Professional Documents
Culture Documents
PH 502
Wavemechanics & Quantum Physics
COURSE SCRIPT
Syllabus
7.3
8.
Wavefunction
Physical Systems
Schrdingers Equation
Probability and Expectation Value
7.4
7.4
Dirac Notation
Confined Particles
8.1
9.
8.2
8.3
Classical Description
1D Harmonic Oscillator in QM
9.2.1
Solution of Hermites Differential Equation
9.2.2
Correspondence Principle
9.3
9.4
10.2
10.3
10.4
10.5
Angular Momentum
11.2
11.3
11.4
11.5
11.6
Electron Spin
11.5.2 Selection Rules and Atomic Spectra
page: 1
Useful constants
Planck constant:
6.62607610-34 Js = 4.13566910-15 eV s
h
h=
h
2
1.05457310-34 Js = 6.58212210-16 eV s
2.99792458108 m/s
Electron mass:
me
9.10939010-31 kg
510.9991 keV/ c2
Proton mass:
mp
1.67262310-27 kg
938.2723 MeV/ c2
Neutron mass:
mn
1.67492910-27 kg
939.5656 MeV/ c2
Fundamental charge:
1.60217710-19 C
Compton wavelength:
C =
Boltzmann constant:
kB
1.38065810-23 J/K
Vacuum permittivity
8 .854188 10 12 C2 /(Jm )
4 0
1 .112650 10 10 C2 /(Jm )
h
me c
2.4263105810-12 m
combinations:
hc = 1.986410-25 Jm = 1239.8 eV nm
h c = 3.161510-26 Jm = 197.33 eV nm
page: 2
Greek Alphabet/Units
page: 3
alpha
nu
beta
xi
delta
omicron
gamma
pi
epsilon
rho
zeta
s, V
eta
tau
q, J
theta
upsilon
iota
kappa
chi
lambda
psi
mu
omega
sigma
f, phi
Units
Prefix
Exponent Symbol
deci
centi
milli
micro
nano
pico
femto
atto
zepto
yocto
-1
-2
-3
-6
-9
-12
-15
-18
-21
-24
d
c
m
n
p
f
a
z
y
Prefix
Exponent Symbol
deca
hecto
kilo
mega
giga
tera
peta
exa
zetta
yotta
1 = 10-10 m
1
2
3
6
9
12
15
18
21
24
da
h
k
M
G
T
P
E
Z
Y
page: 4
COURSE SPINE
What is that?
This section is designed to give you a coarse outline of the learning outcomes of each section. You
might find this useful for exam preparations and revision.
From the PDF-File you can always link to
Blue colour provides links to the particular section!
The following symbols mean:
Workshop or exercise
7.
To know:
1. Properties of operators, notation
2. Eigenfunctions and eigenvalues, degeneracy, physical meaning
3. Linear combination
4. Orthogonality, normalisation and orthonormality
5. Hermiticity, real eigenvalues
6. Commutators
7. Important operators: x , y , z , p x ,p y , p z , L x , L y , L z , H , K , U
7.3/7.4
To know:
1. Wavefunction, probability interpretation, normalisation
2. Hermitian operators and physical observables
3. Expectation values and deductions
4. Schdinger equations
5. Uncertainty principle (as in (7.4.1))
page: 5
7.2 Operators
page: 6
m12
m22
m32
a( x )
b( x) = M
In quantum mechanics, physical observables (e. g. energy, position, linear and angular momentum)
are represented mathematically by operators. For example the total energy (kinetic energy K plus
, where H is called Hamiltonpotential energy U) is represented by an operator H = K + U
operator or simply Hamiltonian. It is defined as:
7.2 Operators
page: 7
h2 2
H=
+U
2m
We will later derive this formula. The link to experimentally assessable values of the energy (its
average value) of a system described by H and in a state described by a wavefunction (r) is given
.
by its expectation value H
= * (r) H
(r ) dr
H
(A + B ) f = A f
(A B ) f = A f
f
+ B
f
B
[ ]
B
f =A
B
f
A
(7.2.1)
(7.2.2)
(7.2.3)
(7.2.4)
[ ]
2f = A
A
f =A
A
f
A
(7.2.5)
(7.2.6)
7.2 Operators
page: 8
( ) ( )
B
C
f = A
B
C f
A
(7.2.7)
B
f B A
f
A
(7.2.8)
see commutators!
Examples:
f ( x) = f ( x) + c
1) Addition of a constant: A
d
2) Derivative (wrt x): D f ( x) =
f ( x ) = f ( x ) (could be named prime operator or dot
dx
operator for derivatives wrt time)
3) For D operating on f(x) =exp(ax):
d
D f(x) = D exp (ax ) =
exp (ax ) = a exp (ax ) = a f(x)
dx
4) For D 2 operating on f(x) =exp(ax 2):
exp ax 2 = D
d exp ax 2 = D
2 a x exp ax 2
D 2 f(x) = D D
dx
d
= 2a
x exp ax 2 = 2a exp ax 2 + ax 2 exp ax 2 = 2a 1 + ax 2 f(x)
dx
d2
D 2 corresponds to the second derivative D 2 =
dx 2
[ ( )]
( ) [
( ( )) ( ( )
( )]
( ) (
HINT: To calculate pure operator expressions it is good advice to include a virtual function
(e.g. f) to remind you that you have to operate!
d
Example: For the two operators D
and x x (meaning just multiply with x) calculate
dx
:
D x - x D
[x f ] = D
[xf ] = f + xf
D
f = x [ f ] = xf
x D
x - x D
f = f and D
x - x D = 1
hence : D
D x :
x D :
[ ]
(
7.2 Operators
page: 9
Linear Operators:
Almost all operators encountered in quantum mechanics are linear operators.
Definition: A linear operator has to satisfy the following two conditions:
( f + g)= A
f +A
g
A
(af ) = aA
f
A
(7.2.9)
f = af
A
operator
eigenfunction
eigenvalue
(7.2.10)
eigenfunction
( 7. 2. 9 )
g = cb g = dg
cA
7.2 Operators
page: 10
d ax
Example: exp( ax ) = e ax is eigenfunction of D because
e = a e ax
dx
2
2
d ax 2
( e ax is not an eigenfunction, because
e
= 2ax e ax ).
dx
(7.2.12)
Deduction:
Any linear combination of a set of degenerate eigenfunctions is also an eigenfunction.
Proof:
same
eigenvalue
a.
cn n
n =1
g =A
c n n =
n =1
n =1
= a
cn A
n
{
a n
cn n = ag
n =1
qed!
Physical meaning:
If a system is in an eigenstate of observable A (i.e., when the wavefunction is an eigenfunction of the
) then the expectation value of A is the eigenvalue of the wavefunction.
operator A
g=
c
n
(7.2.13)
The new function g is a linear combination of eigenfunctions with the combination coefficients c n
(which might be complex!). (in mathematical terms: The eigenfunctions of an operator form a
complete set, from which other functions can be build up)
Dr. P. Blmler, School of Physical Sciences, University of Canterbury, UK
7.2 Operators
page: 11
( 7 .2 .13 )
c = c A = c{a = b
n
n n
bn
= g~
Hence this generates another function g~ , which also is a linear combination of the eigenfunctions;
however with different coefficients.
A special case is given we use linear combinations of degenerate eigenfunctions (see chapter
11.4):
*
m n
dV = 0
(7.2.14)
2) Normalisation:
A function n is normalised if the integral over all space (V) is one.
*n n dV = 1
(7.2.15)
3) Orthonormality:
If both properties are fulfilled, the functions are called orthonormal.
0 if n m
*m n dV = nm =
1 if n = m
(7.2.16)
7.2 Operators
page: 12
7.2.5 Hermiticity:
Definition of Hermiticity:
is called hermitian, if it fulfils
An operator A
dV = * A
*m A
n
n m dV =
* * dV =
nA
m
* dV
n A
m
(7.2.17)
Deduction:
A hermitian operator has real eigenvalues
(7.2.18)
Physical importance:
All observables (real numbers) correspond to hermitian operators, but -of course- not all hermitians
to observables. (see also expectation values)
(7.2.19)
Deduction:
Eigenfunctions corresponding to different eigenvalues of hermitian operators are orthogonal.
A
2) A
1)
*
m
*
n
and
dV = a
=a
A
m
m m
dV = a
with a n am and a n , am
dV = *m a n n dV = an *m n dV
*
n m m
*
n
dV
IR
7.2 Operators
page: 13
*
*
*
dV * A
*m A
n
n m dV = a n m n dV am n m dV
to be hermitian, the two terms on the left side are identical, hence:
Because we have assumed A
a n *m n dV a*m n *m dV = an a *m
) *m n dV
= 0
*
m n
dV = 0 , which
7.2.6 Commutators:
The key feature of operators -and the reason why they are used in quantum mechanics- is the fact
that they do not generally commute (as we will later see with the uncertainty principle). Another
approach -originally proposed by Heisenberg- used matrices, which do not commute either.
Because of this it is useful to introduce a quantity which expresses this essential property. It is called
the commutator operator [ , ] and defined as follows:
Definition of commutator operator:
,B ] A
B
B
A
[A
(7.2.20)
Notation:
B ] = 0 it is said that A
and B commute!
If [ A,
B ] 0 it is said that A
and B do not commute!
If [ A,
( example above)
Commutator algebra:
,A
] = 0
[A
(7.2.21)
,B
] = [B
,A
]
[A
(7.2.22)
, bB ] = b [ A
,B
]
[A
(7.2.23)
,B
C ] = A
B
C
B
C
A
=A
B
C
B C
A
B
A
C
+B
A
C
[A
B
C
B
A
C
+B
A
C
B
C
A
=A
)
(7.2.24)
+B
]
,B
]C
[A
,C
= [A
B , C
] = [A
,C
]B
+A
[B , C
]
[A
(7.2.25)
7.2 Operators
) (
page: 14
,B
+C
]= A
B + C
B + C A
=A
B + A
C
B
A
C
A = [ A
,B
] + [A
,C
]
[A
(7.2.26)
+ B , C
] = [A
,C
] + [A
,B
]
[A
(7.2.27)
observable
classical
QM-operator
r = ( x, y , z )
r = (x , y , z ) = ( x, y, z )
x
px
t
y
p = p y = mr& = m
t
p
z
z
t
p x
x
h
=
p = p y = ih =
y
i
p
z
z
position r = ( x, y , z )
linear momentum
p = px , p y , pz
angular momentum
L = Lx , Ly , Lz
Lx
L = Ly = r p
L
z
1
p 2 + p 2y + p z2
2m x
kinetic energy, K
K=
potential energy, U
U(x,y,z), e.g. U ( x) =
total energy, E
L x
L = L y =
Lz
E=K+U
k 2
x
2
y z
y p z z p y
z
y
h
z p x x p z = z x
i
z
x p y p
x y
x
y
x
y
h2 2
h 2 2
2
2
K =
=
+
+
2m
2m x 2 y 2 z 2
= U ( r ) e.g. U = k r 2
U
2
2
h 2
H =
+U
2m
h
h
Lz
7.3
7.3 Postulates
page: 15
The postulates cannot be proven. They are axioms of quantum mechanics. However, they are
plausible in the context of defining a fixed framework (mathematical rigorous and without
contradiction) for the description of quantum systems.
