −q
 
2! da  2 × 2! da

a=0

which equals zero. The next term gives and so it goes.

cos ϑ 3 − 5 cos2 ϑ 3
q a
8r4
 21

I. Alternative Formulations for Angular

Momentum Operators  
∂φ cos φ
= (6.46)
To proceed, it is of value to inspect the angular mo- ∂y x,z r sin ϑ
mentum operator in terms of angles rather than Carte-
sian coördinates.  
Remember that ∂φ
=0 (6.47)
∂z x,y
x = r sin ϑ cos φ
y = r sin ϑ sin φ which we employ on the defined x-component of the an-
z = r cos ϑ gular momentum, thus
 
We start with a feast of partial derivatives: ∂ ∂
Lx ≡ ypz − zpy = −ıh̄ sin ϑ sin φ − −ıh̄ cos ϑ
  ∂z ∂y
∂r
= sin ϑ cos φ (6.39) where
∂x y,z
     
∂ ∂r ∂ ∂ϑ ∂ ∂φ ∂
= + +
∂z ∂z ∂r ∂z ∂ϑ ∂z ∂φ
 
∂r x,y x,y x,y
= sin ϑ sin φ (6.40)
 
∂ sin ϑ ∂ ∂
∂y x,z = (cos ϑ) + + (0)
∂r r ∂ϑ ∂φ

∂r
 and
= cos ϑ (6.41)
∂z x,y ∂

∂r

∂

∂ϑ

∂

∂φ

∂
= + +
∂y ∂y x,z ∂r ∂y x,z ∂ϑ ∂y x,z ∂φ
 
∂ϑ cos ϑ cos φ
= (6.42)
∂x y,z r ∂

cos ϑ sin φ

∂

cos φ

∂
= (sin ϑ sin φ) + +
∂r r ∂ϑ r sin ϑ ∂φ
 
∂ϑ cos ϑ sin φ
= (6.43) and parenthetically,
∂y x,z r
     
∂ ∂r ∂ ∂ϑ ∂ ∂φ ∂
= + +

∂ϑ

sin ϑ ∂x ∂x y,z ∂r ∂x y,z ∂ϑ ∂x y,z ∂φ
=− (6.44)
∂z x,y r
   
∂ cos ϑ cos φ ∂ sin φ ∂
and finally = (sin ϑ cos φ) + + −
∂r r ∂ϑ r sin ϑ ∂φ
 
∂φ sin φ
=− (6.45) so,
∂x y,z r sin ϑ

   
∂ sin ϑ ∂
Lx ≡ ypz − zpy = −ıh̄r sin ϑ sin φ (cos ϑ) + −
∂r r ∂ϑ
      
∂ cos ϑ sin φ ∂ cos φ ∂
− −ıh̄r cos ϑ (sin ϑ sin φ) + + (6.48)
∂r r ∂ϑ r sin ϑ ∂φ

At constant r, the partial with respect to r looses mean- ing, and one has
   
Lx sin ϑ ∂
= r sin ϑ sin φ + −
−ıh̄ r ∂ϑ
    
cos ϑ sin φ ∂ cos φ ∂
− cos ϑ + (6.49)
r ∂ϑ r sin ϑ ∂φ
 22

which leads to (combining the ϑ partial derivative terms and knowing

and cancelling the r terms)
  Lx Ly −Ly Lz = (ypz −zpy )(zpx −xpz )−(zpx −xpz )(ypz −zpy )
∂ cos ϑ cos φ ∂
Lx = ıh̄ sin φ + (6.50)
∂ϑ sin ϑ ∂φ which is

and similar arguments lead to = ypz zpx − ypz xpz − zpy zpx + zpy xpz − zpx ypz
  +zpx zpy + xpz ypz − xpz zpy = ypz zpx
∂ cos ϑ sin φ ∂
Ly = ıh̄ cos φ − (6.51)
∂ϑ sin ϑ ∂φ which is, reordering:

= ypz zpx − zpx ypz − ypz xpz + xpz ypz

∂
Lz = ıh̄ (6.52) −zpy zpx + zpx zpy + zpy xpz − xpz zpy
∂φ
with which is, cancelling:

= ypz zpx − zpx ypz + zpy xpz − xpz zpy

" #!
∂
2 2 1 ∂ sin ϑ ∂ϑ ∂2
L = −h̄ + (6.53) = ypx (pz z − zpz ) + xpy (zpz − pz z)
sin2 ϑ ∂ϑ ∂φ2
= ypx (−ıh̄) + xpy (ıh̄)
We see that the operator associated with Legendre’s = ıh̄(xpy − ypx ) = ıh̄Lz = [Lx , Ly ]
Equation (multiplied by a constant) has emerged, mean-
ing that Legendre Polynomials and angular momentum that
are intimately asssociated together.
[Lx , Ly ] = ıh̄Lz (6.54)