Note that different authors very often use different formulations and orders of these postulates!
( r1 , r 2 ,...; t )
with r1, r2,... as the co-ordinates of particle 1, 2, ... and time t.
In this lecture we will deal exclusively with single particles and time-independent problems,
hence the wavefunction simplifies to:
(r )
(7.3.1)
P(r ) d r *(r ) ( r ) d r
(7.3.2)
Normalisation condition:
* ( r ) ( r ) d r 1
(7.3.3)
7.3 Postulates
page: 16
= a
A
n
n
(7.3.4)
) the
Hence, in any physical measurement of the observable A (associated with operator A
only values that will ever be observed are the eigenvalues a, which satisfy eq. (7.3.4).
A particular and extremely important case is the Schrdinger equation, which
corresponds to the total energy as the observable. (cf. heuristic derivation)
in
2b) Every (classical) physical observable corresponds to a linear hermitian operator A
quantum mechanics, because it must yield real eigenvalues!
(7.3.5)
(see 7.2.5 and 7.2.7)
This postulate captures the central point of quantum mechanics: The values of dynamical variables
can be quantized!
with eigenvalue a, then any measurement of the
If the system is a state, which is an eigenstate of A
quantity A will result in value a.
Although measurements must always yield an eigenvalue, the state doesnt have to be an eigenstate
.
of A
(see
An arbitrary state g can be expanded in a complete set of eigenfunctions (or eigenvectors) of A
linear combinations in 7.2.3).
=a
Hence,
i with A
i
i i
n
g=
c
i
(7.3.6)
where n may go to infinity. In this case we only know that the measurement of A will yield one of the
values ai, but we do not know which one in particular. However, we do know the probability that
eigenvalue am will occur. It is the squared magnitude of the coefficient c m
7.3 Postulates
page: 17
A dV
*
dV
*
or for normalised wavefunctions: A
(7.3.7)
A dV
*
(7.3.8)
Deductions:
A dV = a dV = a dV = a
*
=
A
=
c A
c dV =
dV
c *m cn *m A
m
m
n
n
n
m n
m
n
c*m cn *m a n n dV
m
c*m cn an *m n dV
m
7.3 Postulates
page: 18
orthogonal (deduction 2b) and cancel for all n m . In other words, only the terms with n = m
remain. If we further assume normalised wavefunctions, we yield
c*m cm am *m m dV = c*m cm am = cm 2 am
=
A
This is a sum of eigenvalues for that state, weighted by the square magnitude of the combination
coefficients of the probing wavefunction.
Example: Assume a wavefunction which is a linear combination of two eigenstates
with the (real) eigenvalues a and b. Hence,
(eigenfunctions) of the linear, hermitian operator A
= a and A
= b . Calculate the expectation value of A:
= c a a + cb b where A
a
a
b
b
[c + c ] A [c + c ]dV
+c A
] dV = [c + c ] [c a + c b ] dV
= [c + c ] [c A
= ac c dV + bc c dV + ac c dV + bc c dV
14243
142
14243
14243
4 43
4
= *A
dV =
A
a a
*
a a
*
a
b b
*
a b
=1 (normal.)
*
a b
*
b a
= 0 ( 7 .2 .19 )
*
b
= 0 ( 7. 2. 19 )
*
b b
*
b b
=1 (normal.)
= a ca 2 + b cb 2
The average (or expectation) value of the observable A is a weighted average of eigenvalues, where
the weights correspond to the square of the coefficients of the eigenfunctions the wavefunction is
composed from.
Tools:
=
If the mean value of A is defined by A
A dV
*
2
Consequently the mean value of A2 is defined by A
A dV
*
(A A )
(A A ) dV = (A A )(A A ) dV
= (A
A )(A A ) dV = A A A A A + A
= A
2 A A + A dV
= A
dV 2 A A dV + A dV
2
dV
(7.3.11)
( )
2 2 A
2+ A
2 = A
2 A
2 A
A
2
= A
7.3 Postulates
page: 19
POSTULATE 4:
The wavefunction or state function of a system evolves in time according to the
time-dependent Schrdinger-equation:
(r , t ) = ih (r , t )
H
t
(7.3.12)
However, in the following chapters we will exclusively work with time-independent systems, for
which the time-independent Schrdinger-equation has been already introduced in postulate 2 eq.
(7.3.4):
time-independent Schrdinger-equation:
2
h
(r , t ) =
( r , t ) = E ( r , t )
2 ( r , t ) + U
H
2m
(7.3.13)
Comment:
These are the postulates to an extend necessary to understand the course. Other postulates e.g.
concerning the symmetry of the wavefunction or alternatively the Pauli-exclusion principle may be
defined.
page: 20
1 2u
(1)
v 2 t 2
2 ( x)
f (t )
x 2
( x) 2 f ( t )
v2
t 2
(2)
If we introduce one of the standard wave equation solutions for f (t ) such as exp( it ) (the constant
can be taken care of later in the normalisation), we obtain
d 2 ( x )
dx 2
v2
( x)
(3)
Now we have an ordinary differential equation describing the spatial amplitude of the matter wave as
a function of position. The energy of a particle is the sum of kinetic and potential parts
E =
p2
+ U (x )
2m
(4)
(5)
Now we can use de Broglies formula to get an expression for the wavelength
=
h
=
p
h
2m( E U ( x) )
(6)
The term 2 /v 2 in equation (3) can be rewritten in terms of if we recall that = 2 and
= v :
2
v
4 2 2
v
4 2
2
2m(E U ( x) )
h2
(7)
page: 21
When this result is substituted into equation (3) we obtain the famous time-independent
Schrdinger equation
d 2 ( x )
dx 2
2m
h2
(E U ( x) ) ( x)
= 0
(8)
+ U (x ) (x ) = E (x )
2m dx 2
(9)
This single-particle one-dimensional equation can easily be extended to the case of three dimensions,
where it becomes
h2 2
(r ) + U ( r ) ( r ) = E ( r )
2m
(10)
A two-body problem can also be treated by this equation if the mass m is replaced with a reduced
mass .
It is important to point out that this analogy with the classical wave equation only goes so far. We
cannot, for instance, derive the time-dependent Schrdinger equation in an analogous fashion
(for instance, that equation involves the partial first derivative with respect to time instead of the
partial second derivative). In fact, Schrdinger presented his time-independent equation first, and
then went back and postulated the more general time-dependent equation.
page: 22
2 ( r , t )
t 2
h2 2
=
( r , t ) + U (r ) ( r , t )
2m
(11)
where U is assumed to be a real function and represents the potential energy of the system. Wave
Mechanics is the branch of quantum mechanics with eq. (11) as its dynamical law. Note that
equation (11) does not yet account for spin or relativistic effects.
Of course the time-dependent equation can be used to derive the time-independent equation. If we
write the wavefunction as a product of spatial and temporal terms, (r , t ) = ( r ) f (t ) , then eq.
(11) becomes
h2 2
= f (t )
+ U ( r ) ( r )
2m
(12)
ih 2 f ( t )
1 h2 2
=
+
U
(
r
)
( r )
f (t ) t 2
(r ) 2m
(13)
(r ) i h
or
2 f (t )
t 2
Since the left-hand side is a function of f(t) only and the right hand side is a function of r only, the
two sides must equal a constant. If we tentatively designate this constant E (since the right-hand side
clearly must have the dimensions of energy), then we extract two ordinary differential equations,
namely
1 d 2 f (t )
iE
=
2
f (t ) dt
h
and
h2 2
(r ) + U ( r ) ( r ) = E ( r )
2m
(14)
(15)
The latter equation is once again the time-independent Schrdinger equation. The former
equation is easily solved to yield
page: 23
Et
f (t ) = exp i
h
(16)
The Hamiltonian in eq. (15) is a Hermitian operator, and the eigenvalues of a Hermitian operator
must be real, so E is real. This means that the solutions f(t) are purely oscillatory, since f(t) never
changes in magnitude (recall Euler's formula e i = cos i sin ). Thus if
Et
(r , t ) = ( r ) exp i
h
(17)
then the total wave function (r , t ) differs from (r ) only by a phase factor of constant magnitude.
There are some interesting consequences of this. First of all, the quantity (r , t )
is time
(18)
Secondly, the expectation value for any time-independent operator is also time-independent, if
(r , t ) satisfies eq. (17). By the same reasoning applied above,
=
A
( r , t ) dV =
*( r , t) A
(r ) dV
* (r ) A
(19)
For these reasons, wave functions of the form (17) are called stationary states. The state (r , t ) is
stationary, but the particle it describes is not!
Of course eq. (17) represents a particular solution to eq. (11). The general solution to eq. (11) will
be a linear combination of these particular solutions, i.e.
(r , t ) =
E jt
c j j ( r ) exp i
(20)
page: 24
2 A
2 defines the uncertainty of the observables (specifically the root mean
A
(7.4.2)
If two eigenstates are to have simultaneously precisely defined values, their corresponding operators
B
= B A
or using definition (7.2.20): [ A
,B
]= 0
must commute. That is A
hence, A
B
= B A
or A
B
B A
=0
and A
But what we have deduced so far is only the necessary condition. We have not proven the
B
B A
= 0 implies that A
and B have the same eigenfunction. Or if
sufficient part. This is A
= a is also eigenfunction to B for A
B
B A
=0?
A
= a , then B A
= B a = aB
Given: A
page: 25
B
= B A
we can rewrite:
because of A
= A
B
= aB
B A
( ) ( )
B
= a B
A
is member
Now we can appreciate the importance of operators, the fact that they do not generally commute
and why a commutator was defined!!
For a more general description of the uncertainty principle, we want to investigate now the case,
that the two operators do not commute and suppose this property in the most general form:
,B
] = iC
[A
Furthermore, the system is in an arbitrary state described by , which is not necessarily an
or B .
eigenfunction ofA
When we measure the system, we get
=
dV
A
* A
=
B
and
B dV
*
but we are more interested in the spread of values around each of these mean values (or in other
words, how broad the values are distributed, or how certain the measurement is). Hence, we use
another deviation operator
B
A
and b = B
a A
and calculate its commutator:
A
,B
B
]= A
A
B B
B B
[ a , b ] = [ A
B B
A
A
B
+ A
= A
)( ) ( )(A A )
A
A
B
B A
+ A
B
)
B ) (B
B
B
A
= [A
,B
] = iC
=A
and B be hermitian, hence a and b
Because we are only interested in real outcomes, we let A
are also hermitian and we define an integral
I=
(a ib )
) (
2
dV where a i b = a * +i b * * a ib
with as a real but arbitrary parameter. This integral relates the two deviations and measures how
much a is different from b (via ) and it is always greater then zero (because of the square).
I=
(a ib )
dV =
page: 26
I=
= 2 a 2 + b 2 i [ a , b ] = 2 a 2 + b 2 + C
2
C
C
+ b 2
I = a 2 +
0
2
2
2
a
4
This statement is true for all real , but we are searching for the minimum deviation (minimal
uncertainty). Because only scales the first term, the minimum is for this term to vanish or
C
=
.
2
2 a
2
The minimum deviation is hence found for b
2
C
4 a 2
a 2 b 2
A
Because a A
1 2
C
4
( )
A
2 = A
2 A
2 A
A
2
, a 2 = A
(A )2 (B )2
1 2
C
4
B
1 C
A
2
This is a more general form of the uncertainty principle!
,B
] = iC
[A
(7.4.3)
the observables to these operators are complementary. and hence cannot be determined with
arbitrary precision. The highest precision that can be achieved is then given by:
B
1 C
A
2
=
where A
2 A
2 defines the uncertainty of the observables.
A
(7.4.4)
page: 27
Examples:
1) Calculate the uncertainty of space and associated linear momentum (e.g. in x-direction):
h
h
h
h h
h
[ x , p x ] = x
x =x
x
= = ih
i x
i x
i x
i x i
i
h
= h and we directly get eq. (7.4.1): p x
hence C
x
2
2) Now calculate the uncertainty of spatial observable y and linear momentum in
h
h
h
h h
x-direction: [ y ,p x ] = y
y =y
y
=0
i x
i x
i x
i x i
Hence, we can determine both simultaneously and precise. The two operators commute and are
therefore not complementary!
page: 28
dV
*m A
n
(7.5.1)
where the symbol n is called a ket and denotes the state expressed by wavefunction n . The
complex conjugate of the wavefunction *n is denoted by the bra n .
When bras and kets are strung together with an operator between them, like as in the bracket
n the integral of eq. (7.5.1) is to be understood. When the operator is simply multiplication
mA
by 1 (or I ) the operator is omitted from the bracket.
Hence the following conditions from the previous postulates become in bracket notation:
normalisation condition:
n n =1
(7.5.2)
orthonormality condition:
n m = nm
(7.5.3)
hermiticity:
n = nA
m*
mA
(7.5.4)
page: 29
Confined particles:
Solutions
To know:
1. Solutions for particle in 1D box
n =
2
a
nx
sin
,
a
En =
n2 h2
8ma
, n = 1, 2, 3, K
To know:
General recipe for solving QM-problems
1. Determine potential energy
2. Determine boundary conditions
3. Find suitable co-ordinate system
4. Try to separate Schrdinger equation
5. Solve differential equation
6. Determine constants/quantum numbers via boundary
conditions, normalisation, zero-point energy.
8 Confined Particles
8.
page: 30
Confined Particles
U(x)
(8.1.1a)
0 for 0 x a
U (x ) =
(8.1.1b)
for x < 0 and x > a
The particle cannot penetrate the infinitely high potentials at x = 0 and x = a. Therefore, the
wavefunction is only non-zero for the region defined by U(x) = 0, resulting in the Hamiltonian we
already know for free translational motion. The Schrdinger equation is then
with:
d2
dx
2mE
h2
(8.1.2)
(8.1.3)
Considering the boundary conditions (0) = (a) = 0 one gets (0) = A = 0 and
(a) = B sin(ka) = 0. The only non-trivial (particle actually exists) solution is ka being an integer
multiple of .
k=
n
, n 1, 2, 3, K
a
(8.1.4)
n2 h 2 2
2ma 2
n2 h2
8ma 2
, n = 1, 2, 3, K
(8.1.5)
To complete the problem, only B remains to be determined. This is done by normalising the
problem. Because the probability to find the particle in the box must be unity, one simply has to solve
the following equation:
8 Confined Particles
*n ( x ) n ( x)
dx =
sin
( )d x = [
2 nx
a
B2 x
2
( )]
a
a
sin 2 n x
4n
a
0
page: 31
B2a
=
2
(8.1.6)
En [8ma 2 /h 2 ]
n =
2
a
sin
nx
,
a
En =
n2 h2
8ma 2
, n = 1, 2, 3, K
(8.1.7)
yn*(x ) yn (x )
yn(x)
n=4
n=3
n=2
n=1
Fig. 8.2: 1D-confined particle: Graphical representation of the wavefunctions (left) and the probability
densities (right) versus the energy of the state for n = 1, 2, 3 and 4.
8 Confined Particles
page: 32
U(x,y)
a1
a2
0
0
Fig. 8.3: Geometry of the rectangular well. Note the graphical necessity to give infinity a value!
2
2
+
x 2 y 2
+ U ( x, y)
0 for 0 x a1 and 0 y a 2
U ( x, y ) =
for all other
Hence the Schrdinger equation is (cf. argument above (8.1.4)):
with:
2 ( x, y )
x 2
2 ( x , y )
y 2
2mE
h2
( x, y)
(8.2.1a)
(8.2.1b)
(8.2.2)
This partial differential equation can only be solved if we can separate the variables ( see
excursus E1). We now substitute (8.2.2) with y(x,y) = X(x)Y(y) and divide by XY as shown in
excursus E1:
X X + Y Y = 2mE h 2
(8.2.3)
8 Confined Particles
page: 33
Each term must be a constant summing up to the energy term, which we therefore separate as well
E = EX + EY. So (8.2.3) can be separated into two similar parts:
X ( x) =
2mE X
h
X ( x)
and
2mEY
Y ( y ) =
Y ( y)
with E X + E Y E
(8.2.4)
Each of these equations has the form of (8.1.2) for the 1D-problem and even the boundary
conditions are the same! So we can readily write the solution as:
n1 , n 2 = X n 1 ( x) Yn 2 ( y) =
n12
a2
1
h
En 1 , n 2 = EnX + E Yn =
1
2
8m
a)
n 22
a 22
(8.2.5)
b)
n2
n1
1
2
c)
y
a2
d)
x
a1
Fig. 8.4: Wavefunctions for a 2D well with a 2 = 2 a 1: a) y11, b) y12, c) y11, d) matrix with grey
shaded pictures for n 1,n 2 = 1,2,3,4. Dark colours represent negative values, bright colours
positive values.
8 Confined Particles
page: 34
A typical feature of 2D/3D systems becomes apparent when the box is geometrically square (cube
for 3D, see Fig. 8.6), that is a1 = a2 = a. Then the energies are
En 1 , n 2 =
h2
8m a
(n12 + n 22 )
(8.2.6)
That means that yab and yba have exactly the same energy, although the wavefunctions are different.
The property of different states having the same energy is called degeneracy, and the corresponding
states to this energy are said to be degenerate.
The wavefunctions of degenerate states typically can be transformed by simple geometrical
transformation (like 90 rotations). Thus, degeneracy can always be traced to an aspect of the
systems symmetry.
8 Confined Particles
page: 35
a)
b)
c)
Fig. 8.5: STM measurements of quantum corrals: a) Different stages in the assembly of a round
corral, b) the finished round quantum corral with a diameter of 143 , c) a rectangular corral.
8 Confined Particles
page: 36
h2 2
( x, y, z ) + U ( x , y , z ) ( x, y , z ) = E ( x, y, z )
2m
(8.3.1)
0 y a2
and
0 z a3
(8.3.2)
En1 , n 2 , n 3
n x n y n z
8
sin 1 sin 2 sin 3
a1a 2 a3
a1 a 2 a3
(8.3.3)
with n1 , n2 , n3 = 1, 2, 3, K
Problem 8.1: Find degenerate states in Fig. 8.4d. Can you tell by looking at the displayed
probability density? Find other states for n1,n2 > 4!
Problem 8.2: Assuming the rectangular corral in Fig. 8.5.c was scanned at 4.2K and contains a
single electron, can you estimate the distance of the iron atoms using eq. (8.2.5)?
Discuss how realistic your result is and what assumptions might be wrong!
Thermal energy E = k BT.
Comparing the thermal and QM energy one can work out the distance of the STM
tip, assuming a particle in a 3D box.
Problem 8.3: How degenerate is (5,6,7) for a particle in a cubic box? (It is not 6)
References:
1.
2.
Mary L. Boas: Mathematical Methods in the Physical Sciences 2nd ed., Wiley,
New York 1983.
M. F. Crommie, C. P. Lutz, D. M. Eigler: Confinement of electrons to quantum
corrals on a metal surface, Science 262(5131) (1993) 218-220; M. F. Crommie, C.
P. Lutz, D. M. Eigler and E. J. Heller, Physica D 83 (1995) 98-108.
8 Confined Particles
page: 37
Fig. 8.6: The first energy levels of a partic le in a cubical box with a 1 = a 2 = a 3 = a together with the
quantum numbers n 1, n 2, n 3 of the corresponding wavefunctions and the degeneracy of each
energy level. The ground state energy is E0 = E111 = 3h 2/(8ma2).
Excursus 1
E1
page: 38
x 2 y 2 z 2
(E1.1)
However, we want to restrict it to two dimensions for the moment and try a solution of independent
variables, that is:
( x, y ) = X ( x ) Y ( y)
(E1.2)
d 2 X (x )
dx 2
+ X (x )
d 2Y ( y )
dy 2
=0
(E1.3)
Note that we can write this now in ordinary rather than partial derivatives, because X depends only
on x and Y on y!
We now divide (E1.3) by XY to separate the terms:
1 d 2 X (x )
2
X (x )
x4
1442d4
3
depends on x only
1 d 2Y ( y )
Y ( y ) dy 2
14
4244
3
=0
(E1.4)
depends on y only
X = k 2 X
with k 0
and Y = k 2 Y
(E1.5)
(E1.6)
The constant k 2 is called the separation constant and (E1.6) has the following solutions:
sin( kx)
X ( x) =
cos( kx)
and
exp( ky)
Y ( y) =
exp( ky)
(E1.7)
The solutions (or linear combinations) of the initial problem are therefore:
e ky sin kx
ky
e
sin kx
( x, y ) = X ( x )Y ( y ) = ky
e cos kx
e ky cos kx
(E1.8)
9.