J. Ladder Operator for Angular Momentum

Defining [Ly , Lz ] = ıh̄Lx (6.55)

L+ = Lx + ıLy

and [Lz , Lx ] = ıh̄Ly (6.56)

L− = Lx − ıLy and knowing

[Lx , L+ ] = Lx (Lx + ıLy ) − (Lx + ıLy )Lx

which is, expanding:

= Lx Lx + ıLx Ly − Lx Lx − ıLy Lx
= Lx Lx − Lx Lx + ı(Lx Ly − Ly Lx )
= +ı(ıh̄Lz )
= −h̄Lz

we derive that and its companion

[Lx , L+ ] = −h̄Lz (6.57)
[Lz , L− ] = −h̄L+ (6.60)

[Ly , L+ ] = h̄Lz (6.58) This last equation tell us that L+ ladders us up, and L−
ladders us down on Lz components. If

[Lz , L+ ] = h̄L+ (6.59) Lz | >= something| >

 23

then or

L− Lz | >= somethingL− | > = L2x + L2y + L + z 2 + ı(−ıh̄Lz ) − L2z

or i.e.
− −
[Lz L + h̄]| >= somethingL | >
L+ L− = L2 + h̄Lz − L2z

Lz L− | >= (something − h̄)L− | > and repeating this for the other order, one has

which means that where ever we were, we are now lower L− L+ = L2 − h̄Lz − L2z
by h̄. If

− Lz | >= Kh̄| >

L =
K with no intentional restriction on the value of K, and

L2 | >= K 0 h̄2 L+ | > (6.61)

K− h L− L+ | > +h̄Lz | > +L2z | >= K 0 h̄2 | >

K−2 h If this particular ket is the highest one, |hi >, then lad-
dering up on it must result in destruction, so we have

K−3 h +h̄Lz |hi > +L2z |hi >= K 0 h̄2 |hi >

K−4 h +(hi)h̄2 |hi > +(hi)2 h̄2 |hi >= K 0 h̄2 |hi >

which means that

K−5 h +(hi) + (hi)2 = K 0

FIG. 2: Laddering Down on the z-component of angular mo- while, working down from the low ket one has
mentum
L+ L− |lo > −h̄Lz |lo > +L2z |lo >= K 0 h̄2 |lo >
+ −
Next, we need an expression for L L , not the com-
mutator. It is: which is
L+ L− = (Lx + ıLy )(Lx − ıLy ) −(lo)h̄2 Lz |lo > +(lo)2 h̄2 |lo >= K 0 h̄2 |lo >

= L2x + L2y + ı(Ly Lx − Lx Ly ) i.e.,

or −(lo) + (lo)2 = K 0

= L2x + L2y + L2z + ı(Ly Lx − Lx Ly ) − L2z which means

−(lo) + (lo)2 = K 0 = +(hi) + (hi)2 = (lo)((lo) − 1) = (hi)((hi) + 1)

which can only be true if (lo) = - (hi). Say (hi) = 7, if (hi) = 7.1 then (lo) can not be achieved by stepping
then K’ = 7*8 and -7(-7-1) which is the same! Note that down integral multiples of h̄.
 24

VII. LAGUERRE POLYNOMIALS or

d2
 
d dy
The radial equation for the H-atom is x 2 + (2 − x) + (α − 1) =0 (7.4)
dx dx dx
h̄2 d2 Ze2
 
2 d `(` + 1)
− + − R(r) − R(r) = ER(r) Doing it again (differentiating), we obtain
2µ dr2 r dr r2 r
d (xy 000 + (2 − x)y 00 + (α − 1)y 0 = 0)
which we need to bring to dimensionless form before pro-
ceeding (text book form). Cross multiplying, and defin- dx
ing  = −E we have which leads to
 2
2µZe2

d 2 d `(` + 1) 2µ y 000 + xy 0000 − y 00 + (2 − x)y 000 + (α − 1)y 00 = 0
2
+ − 2
R(r)+ 2 R(r)− 2 R(r) = 0
dr r dr r h̄ r h̄
which finally becomes
and where we are going to only solve for states with  > 0,
i.e., negative energy states.
 2  2
d d d y
Defining a dimensionless distance, ρ = αr we have x 2 + (3 − x) + (α − 2) =0 (7.5)
dx dx dx2
d dρ d d
= =α Generalizing, we have
dr dr dρ dρ
 2  k
d d d y
so that the equation becomes x 2 + (k + 1 − x) + (α − k) =0 (7.6)
dx dx dxk
 2
2µZe2 α