9.1
Classical description
Detail
page: 39
Classical Description
To know:
1. Potential energy: U = k x 2 with k = m2 = 22 m 2 x2
2
1D Harmonic Oscillator in QM
Detail (solution of differential equation)
MAPLE exercise
To know:
1. Schrdinger-Equation:
h2 d2
m2 2
( x) +
x ( x) = E ( x)
2m dx 2
2
E = s + 1 h = s + 1 h
2
with s = 0, 1, 2,K
with = 2 m h x = m h x
3. Zero-point energy s = 0 E0 = h2
4. Potential tunneling (QM Classical)
5. Correspondence principle
Exercises and workshops
9.3
page: 40
MAPLE exercise
To know:
h2 2
1. Schrdinger-Equation:
(r ) + 22 m 2 r 2 ( r ) = E ( r )
2m
with r = ( x, y, z )
2. Solutions: Energy and quantum numbers*
(sx +
s y + s z + 3 h
2
with s x , s y , s z = 0, 1, 2, 3, K
Es x s y s z =
with = 2 m h x = m h x
3. Degeneracy
*for identical oscillation frequencies in all 3 dimensions.
9 Harmonic Oscillator
9.
page: 41
Fig. 9.1 The one-dimensional harmonic oscillator. Displacement from equilibrium (x = 0) is denoted by
x.
The classical equation of motion of a particle with mass m is given by Hookes law (cf. Fig. 9.1):
m&x& = kx
or
(9.1.1)
&x& + 20 x = 0
with k = m02 as the spring constant and 0 the natural frequency. The energy of the system is then E = K + U =
In the turning points x 0 the particle comes to a rest (cf. Fig. 9.2) and thus the kinetic energy K = 0
and the energy is entirely potential
E = U = k x02
2
(9.1.3)
For x 2 > x02 , the kinetic energy is negative, which is classically forbidden.
Fig. 9.2: The potential U(x) of the harmonic oscillator. The turning points of the oscillation are at
x = x0, where U(x0) = E.
For the classical case. the probability PCL(x) of finding the particle in the interval dx around the point
x at any time is
Dr. P. Blmler, School of Physical Sciences, University of Canterbury, UK
9 Harmonic Oscillator
page: 42
dt
= 0 dt
T0
2
(9.1.4)
( )
2
with T0 as the period. Combining eqn. (9.1.2) and (9.1.3) gives dd xt = 20 x02 x 2 and thus
P CL ( x ) =
0 d t
1
=
2 d x
2 x 02 x 2
(9.1.5)
x0
Thus N = 2 and
CL
N
( x) d x =
2
x0
x 02
1
x2 2
x0
P CL ( x ) =
N
dx =
2
x 0
N
arcsin
= 1
x0 x
2
x 02
(9.1.6)
(9.1.7)
9 Harmonic Oscillator
page: 43
(9.2.1)
( x) + U ( x) ( x) = E ( x )
2m dx 2
d
dx
(x ) +
8 m
h
[E 2 m x ] ( x) = 0
2
2 2
2E
m
and 2
x
h
h
(9.2.3)
d2
() = ( )
d 2
(9.2.4)
(9.2.2)
Note: The presented solution uses operators. Brehm and Mullin provide you with an
alternative way using series expansion (check their pages 244-250).
The left side of equation (9.2.4) naively reminds you of equation a2-b2 = (a+b) (a-b), but be
careful because we are dealing with operators here. So we try:
d
d
+
=
+ ( ) = + 2
d
d
d2
= 2 2 1
d
(9.2.5a)
and analogously
d2
d
d
+ = 2 2 + 1
d
d
d
(9.2.5b)
a
+
d
and
a +
d
(9.2.6)
d2
() = a a + + 1 ()
d 2
(9.2.7a)
9 Harmonic Oscillator
d2
( ) = a + a 1 ( )
d 2
page: 44
(9.2.7b)
From this equation or the corresponding formulation of the Schrdinger equation in (9.2.4), which is
then (where l() denotes the wavefunction to the energy term )
a a + ( ) = ( + 1) ( )
(9.2.8a)
a +a ( ) = ( 1) ()
(9.2.8b)
a +a a a + = 2
(9.2.9)
Now we take the Schrdinger equation of (9.2.8b) and operate on both sides with a
a a + a () = ( 1) a ( )
(9.2.10)
(a +a 2)a () = ( 1) a ()
a + a a () = ( 1) a ( ) + 2a ()
(9.2.11)
a + a a () = ( 3) a ( )
This corresponds to the Schrdinger equation (9.2.8) for the energy (eigenvalue) -2, or
a + a ( 2 ) ( ) = ( 3) ( 2 ) ()
(9.2.12)
(9.2.14)
thus, min.= 1. As demonstrated above, can take integer values differing by 2, so we conclude that
= 1, 3, 5, 7, ... or with 2s + 1, where s = 0, 1, 2, 3,... Now we
re-substitute using the definition in (9.2.3):
9 Harmonic Oscillator
h
= s + 1 h = s + 1 h
2
2
2
with s = 0, 1, 2,K
page: 45
(9.2.15)
What is left is the determination of the wavefunction, which is now rather simple. Therefore, we take
the wavefunction corresponding to the lowest energy level and all we have to do is to operate on it
with a+ to get the next higher wavefunction and continuation of this procedure gives us any s we
like. (Note that now s rather then is used as a label.)
( )
s ( ) = N s a +
0 ()
= N s dd dd dd K dd
()
( s times) 0
(9.2.16)
with Ns as a normalisation constant. So all left to determine is 0, from which we know it must
satisfy (9.2.13) or
d
d 0
a 0 = + 0 = 0 or
= d ln 0 = d
0
d
2
0 ( ) = N 0 e 2
(9.2.17)
(9.2.18)
1 () exp = 2 exp
2
2
d
2
= N1 [2] exp
2 ( ) = N 2 4 2 2 exp
2
2
3 ( ) = N 3 8 3 12 exp
2
(9.2.19)
the polynomials in squared brackets are known as the Hermite polynomials H1, H2, H3,... which
are summarised in Table 9.1 together with the normalisation constants Ns, which are obtained by
the requirement
() ( ) d = 1 .
E = s + 12 h = s + 12 h
2
s ( ) = N s H s () e 2
with s = 0, 1, 2, K
with = 2 m h x = m h x
(9.2.15)
(9.2.20)
9 Harmonic Oscillator
page: 46
Table 9.1: Energy levels Es, quadratic normalisation constants N s2 and Hermite polynomials H
for various s
s
Es [J]
H s(x)
N2
h
2
3h
2
1 4m
2
h
5h
2
1 4m
8
h
42 2
7h
2
1 4 m
48
h
83 12
9 h
2
1 4 m
384
h
164 48 2 + 12
11h
2
1
4 m
3840
h
4m
h
2 s s!
2 ds
2
( 1) s e
e
d s
....
s
4 m
h
(s + 12 )h
Exercise 9.1:
H s( y ) 2 y H s (y) + 2 s H s ( y) = 0
(9.2.21)
H s +1 ( y ) + 2 s H s 1 ( y ) 2 y H s ( y ) = 0
(9.2.22)
H s ( y ) = 2s H s 1 ( y )
(9.2.23)
Show that the normalisation constants in Tab. 9.1 are correct! (Use:
exp( a
0
2 2
x ) dx =
for a > 0)
2a
9 Harmonic Oscillator
E [hn]
11/2
9/2
7/2
5/2
3/2
1/2
y(x)
page: 47
y*(x ) y(x )
Fig. 9.3: Graphs of the wavefunctions (left) and probability density functions (right) of the 1D
harmonic oscillator for the first 5 energy levels versus dislocation y. The red curves on the
right represent the probability density functions of the classical harmonic oscillator (cf. eq.
(9.1.7)).
On the right side of Fig. 9.3 we observe, that with increasing energy or s the quantum mechanical
probability density PsQM () is becoming more and more similar to the classical probability density
PCL () (red curves in Fig. 9.3). This observation is further illustrated in Fig. 9.4, where the two
functions are compared for E20 = 21h/2.
9 Harmonic Oscillator
page: 48
P(x)
Fig. 9.4: Illustration of the correspondence principle: Black curve: Quantum mechanical probability
density function of a high energy state (20())2. Red curve: Classical probability density
function from (9.1.7).
The principle that for large quantum numbers (e.g. energies) the quantum mechanical probability
distribution function PQM resembles the classical probability distribution function PCL is called
correspondence principle (see Fig. 9.4), it describes the limiting regime where quantum mechanics
meets classical mechanics (c.f. de Broglie relation).
correspondence principle:
lim PsQM ( r ) = P CL ( r )
dt
dr
In other words, the system behaves only classically, when there is sufficient energy.
Exercise 9.2: Demonstrate the correspondence principle of a particle in a 1D-box.
(9.2.24)
9 Harmonic Oscillator
page: 49
The solution of the 2D harmonic oscillator is omitted (see eq. (8.2.3) and the following treatment
of the 3D harmonic oscillator). The results are:
Es x s y = s x + 1 h x + s y + 1 h y = s x + 1 h x + s y + 1 h y
2
with s x , s y = 0, 1, 2,K
2
2
s x s y ( , ) = N s x s y H s x () e 2 H s y () e 2
(9.3.1)
(9.3.2)
with = 2 m x h x = m x h x
and = 2 m y h y = m y h y
N s2 s =
x y
4m
(s +s )
2 x y sx ! s y ! h
1
x +y
(9.3.3)
quantum numbers
( s x , sy )
degeneracy
(01) (10)
(00)
E/hn
energy
levels
9 Harmonic Oscillator
page: 50
0
Fig. 9.6: Energy diagram for a symmetrical (Ux = Uy = U) two-dimensional harmonic oscillator.
Compare with Fig. 9.7.
s2
s1
0
1
2
3
Fig. 9.7: Wavefunctions for a symmetrical (Ux = Uy = U) two-dimensional harmonic oscillator. Black
for negative, white for positive values. Compare the pattern for the degenerate states from
Fig. 9.6.
9 Harmonic Oscillator
page: 51
Fig. 9.8: The first four wavefunctions with identical quantum numbers (00, 11, 22, 33) for a
symmetric potential. (Note: The states with mixed quantum numbers are not shown and not
both potentials are drawn. The red line is an diagonal average).
9 Harmonic Oscillator
page: 52
k 2
r = 22 m 2 r 2
2
(9.4.1)
r 2 = x 2 + y2 + z 2
with
(9.4.2)
h2 2
( r ) + 2 2 m 2 r 2 ( r ) = E ( r )
2m
(9.4.3)
and:
2E
2
=
E X + E Y + E Z X + Y + Z
h h
m
x,
h
m
y,
h
m
z
h
(9.4.4)
(9.4.5)
2
2
2
+
+
(, , ) 2 + 2 + 2 (, , ) = ( , , )
2
2
2
(9.4.6)
As in the 2D and 3D treatment of a confined particle in zero potential (cf. chapter 8.2 and 8.3) we
assume the solution to be a product of functions X,Y,Z, which depend on one spatial co-ordinate
only:
(, , ) = X ( ) Y ( ) Z ( )
(9.4.7)
2 + X
X ( ) d 2
1 dY ( )
+
2 + Y
Y ( ) d 2
1 dZ ( )
2
Z
+
=0
Z ( ) d 2
(9.4.8)
When each term becomes zero the equation is fulfilled, for instance for the first term:
1 dX ( )
2 + X = 0
2
X ( ) d
dX ()
d
2 X () + X X () = 0
dX ( )
d
(9.4.9)
2 X () = X X ( )
This the same expression as eq. (9.2.4), of which we know the results (eq. (9.2.15) and (9.2.16)).
Hence, the solution of the Schrdinger equation for the 3D harmonic oscillator is given by:
Dr. P. Blmler, School of Physical Sciences, University of Canterbury, UK
9 Harmonic Oscillator
page: 53
Es x s y s z = s x + 1 h + s y + 1 h + s z + 1 h = s x + s y + s z + 3 h
2
2
2
2
with s x , s y , s z = 0, 1, 2, 3, K
2
2
2
s x s y s z (, , ) = N s x s y s z H s x ( ) e 2 H s y () e 2 H s z ( ) e 2
with = 2 m h x , = 2 m h y , and = 2 m h z
N s2 s s =
x y z
1
2
(s x + s y + s z )
3
4m 2
sx ! s y !sz !
Fig. 9.9: Energy diagram for the first 4 states of a symmetrical 3D harmonic oscillator.
(9.4.10)
(9.4.11)
(9.4.12)
(9.4.13)
page: 54
Classical Description
To know:
1. Linear / Rotational Motion:
2. Reduced Mass
3. Boundary Conditions
To know:
1. Schrdinger-Equation in circular co-ordinates
2. Solutions and quantum numbers
1
e im
2
m ( ) =
Em = m
2 h
2I
and
with m = 0, 1, 2, 3, K
Lm
z = mh
3. Uncertainty and m = 0
4. Shape of wavefunctions, auxiliary functions, Lz
page: 55
MAPLE exercise
To know:
1. Schrdinger-Equation in spherical co-ordinates.
2. Integration limits for spherical co-ordinates.
3. Solutions (spherical harmonics)*
h2
energy El = l (l + 1)
2I
4.
5.
6.
7.
* not in detail, but their composition and dependence (the most simple you could remember!)
page: 56
To know:
1. Linear definitions, eg. . L x = y p z z p y = hi y z z y
2
2 2
2. L z = i and L = h
applications to spherical harmonics
3. Spatial quantization: L = h l (l + 1)
10 Rotational Motion
page: 57
Cartesian: r = x,y
circular: r, f
mass
moment of inertia:
I = mi ri2
i
linear velocity:
v x = x& , v y = y&
angular velocity:
v rv v
=
, = 2
r2
v = r&
momentum
force:
kinetic energy:
linear momentum:
p = mv
L = r p , L = I
linear force:
F = p& = ma
torque:
T = r F = I &
K=
angular momentum:
1 2 p2
mv =
2
2m
K=
1 2 L2
I =
2
2I
The typical case, where we wish to apply the following treatment, is the rotation of a (at least)
diatomic molecule around its centre of mass (Fig. 10.1).
z
L
r0
m
a)
b)
Fig. 10.1: a) Rotation of a diatomic molecule around its centre of mass (CM). b) Without loss of
generality the system can be described by the rotation of a reduced mass m.
Dr. P. Blmler, School of Physical Sciences, University of Canterbury, UK
10 Rotational Motion
page: 58
First we want to reduce the problem to a single mass problem. Since the rotation is around the
centre of mass. In good agreement with nature we can assume point-masses:
m1 r1 = m2 r2
m1 r1 = m2 ( r r1 ) or r1 =
hence:
and r0 = r1 + r2
m2
m1
r0 and r2 =
r
m1 + m2
m1 + m2 0
I = m1r12 + m2 r22
(10.1.1)
(10.1.2)
(10.1.3)
m12
m m (m + m1 ) 2
m1 m2
r02 + m2
r02 = 1 2 2
r0 =
r02
2
2
2
(
m
+
m
)
(m1 + m2 )
(m1 + m2 )
(m1 + m2 )
1
2
I = r02
m22
with =
(10.1.4)
m1 m2
as the ' reduced mass '
(m1 + m2 )
Analogously we can reduce the masses of a many-body problem, without loss of generality!
p2
L2
L2
=
=
2m
2I
2r02
(10.2.1)
h
hr
or L =
(10.2.2)
The consequence of (10.2.2) is that the angular momentum is inversely proportional to the
wavelength of the particle. Thus, the shorter the wavelength the higher the angular momentum, the
higher the energy. If can take arbitrary values, after one round the particle-wave will have a
different amplitude at the same position. Therefore, the corresponding wavefunction will have
different values at the same position, which is a violation of the first QM-postulate. Hence, only such
are quantum mechanically allowed, which form standing waves on the orbit (see Fig. 10.2). In
other words has to be an integer multiple of the path length:
2r0 = m
(10.2.3)
10 Rotational Motion
page: 59
combine eqs.(10.2.1)-(10.2.3):
Em =
L2 h 2 r02 h 2 r02 m2
=
=
=
2 I 2 I 2
82 I r02
h2 m 2
with m = 1, 2, 3, K
2I
(10.2.4)
Note: Careful that you do not confuse m with mass. At the moment we might call it the azimuthal
quantum number !!
x 2
y 2
h2
(10.3.1)
However, Cartesian co-ordinates are not a suitable co-ordinate frame. 2D-polar or circular coordinates are better suited for the problem, because r = r0 is a constant. Therefore we transform eq.
(10.3.1) into circular co-ordinates (r,) (see Excursus 2: eq. (E2.16)) and replace the mass m by
the reduced mass .
10 Rotational Motion
page: 60
The Schrdinger equation for the rigid rotator in circular co-ordinates is therefore:
1 2
1
2E
(r , ) =
(r , )
r ( r, ) + 2
2
r r r
r
h2
2 (r , ) =
(10.3.2)
This looks more complicated then (10.3.1), but since r is not a variable but a constant r0 (supplied
by the rigidity of the rotators axis), it simplifies to (using eq. (10.2.4) to express E as L2 and m,
eq. (10.1.4) for I):
1 d2
r02
d
() =
() =
2E
h2
2Er02
2
h
()
(10.3.3)
() =
L
h
() = m 2 ()
We acknowledge that the partial differentials became ordinary, and that this equation has been
solved already for the free one-dimensional particle (see part 5c of John Dores part). At the
moment we still consider m as unrestricted in value despite our previous considerations.
The solutions to (10.3.3) are (for generality we allow complex solutions):
() = Ae im + Be im
(10.3.4)
(10.3.5)
Aeim + Be im = Ae im ( 2 + ) + Be im ( 2 + )
Ae
im
+ Be
im
= Ae
im 2
im
+ Be
im 2
im
(10.3.6)
This can only be fulfilled if m is an integer (because e im 2 = 1 for m = 0, 1, 2, ...), which also
satisfies our previous considerations, and the wavefunction reproduces itself on the next circuit.
Energy and angular momentum are quantized.
Since we allow m to have negative and positive integer values, let us compare the cases m and -m in
eq. (10.3.4). Obviously they correspond to the same state when A and B are swapped. The
corresponding energies (Em m 2 ) are the same, so the states are degenerate. However, the
angular momentum L is pointing in opposite directions (+z and -z). Hence, the two states correspond
to rotation in opposite directions. Because the same considerations hold for the wavefunction with
either A = 0 or B = 0, we want to eliminate this unnecessary ambiguity and rewrite eq. (10.3.4) as:
m () = Aeim
(10.3.7)
10 Rotational Motion
page: 61
This also allows the determination of A by normalisation (Note the integration limits for ):
2
* () () d =
0
Hence, A = 1
im
im
d = A
1 d = 2 A
=1
(10.3.8)
and we could readily write down the solution. However, there is still the unsolved
case m = 0! For the discussion of a particle in a box we have used two arguments, as to why the
corresponding state doesnt exist: 1) It wouldnt agree with Heisenbergs uncertainty principle
(since the particle was confined, it has had a finite position range, so the momentum/energy must not
become zero). 2) The corresponding wavefunction (cf. eq. (8.1.7)) would become zero, which
would violate the first postulate.
Lets now consider the same arguments for the case of the rigid rotator. The second can be solved
directly and 0 ( ) = 1 0 . It is constant but not zero, so it exists! The first argument based on
2
the fact, that the particle is confined. Actually it is not, it moves free on its circular path in zero
potential. The idea that it is constrained to the ring originates from the viewpoint of its path in 2DCartesian co-ordinates, but we have solved the problem in circular co-ordinates, and we will find
expressions for its wavefunction for any . (The Schrdinger equation is identical with that of an
unconstrained particle in 1D Cartesian co-ordinates.)
Therefore, we allow m = 0 as an eigenvalue of the Schrdinger equation1.
Normalised solution of the Schrdinger equation for the rigid rotator
m () =
1
e im
2
Em = m 2
h2
2I
with m = 0, 1, 2, 3, K
(10.3.9)
Lmz = mh
(10.3.10)
and
From the solutions we directly observe, that each state is doubly degenerate due to rotation in
opposite directions (Lz pointing up- or downwards) but with the same speed (energy). Except for
m = 0 where Lz is zero and the question of direction doesnt arise.
There are several ways to depict the wavefunctions. A problem arises, because they are complex
e im = cos( m) + i sin( m) . We get somewhat around the problem by plotting the real part only
(since the imaginary is just 90 phase shifted) as done in Fig. 10.3.
The treatment in eq. (10.2.4) might still be treated as correct, because we didnt explicitly consider circular coordinates at that point. However, it has no particular meaning, because we didnt solve the Schrdinger equation
for 2D-Cartesian co-ordinates.
10 Rotational Motion
page: 62
Fig. 10.3:
Graphical representation of the
wavefunction (real part) for the rigid
rotator
(m = 0,1,2,3).
The
wavefunctions for m = -1, -2, -3
look identical, but the direction of
rotation alters (small arrow) and the
angular momentum (bold arrow)
points down.
However, this gives a wrong picture, because it suggests, that the particle is distributed non-uniformly
over the circle. As shown in eq. (10.3.8) the probability distribution is independent of and hence
uniform. Hence, when the particle is in a state of definite angular momentum its distribution is
completely uniform!
A better representation is shown in Fig. 10.4, where an auxiliary function is constructed by
connecting each point of the perimeter to the centre and to mark on this radius a length proportional
to the amplitude of the wavefunction (see Fig. 10.4a). In other words, this is the plot of the amplitude
in polar co-ordinates. Hence, for 0 we would obtain a circle.
This procedure allows us to project the wavefunction into a plane and the overlay of its real and
imaginary part (see Fig. 10.4c and d). The sum of their magnitudes than somehow resembles the
probability distribution (note overlapping parts in Fig. 10.4c/d, they correspond to the gaps from a
circle).