2 d 2 d `(` + 1) 2µ
α + − R(ρ)+ 2 R(ρ)− 2 R(ρ) = 0
dρ2 ρ dρ ρ2 h̄ ρ h̄
VIII. PART 2
2
which is, upon dividing through by α ,
 2 Consider Equation 7.1
2µZe2

d 2 d `(` + 1) 2µ
+ − R(ρ)+ 2 R(ρ)− 2 2 R(ρ) = 0  d2 2 d `(` + 1

2µZe2
dρ2 ρ dρ ρ2 h̄ ρα h̄ α + R(ρ) − R(ρ)+ 2 R(ρ)−R(ρ) = 0
dρ 2 ρ dρ ρ 2
h̄ ρα
Now, we choose α as (8.1)
if we re-write it as
2µ
α2 = 2  2
d d `(` + 1

2µZe2
h̄ ρ 2 +2 − R(ρ) + 2 R(ρ) − ρR(ρ) = 0
dρ dρ ρ h̄ α
so
(8.2)
 2
d 2 d `(` + 1)

2µZe2 (for comparison with the following):
+ − R(ρ) + R(ρ) − R(ρ) = 0
dρ2 ρ dρ ρ2 ρ
k − 1 x k2 − 1
 
(7.1) 00 0
xy + 2y + n − − − y = 0 (8.3)
To continue, we re-start our discussion with Laguerre’s 2 4 4x
differential equation:
Notice the similarity if ρ ∼ x, i.e., powers of x, x−1 etc.,
d2 y dy
x 2 + (1 − x) + αy = 0 (7.2) 2µZe2 k−1
dx dx *
)n− (8.4)
h̄2 α 2
To show that this equation is related to Equation 7.1 we
differentiate Equation 7.2
x
 2  *
ρ) (8.5)
d y dy
d x dx2 + (1 − x) dx + αy = 0 4
(7.3)
dx
which gives k2 − 1 `(` + 1)
*
) (8.6)
4x ρ
y 00 + xy 000 − y 0 + (1 − x)y 00 + αy 0 = 0
We force the asymptotic form of the solution y(x) to
which is be exponentially decreasing, i.e.,

xy 000 + (2 − x)y 00 + (α − 1)y 0 = 0 y = e−x/2 x(k−1)/2 v(x) (8.7)

 25

and “ask” what equation v(x) solves. We do this in two So, assuming the first part of Equation 8.7, we have
steps, first assuming

y = e−x/2 w(x) k − 1 (k−3)/2

y 0 (x) = x w(x) + x(k−1)/2 w0 (x)
2
and then assuming that w(x) is

w = x(k−1)/2 v(x) and

k − 1 k − 3 (k−5)/2 k − 1 (k−3)/2 0 k − 1 (k−3)/2 0

y 00 (x) = x w(x) + x w (x) + x w (x) + x(k−1)/2 w00 (x) = 0
2 2 2 2

which we now substitute into Equation 8.3 to obtain

k − 1 k − 3 (k−3)/2
xy 00 = x w(x) + (k − 1)x(k−1)/2 w0 (x) + x(k+1)/2 w00 (x)
2 2
k − 1 (k−3)/2
2y 0 =2 x w(x) + 2x(k−1)/2 w0 (x)
2
ny = nx(k−1)/2 w(x)
k−1 k − 1 (k−1)/2
− y =− x w
2 2
x x(k+1)/2
− y =− w
4 4
k2 − 1 k 2 − 1 (k−3)/2
− y =− x w→0 (8.8)
4x 4

or so, substituting into Equation 8.8 we have

x 
xw00 = e−x/2 v − xv 0 + xv 00
 
k−1 x
xw00 + (k + 1)w0 + n − − w=0 (8.9) 4
2 4 
k+1

0 −x/2
(k + 1)w = e − v + (k + 1)v 0
2
   
k−1 x k−1 x
n− − w = e−x/2 n − − v (9.1)
IX. 2 4 2 4
= 0 (9.2)
Now we let
so, v solves Equation 7.6 if α = n. Expanding the r.h.s.
−x/2
of Equation 9.1 we have
w=e v(x)  
x x 00 0 k−1 k+1
v − +xv +(k +1−x)v + n − − v=0
(as noted before) to obtain 4 4 2 2
i.e.,
1
w = − e−x/2 v + e−x/2 v 0
0
xv 00 + (k + 1 − x)v 0 + (n − k)v = 0
2
which is Equation 7.6, i.e.,