10 Rotational Motion
b)
a)
+
_
+ _
+
_
_
c)
page: 63
d)
Fig. 10.4: a) Construction of the auxiliary function from the real part of 1. b) Same for 2. c)
The auxiliary function from a). Dark colour denotes the real part, brighter grey the
imaginary part. d) Same as c) for 2 or b) respectively.
10 Rotational Motion
page: 64
Hence, we have to express the Schrdinger equation in 3D polar (spherical) co-ordinates with a
wavefunction that depends on both spherical angles (and the radius), y(r, , ). Fig. 10.6 shows
their interdependence.
Fig. 10.6: Relation between Cartesian and spherical co-ordinates and their range.
In the excursus to this chapter it is shown, how the Laplacian operator is converted into spherical coordinates (see eq. (E2.2.24))
2 =
1 2
1
1
2
r
+
sin +
r 2 sin 2 2
r 2 r r
r 2 sin
Following the same assumptions as in section 10.3, we can directly derive the Schrdinger
equation (r = r0 = const., mass: m ):
1
2
2E
sin ( , ) +
( , ) =
(, ) (10.4.1)
r02 sin
r02 sin 2 2
h2
1
After multiplication with r02 sin 2 and setting r02 = I the equation becomes:
2
2 IE
sin sin (, ) +
( , ) =
sin 2 ( , )
2
2
(10.4.2)
(10.4.3)
10 Rotational Motion
page: 65
(10.4.4)
We rearrange the terms that depend on or on either side of the equation and substitute the
energy term by 2IEh 2 :
sin d
d
1 d2
2
( )
sin ( ) + sin =
() d
d
( ) d 2
(10.4.5)
To be solvable each side must correspond to a constant! We take a closer look at the right term and
recognise it as identical to eq. (10.3.3) when choosing the separation constant to be m2. So we
have to solve two equations depending on and :
d2
d
and
( ) = m2 ( )
sin d
d
2
2
sin ( ) + sin = m
() d
d
(10.4.6)
(10.4.7)
The first we have already solved in the previous section and can copy its solution from (10.3.9):
m ( ) =
1
e im with m = 0, 1, 2, K
2
(10.4.8)
Equation (10.4.7) however is new and more complicated. First we multiply with ( ) / sin 2 and
rearrange:
1 d
d
m 2
( ) = 0
sin ( ) +
2
sin d
d
sin
(10.4.9)
Now we want to get rid of all the sin terms. Because we have to take its derivative we choose for
substitution a new function
P( ) = cos
(10.4.10)
and with d = sin d and the Pythagoras relation 2 = 1 sin 2 equation (10.4.9) becomes:
d
dP ( )
m2
1 2
+
P ( ) = 0
2
d
d
1
(10.4.11)
10 Rotational Motion
(1 2 )P () 2P () + 1 m 2 P ( ) = 0
2
or
page: 66
(10.4.12)
This type of equation is called associated Legendre differential equation (see e.g. [1] page 504)
with eigenvalues = l ( l + 1) and has solutions Pl m () for l m l and l as an integer called
the associated Legendre functions. They have the following general form of (Legendre) polynomials
of degree l and order m:
( 2)
P m ( ) = ( 1) m 1
m
2
2l l !
l
(
2 1)
d l + m
dl + m
(10.4.13)
and can be normalised by the following integral over the entire range (N as the normalisation
constant):
1
(N P
2
2 ( l + m )!
(
)
d = N 2
=1
l
2l + 1 (l m )!
m
Nlm =
(2l + 1)(l m )!
(10.4.14)
2(l + m )!
the harmonic oscillator. However the solution is quite lengthy due to the interdependence of m and l)
We have seen that the dependent part of the Laplacian operator in spherical co-ordinates
produces differential equations of the Legendre type. It is common practice and useful to denote the
part with the angular derivatives of the Laplacian by a new operator, which we will call Legendre
operator (or Legendrian)2:
2 =
1
1 2
sin +
sin
sin 2 2
(10.4.15)
The Legendrian measures the curvature of a function relative to the surface of a sphere. Thus its application to a
wavefunction that represents a sphere is zero, because a sphere in spherical co-ordinates is flat.
10 Rotational Motion
page: 67
with
2IE
h2
( , ) = ( ) ( )
and
(10.4.16)
1
eim with m = 0, 1, 2, K
2
(10.4.17)
Because m is limited by l, it is often referred to as ml. As eigenvalues for the associated Legendre
equation we found with the condition that l is an integer:
=
2IE
h2
= l (l + 1) with l = 0, 1, 2, K
(10.4.18)
(2l + 1)(l m )!
4(l + m )!
(10.4.19)
l m l
This lengthy expression is typically abbreviated by (again) new functions Yml ( , ) called spherical
harmonics. We summarise the mathematical treatment:
Particle rotating on a sphere:
Schrdinger equation:
2 Yml (, ) = l ( l + 1) Yml ( , )
Yml ( , ) =
Energy:
El = l (l + 1)
Quantization conditions:
l = 0, 1, 2, K and
(10.4.20)
(2l + 1)(l m )!
4 ( l + m )!
Pml (cos ) e im
h2
2I
l m l
(10.4.21)
(10.4.22)
(10.4.23)
10 Rotational Motion
page: 68
Table 10.2: Composition of the spherical harmonics Yml ( , ) for the quantum numbers
l 3 and -l m +l
l
El *
Pml () **
Pml (cos )
l
Nm
***
cos
1
2
m 1 2
m sin
1
4
1
2
1
2
m 3 1 2
m 3 cos sin
1
12
30
31 2
3 sin 2
1
24
30
12
1
2
12
12
12
(3 2 1)
(
1
2
(3 cos 2 1)
(5 3 3 )
1
( 2 5 sin 2 ) cos
2
m 32 5 cos2 1 sin
1
12
21
15 sin 2 cos
1
120
210
m 15 sin 3
1
120
35
( )
32
m 15(1 2 )
15 1 2
1
4
1
2
m 3 5 2 1 1 2
Yml ( , )
3
4
cos
m 83 sin e i
5
16
(3 cos 2 1)
sin 2 e i 2
7
( 2 5 sin 2
16
21
64
) cos
(5 cos 2 1)sin ei
105
sin 2 cos e i 2
32
35
sin 3
64
e i 3
* in multiples of h 2 (2 I ) = l(l+1)
** from eq. (10.4.13)
***from eq. (10.4.14)
So far for math, but what do the spherical harmonics look like? We have solved a problem for a
particle on a sphere, and the result is a (sometimes) complex wavefunction on the surface of the
sphere. How can one visualise this? The problem is identical with projecting the globe of the earth
onto a 2D-map! We are used to projections of the earth onto a cylinder, so that the poles become
as large as the equator, although they are points (pseudo Mercator-projectons). In our case the
latitudes correspond to from 0 (south pole) to p (north pole) and the longitudes to from 0
(Greenwich) to 2 (again Greenwich, so UK is at the left and right edge). Although this is a rather
contorted view of reality, you might find it instructive (see Fig. 10.7).
These wavefunctions projected on a sphere are more realistic, but you are restricted to the view on a
single hemisphere (see Fig. 10.8).
Another possibility is to plot auxiliary functions (like in the 2D case, cf. Fig. 10.4), but this time we
will plot the amplitude of the wavefunctions in spherical co-ordinates (see Fig. 10.9). Like in the
2D-case this disables us to distinguish the sign of , but it is a more realistic representation of the
wavefunction. (To understand the relation between Fig. 10.7, 10.8 and 10.9, find the nodes -now
nodal planes- in Fig. 10.7 and mark them on the projections onto a globe, as in Fig. 10.8.
Realise that is spanning the entire diameter, so that nodes separated by -or half the image widths
Dr. P. Blmler, School of Physical Sciences, University of Canterbury, UK
10 Rotational Motion
page: 69
in Fig. 10.7- correspond to the same nodal plane in Fig. 10.8 and 10.9. For a better
comparison switch to graph 10.10, where all three representations are compared next to each
other.)
Finally we want to calculate the probability density function:
QM
Plm
( , ) = Yml * (, ) Yml (, )
lm* () *m ( ) lm ( ) m ( )
(10.4.24)
Consulting eq. (10.4.8) We directly see that *m ( ) m ( ) = 1 and from eq. (10.4.13) that
lm* ( ) = lm ( ) , because lm () is always real. Hence, (10.4.24) becomes:
QM
Plm
( ) = lm ( )
(10.4.25)
We have just proven that the probability density function is (like in the 1D case) homogeneous with
QM
respect to . In other words, Plm
() is not homogeneous with respect to .
QM
Fig. 10.10 shows Plm
() in the three different visualisation modes we have used (map
corresponds to Fig. 10.7, globe to 10.8, aux. to 10.9). Note the similarity of the probability
density functions along the diagonal of Fig. 10.10! The auxiliary functions look almost alike (e.g.
QM
QM
P1QM
1 () P2 2 () P3 3 () ), however they are not identical! They have the same symmetry,
but since they correspond to different energy levels, their curvature along must increase with the
quantum numbers, as it can be clearly seen form the map-graphs! Also note that the nodal planes
now correspond to black colour and since we have taken the square of the wavefunction, we might
no longer consider them as planes but nodal cones (planes are then cones for =/2 and 0)!
The last question we are going to consider is the energy term scheme (see Fig. 10.11) for the
particle on a sphere and the degeneracy of its levels. As shown by eq. (10.4.22), the energy of the
system doesnt depend on m, although m is always spanning a range -l m +l resulting in the
same energy. Hence the degeneracy of each energy level is 2 times l (each m is covering the range of
l twice) +1 (for the zero):
Degeneracy of El
is
2l+1
for l = 0, 1, 2, ....
(10.4.26)
(10.4.27)
h
4cI
(10.4.28)
10 Rotational Motion
Yl m ( , )
m=0
Real
m = +1
Real
Imag
m = -1
Real
Imag
page: 70
m = +2
Real
Imag
m = -2
Real
Imag
m = +3
Real
Imag
m = -3
Real
Imaginar
y
l=0
l=1
l=2
l=3
Fig. 10.7:
Cylindrical projections of the wavefunctions Yl m ( , ) of a particle on a sphere (pseudo Mercator-projections). The vertical axis is from 0 to
, the horizontal axis is from 0 to 2. White colours depict positive and black negative amplitudes. Nodes are in mid-grey! (Note that this is a
rather distorted view of the wavefunctions).
10 Rotational Motion
Yl m ( , )
m=0
Real
l=0
m = +1
Real
Imag
m = -1
Real
Imag
page: 71
m = +2
Real
Imag
m = -2
Real
Imag
m = +3
Real
Imag
m = -3
Real
Imaginar
y
f=0
q=0
l=1
l=2
l=3
Fig. 10.8:
Projections of the wavefunctions from Fig. 10.7 on a sphere. The view is from = 30 and = 270 from 0 to 2 (on the globe of the earth this
would correspond to somewhere over the southern Pacific). White colours depict positive and black negative amplitudes. Nodes are in mid-grey!
Yl m ( , )
m=0
Real
10 Rotational Motion
m = 1
Real
page: 72
m = 2
Imaginary
Real
m = 3
Imaginary
Real
Imaginary
l=0
l=1
l=2
l=3
Fig. 10.9:
Projection of the magnitude of the wavefunctions Yl m ( , ) in spherical co-ordinates, generating a 3D-version of the auxiliary functions as
described in Fig. 10.4.
QM
Plm
( )
map
l=0
10 Rotational Motion
globe
page: 73
m = 1
aux.
map
globe
m = 2
aux.
map
globe
m = 3
aux.
map
globe
aux.
f=0
q=0
l=1
l=2
l=3
Fig. 10.10: The probability density functions of the particle on a sphere. As shown in eq. (10.4.25) it is only a function of . The three different
visualisations are directly compared. Map corresponds to the pseudo Mercator-projection of Fig. 10.7. Globe is a plot of the wavefunction on
the sphere as in Fig. 10.8, and aux. is the plot of the amplitude in spherical co-ordinates (cf. Fig. 10.9). Note that the levels gray scale now
from 0 to max, so the nodal planes/cones are black!
Dr. P. Blmler, School of Physical Sciences, University of Canterbury, UK
10 Rotational Motion
quantum
number
degeneracy
l=4
12
l=3
l=2
2
0
l=1
l=0
3
1
E
2
[2I/ h ]
20
Fig. 10.11:
l (+1)
l
page: 74
2l+1
ey
y
py
yp z zp y
ez
z = zp x xpz
pz
xp y yp x
(10.5.1)
Also from Tab. 10.1 we see that the kinetic energy can be expressed in terms of angular momentum.
Since we didnt have to consider potential energy in section 10.4, we can express the total energy
as kinetic energy.
E =K=
with
L2
2I
(10.5.2)
(10.5.3)
In the context of eigen-equations and the entire context of quantum-mechanics, it is more useful to
express angular momentum Lx, Ly, Lz and L2as operators. The straightforward way is to rewrite
(10.5.1) for each component of L by the operators of position and linear momentum:
L x = y p z z p y = hi y z z y
(10.5.4)
(10.5.5)
L z = x p y y p x = hi x y y x
(10.5.6)
L y = z p x x p z = hi z x x z
10 Rotational Motion
page: 75
As before we rewrite these equations in polar notation, which is more suitable to the problem. From
Fig. 10.6 we get the expressions for the Cartesian co-ordinates:
x = r sin cos
y = r sin sin
z = r cos
(10.5.7)
Hence,
L x = h r sin sin r cos
i
z
y
L = h r cos r sin cos
y
(10.5.8)
Now, we express the partial derivatives of the polar co-ordinates by the Cartesian (you might use
the recipe of E2.2 - the other way around) or work it out:
=
r
x y z
+
+
= sin cos + sin sin + cos
r x r y r z
x
y
z
x y z
+
+
= r cos cos + r cos sin r sin
x y z
x
y
z
x y z
+
+
= r sin sin + r sin cos +
x y z
x
y
(10.5.9)
After some reformulation (for which we do not need r , because L must not depend on r), we
get:
L x = h sin cot cos
i
L y = h cos cot sin
i
L = h
z
(10.5.10)
Check the result by substituting the expressions of (10.5.9) into (10.5.10) and try to derive (10.5.8).
This is easier then the other way around ( cot = cos / sin ).
We are left with deriving an equation for L 2 = L (L ) = L x (L x ) + L y (L y ) + L z (L z ) . When we do
by using the expressions in (10.5.10), this becomes quite lengthy, but try it. The approach here is
different by looking at eq. (10.5.2). If the system has no potential energy (like in section 10.4),
the total energy is purely kinetic and expressed by L2. In operator notation we conclude, that the
following eigen-equation must be valid for L 2 :
L 2 (, ) = 2 IE ( , )
Dr. P. Blmler, School of Physical Sciences, University of Canterbury, UK
(10.5.11)
10 Rotational Motion
page: 76
(10.5.12)
we find (cf. (10.4.15), the same result we would get using the explicit expressions in (10.5.10)) :
2
1 2
h 1
L 2 = h 2 2 =
sin
i sin
sin 2 2
(10.5.13)
Since we have solved eq. (10.5.12) in section 10.4 (see (10.4.20)), all we have to do is to
substitute:
L 2 Yml (, ) = h 2 2 Yml (, ) = h 2 l (l + 1) Yml ( , )
with l = 0, 1, 2, K (10.5.14)
i
i
h
= Nlm ( )
m e im = hm Yml ( , )
2
L z Yml (, ) =
N lm ( )
e im
2
with
(10.5.15)
l m l
We have just seen that the magnitude of the total angular momentum L is quantized having
eigenvalues h l (l + 1) and that additionally L z is quantized by eigenvalues hm .
As long as we do not assign the problem to specific Cartesian direction - for instance by applying an
additional (magnetic) field- there is nothing specific about the z co-ordinate. The simple dependence
on occurs because we had chosen it to be perpendicular to z. We will later see, that this choice
makes the x and y component of the angular momentum non-measurable, as their contribution to the
probability density function vanishes (see eq. (10.4.25)). However, if we would have chosen x or
y to lie perpendicular to the angular momentum relative to them would turn out to be quantized,
and the remaining other would be non-determinable.
These results are really important, because we can now conclude, that any system with a spherically
symmetric potential U(r) -hence a central force problem, which can be separated from the angular
terms- has solutions for its angular terms as above.
Graphically we can summarise these considerations. As an example we consider l = 2 and hence
L = h l (l + 1) = h 6 . If a preferred direction exists (as in the presence of an external field) or
we simply chose it to be z, L z = hm with m = -2, -1, 0, 1, 2. Thus 5 different directions of the
angular momentum vector exist relative to the z-axis, as visualised in Fig. 10.12. Note that a
Dr. P. Blmler, School of Physical Sciences, University of Canterbury, UK
10 Rotational Motion
page: 77
completely parallel or anti-parallel alignment to the z-axis is not possible, since its maximum
projection on the preferred axis is lh , while having a length of h l (l + 1) > hl . (Only for very large
l ( l + 1) l , so that the object is able to rotate solely
z
m = +2
2h
m = +1
h
h 6
0
h
m=0
m = -1
2h
m = -2
Excursus 2
Excursus 2:
page: 78
=
+
x 2
y 2
The Laplacian is
(E2.1)
with r 2 = x 2 + y 2
(E2.2)
r = x2 + y2
or
y
sin
=
= tan
x
cos
and
y = r sin
and
or
= arctan
(E2.3)
y
x
(E2.4)
(E2.5)
f
f r
f
=
+
x
r x
x
(E2.6)
df =
hence
and
f
f r
f
=
+
y
r y
y
(E2.7)
12
1 2
r
2
1
=
x + y2
= 2x x 2 + y 2
=
x
x
2
x
2
x +y
x
= cos
r
y
y
1
y
y
sin
=
arctan
=
=
=
=
2
x
x
x
r
x2 1 + y
x2 + y 2
r2
2
(E2.8)
(E2.9)
r
=
y
y
= = sin
r
x2 + y2
1
=
y
x
1
1+
x2
x
x2 + y2
(E2.10)
cos
r
(E2.11)
Excursus 2
page: 79
Therefore, we can express the first partial derivatives of f by substituting (E2.8/9) into (E2.6) and
(E2.10/11) into (E2.7):
f
f
f sin
f
sin f
fx
=
cos
= cos
x
r
r
r
r
(E2.12)
f
f
cos f
f y
= sin
+
y
r
r
(E2.13)
f x
f x r
f x
=
+
r
r x
x
(E2.14)
f y
f y r
f y
=
+
y
r y
y
(E2.15)
We have to be careful, because f x and f y depend on r and ! Substituting (E2.12) and (E2.8/9)
into (E2.14):
2 f
x 2
f sin f
cos
r
r
f sin f
cos
cos
r
r
2 f
sin f
sin 2 f
= cos
+
r r
r 2
r 2
cos
f
2 f sin cos f
sin 2 f
sin + cos
r
r
r
r 2
= cos 2
2 f
r 2
sin
sin
+
+
+
r r
r
r
r
r
r2
r2
2
f
cos f
f
cos f cos
sin
+
sin +
sin
+
r
r
r
r
r r
= sin 2
2 f
r 2
+
+
r r
r
r
r
r
r2
r2
2
Excursus 2
page: 80
The Laplacian is then given by the sum of the last two equations:
2
2
2
2
2
2
2
2
2 f = f + f = sin 2 + cos 2 f + sin + cos f + sin + cos + f
r
r
x 2
y 2
r 2
r2
2
x 2 y 2
r 2
1
1 2
+
r r r 2 2
1 2
1
r + 2 2
r r r
r
(E2.16)
where we used a nice abbreviation for the two terms that depend on r.
(
(
)
)
h h
x
r
1
y = h h h hr h
r (h h )
z
r
( )
( )
( )
( )
(E2.17)
2
2
2
y
hr2 = x + + z
r
r
r
with
2
2
2
y
h2 = x + + z
y
h2 = x + + z
(E2.18)
2
and
2 = 2 + 2 + 2 =
2
2
2
x
1
hr h h
h h
hr h
h h
+ r
+
r hr r h h
(E2.19)
( x, y, z )
=
( r , , )
x
r
y
r
z
r
(E2.20)
x
= hr h h
(E2.21)
Excursus 2
page: 81
( )
2
2
y
hr2 = x + = (cos )2 + (sin )2 = 1
r
r
2
y
h2 = x + = ( r sin )2 + (r cos )2 = r 2
2
1 r r
2 = 1 r + 1 1 = 1 r + 1
and hence: = and
r
r r r r r r r r r 2 2
(E2.22)
hr2 = xr
( )
( )
2
2
2
y
h2 = x + + z = (r cos cos ) 2 + (r cos sin )2 + ( r sin )2 = r 2
y
h2 = x + + z = ( r sin sin )2 + (r sin cos )2 = r 2 sin 2
and hence (see (E2.17))
r 2 sin
sin r 2
r
r
1
1
= 2
(r sin ) = 2
r sin
r sin
r sin
(r )
r
(E2.23)
2
1
1
2 =
r
sin
+
(
sin
)
+
sin
r 2 sin r
=
1 2
1
1
r
+ 2
sin + 2
2 r
2
r
r sin 2
r
r sin
(E2.24)
(E2.24)
Excursion 2
page: 82
Tab. E2.1: You might find the following table generally useful:
co-ordinates:
Cartesian
x,y,z
hr
h
h
dV
2
x2
2D circular
x = r cos
3D-cylindrical
x = r cos
3D-spherical
x = r sin cos
y = r sin
y = r sin
y = r sin sin
z = z
1
h z = 1
r
z = r cos
1
r
rsinq
1
r
x
y
z
1 r r
r
r
r
1
r
z
sin r 2
r
1
r sin
r 2 sin
1 1 2
r +
r r r r 2 2
1 1 2
2
+
r +
r r r r 2 2 z 2
1 2
1
1
2
r
+
sin +
r 2 sin 2 2
r 2 r r
r 2 sin
r dr d dz
r 2 sin dr d d
2
2
+ 2+ 2
dx dy dz
r dr d
page: 83
MAPLE exercise
To know:
1. Coulomb potential, effective potential, angular momentum barrier, energy
spectrum, Schrdinger-equation for central force problem.