1 −x/2 dk y
w00 = e v − e−x/2 v 0 + e−x/2 v 00 v=
4 dxk
 26

and and proceed as normal

dk y ∗ xy 00 = 2a2 x + (3)(2)a3 x2 + (4)(3)a4 x3 + · · ·

y = e−x/2 x(k−1)/2
dxk +y 0 = (2)a2 x + (3)a3 x2 + (4)a4 x3 + · · ·
or −xy 0 = −a2 x − (2)a2 x2 − (3)a3 x3 − · · ·
+γy = γa0 + γa1 x + γa2 x2 + · · · = 0 (10.1)
1 −x/2
w00 = e v − e−x/2 v 0 + e−x/2 v 00
4 which yields
so, substituting into Equation 8.8 we have a1 = −γa0
x 1−γ
a2 = a1

xw00 = e−x/2 v − xv 0 + xv 00 4
 4  2−γ
0 −x/2 k + 1 0 a3 = a2
(k + 1)w = e − v + (k + 1)v 9
2 3−γ

k−1 x
 
k−1 x
 a4 = a3 (10.2)
n− − w=e −x/2
n− − v (9.3) 16
2 4 2 4 or, in general,
= 0 (9.4)
j−γ
aj+1 = aj
so, v solves Equation 7.6 if α = n. Expanding the r.h.s. (j + 1)2
of Equation 9.3 we have
which means
x x

k−1 k+1
 γ
00
v − +xv +(k +1−x)v + n −0
− v=0 a1 = − a0
4 4 2 2 1
(1 − γ)γ
a2 = − a0
i.e., (4)(1)
(2 − γ)(1 − γ)γ
xv 00 + (k + 1 − x)v 0 + (n − k)v = 0 a3 = − a0
(9)(4)(1)
which is Equation 7.6, i.e., (3 − γ)(2 − γ)(1 − γ)γ
a4 = − a0
(16)(9)(4)(1)
dk y ··· (10.3)
v=
dxk
which finally is
and
Πj−1
k=0 (k − γ)
k ∗ aj = − a0
y = e−x/2 x(k−1)/2
d y Πjk=1 (k 2 )
dxk
and
or
Πj−1
k=0 (k − γ) j−γ
−ρ/2 (k−1)/2) aj+1 = −(j − γ) a0 = aj
R(ρ) = e ρ Lkn∗ (ρ) (j + 1)2 Πjk=1 (k 2 ) (j + 1)2

where y ∗ and R(ρ) are solutions to Laguerre’s Equation which implies that
of degree n. Wow.
aj+1 j−γ 1
= ∼
aj (j + 1)2 j
X. PART 3 as j → ∞. This is the behaviour of y = ex , which would
overpower the previous Ansatz, so we must have trunca-
Now, all we need do is solve Laguerre’s differential tion through an appropriate choice of γ (i.e., γ = n∗ ).
equation.

xy 00 + (1 − x)y 0 + γy = 0 XI.

where γ is a constant (to be discovered). We let If γ were an integer, then as j increased, and passed
into γ we would have a zero numerator in the expression
?
(j − γ)
X
y= aλ xλ aj+1 = aj
λ=0 (j + 1)2
 27

and all higher a’s would be zero (i.e., not a power series and therefore Equation 8.3 tells us that
but a polynomial instead)! But
2µZe2
 
∗ k−1
2µ 2µE n − = n∗ − ` =
α2 = 2 h̄2 α
2 =−
h̄ h̄2
implies
so, from Equation 8.6 we have
2µZe2
k2 − 1 α=
= `(` + 1) h̄2 (n∗ − `)2
4

k 2 − 1 = 4`2 + 4` 2µE µZ 2 e4
α2 = − 2 = 4 ∗
h̄ h̄ (n − `)2

k = 2` + 1 i.e.,

so µZ 2 e4
E=− 2
2h̄ (n∗ − `)2
k2 − 1 2` + 1 − 1
= =`
2 2

[1] the only thing that happens in making the complex con- [3] notice the rewriting into momentum language!
jugate is that each ı is changed to −ı (and each −ı → ı). [4] The limits correspond to ϑ = 0 to ϑ = π
[2] the ‘bracket’, hah, hah.