2. Bohr-model, -radius
3. Radial waveform (Laguerre polynomials)*
4. Solutions*
(r , , ) = N nlmR (r ) () ()
En =
Z 2 e 4
802h 2 n 2
l n 1 and
l m l
* not in detail, but their dependence (the most simple you could remember!)
11 Hydrogenic Atom
page: 84
Ze 2
40 r
(11.1.1)
with 0 = 8.854188 10 12 C 2 / (Jm ) the permittivity in the vacuum. As introduced in part 10.1
we can simplify the two-particle problem to a quasi one-particle problem using its reduced mass
m m
= e nucleus , where mnucleus is the mass of the nucleus.
me + mnucleus
me mp
me + mp
h2 2
Ze 2
= E
2
40 r
(11.1.2)
Because the problem is similar to the rotation of a particle on the surface on the sphere, but now with
a radial potential, an expression of this equation in spherical co-ordinates, (r , , ) , appears to be
appropriate (using eqn. (E2.24)) :
1 2
1
1
2
2 r
+ 2
sin + 2 2
( r , , ) +
r sin
r sin 2
r r r
2
Ze 2
E +
( r , , )
4 0 r
h 2
(11.1.3)
=
(11.1.4)
11 Hydrogenic Atom
page: 85
d
r 2 dr dr
r 2 sin d
r 2 sin 2 d 2
(11.1.5)
2
Ze 2
E
+
R
40 r
h 2
r2
following the typical separation-procedure eq. (11.1.5) is multiplied by
:
R
1 d 2 dR
1
d
d
1
d 2
2
Ze 2 2
r
+
sin
+
+
E
+
r =
R dr dr
sin d
d
40 r
sin 2 d 2
h 2
0 (11.1.6)
and rearranged:
1
1 d 2 dR
2
Ze 2 2
d
d
1
d2
r
+
E
+
r
=
sin
(11.1.7)
2
2
R dr dr
40 r
sin
h 2
sin
Hence, the problem is separated into a radial part (left) and an angular part (right). The right side
(square brackets) we identify as the differential equation for the particle on a sphere (see eq.
(10.4.15) to (10.4.20)) and can readily insert the solutions (spherical harmonics):
1 d 2 dR
2
Ze 2 2 (10 .4 .15 )
1 2
=
r
+ 2 E +
r
R dr dr
40 r
h
(10 . 4. 19 )
Yml
2 Yml
(10 .4 .20 )
1 d 2 dR
2
Ze 2 2
r
+ 2 E +
r l ( l + 1) = 0
R dr dr
40 r
h
1 2 d2 R
dR
2
Ze 2 2
+ 2r +
r
E +
r l ( l + 1) = 0
R dr 2
dr
40 r
h 2
d 2 R( r )
dr 2
(11.1.8)
l (l + 1)
R
r2
(11.1.9)
2
2 dR( r )
Ze 2
l (l + 1)
+
+ 2 E+
R( r ) = 0
r dr
4 0 r
h
r 2
This is not surprising, because the problem can be understood as a particle moving on a collection of
spherical shells. However, what is left to find is its distribution over these shells, the radial part.
11 Hydrogenic Atom
page: 86
Now we want to go -in a way- some steps backwards, to understand our findings. Therefore, we
express equation (11.1.9) in the standard form of the Schrdinger-equation:
h 2 l ( l + 1)
h2 1 d 2 d
r
R
(
r
)
+
+
U
(
r
)
R( r )
2
2 r 2 dr dr
2r
= ER( r )
h 2
R( r ) + U eff ( r ) R (r ) = ER( r )
2
(11.1.10)
h 2 l (l + 1)
U e ff [10-18J]
U [10 -17J]
a)
0.6
b)
l=3
1
0.4
0.2
a.m.b.
r []
-0.2
10
U eff( r)
U( r)
r []
4
10
-1
-0.8
-1.0
l=1
-0.4
-0.6
l=2
l=0
-2
Fig. 11.1: Potentials for 1H: a) green: Coulomb-potential, U(r); blue: angular momentum barrier
(a.m.b.) and red curve: sum of a) and b), the effective potential Ueff(r). Calculations for
l = 1. b) Ueff(r) for l = 0, 1, 2 and 3.
In Fig. 11.1 we recognise the attracting nature of the Coulomb-potential (negative) and the repelling
character of the angular momentum barrier. If the energy becomes higher then the Coulomb-term
and the angular momentum term dominates, the particle is retracted from the nucleus and must
behave like a free (unbound) particle, hence without quantized energy states (energies form a
continuum, see Fig. 11.2). This process is called ionisation and defined by Ueff. Due to the wider
potential range for higher energies, we can also conclude that the energy levels for the bound states
must converge to that limit. Due to the finite value of the potential at a discrete bound energy state we
can finally expect the particle to tunnel beyond the classical limits (that is where the energy level
meets the potential).
11 Hydrogenic Atom
U eff [10-18J]
page: 87
continuum of unbound,
free-particle states
r []
discrete,
bound
states
10
-1
2
2 dR( r )
Ze 2
l (l + 1)
+ 2 E +
R( r ) = 0
r dr
40 r
h
r 2
2
2
Ze 2
l (l + 1)
R + R + 2 E +
R = 0
r
4 0 r
h
r 2
(11.2.1)
E
h2
and
Ze 2
40 h 2
(11.2.2)
leading to:
R +
2
2
l (l + 1)
R + 2R +
R
R =0
r
r
r2
(11.2.3)
We are now testing the equation for r , meaning the electron leaves the nucleus to become a free
particle. In eq, (11.2.3) that causes the terms that depend on 1/r to vanish:
+ 2R = 0
R
(11.2.4)
11 Hydrogenic Atom
page: 88
R ( r ) = A exp i 2 r + B exp + i 2 r
(11.2.5)
where E represents the energy. This general (complex) solution can only be normalised (to
the real value 1), when we assume E or < 0, so that i 2 = i 1 2 = 2 with a
positive radicand. (Try these normalisations as examples or alternatively substitute with a negative,
see eq. (11.2.8)) This can also be suggested when looking at Fig. 11.2 (but now we have
mathematical proof)! The normalisation condition then reads as:
(A exp (
2 r + B exp 2 r
))2 dr = 1
(11.2.6)
The first term becomes infinite for r and is hence unacceptable, because we know that the
wavefunction must vanish, while the second (becoming zero) is acceptable for these conditions.
Therefore, A = 0 and
R ( r ) = B exp 2 r
(11.2.7)
2E
h
(11.2.8)
Now, we try a first approach for a general solution by incorporating the knowledge for r and
ignore B, knowing that the final normalisation of the wavefunction must recover it.
R( r ) = R f ( r ) = e r f ( r )
(11.2.9)
= e r f ( r ) 2f ( r ) + 2 f (r )
(11.2.10)
These results are used in the radial part of the Schrdinger-equation (eq. (11.2.3)):
e r f ( r ) 2 f (r ) + 2 f ( r ) +
2 r
2 l (l + 1) r
e ( f ( r ) f ( r ) ) + 2 +
f (r ) = 0
e
r
r
r2
11 Hydrogenic Atom
2
2
2 l ( l + 1)
f ( r )
f (r ) + 2 +
2 f (r) = 0
r
r
r
2( + ) l (l + 1)
1
f ( r ) + 2 f ( r ) + 2 + 2 +
f (r) = 0
r
r
r2
page: 89
f ( r ) 2 f (r ) + 2 f (r ) +
(11.2.11)
f ( r ) + 2 f (r ) +
f (r) = 0
r
r
r2
(11.2.12)
So far things do not really look simplified. We try another approach for the evaluation of the
unknown part of this equation f(r) by expressing is as a power series (with coefficients b) and trying
to find an upper limit (quantum number) to meet the boundary conditions, which means that the series
is transformed into a polynomial of a degree given by this upper limit:
f (r )
bq r q
and hence R( r ) = e r
bq r q
(11.2.13)
q =0
q= 0
for substitution into eq. (11.2.12), we calculate the first and second derivative:
f (r ) =
bq r q
q
f ( r ) =
q bqr q 1
(11.2.14)
f ( r ) =
q(q 1) bq r q 2
q
[q (q 1) + 2q l (l + 1)]bq r
q 2
[(2 2 ) 2q ]bq r
q 1
=0
=0
(11.2.15)
In order to become zero the coefficients for each power of r have to cancel. Hence, we match the
indices for r q and set them to be zero:
[( q + 2)( q + 1)
bq + 2
bq + 1
+ 2(q + 2) l ( l + 1) ]bq + 2 2 [( q + 1) ( )] bq +1 = 0
2 [( q + 1) ( ) ]
2( q + 2 )
=
[(q + 2)( q + 1) + 2( q + 2) l (l + 1) ] ( q + 2)( q + 3) l (l + 1)
(11.2.16)
11 Hydrogenic Atom
page: 90
We further simplify this by reducing the index, q, by 1, which results in the following recursionformula:
bq +1 = bq
2((q 1) + 2 )
( q + 1)
= 2bq
( q + 2)( q + 1) l (l + 1)
(q + 2)( q + 1) l (l + 1)
(11.2.17)
The next step is to find a limiting case for q and hence dominating and l2, hence assuming q >
l:
2bq
(q + 1)
lim bq +1 2bq
=
(q + 2)( q + 1)
q
q+2
(11.2.18)
The purpose of a recursion-formula is to address the next higher coefficient by a lower. So we want
to try to express bq by b0 (the lowest coefficient relating to a constant). We start by calculating b1, b2
etc... to end up with a general expression for large q:
2
1
b1 b0 = (2 )b0
3
3
1
1
b2 (2)b1 =
(2)2 b0
4
3 4
1
1
b3 (2)b2 =
(2)3 b0
5
3 4 5
...
2
bq = (2)q b0
q!
(11.2.19)
ar
q =0
hence R( r )
( 4 .2 .13 )
(ar )q
q!
e r
bq r q
( 4 .2 .19 )
q =0
e r 2b0
q =0
(2r )q
q!
(11.2.20)
R( r ) = 2b0 e r e 2 r = 2b0 e r
Without loss of generality, we can deposit the 2 in the normalisation constant and get
R (r ) = b0 e r
(11.2.21)
However, this cannot be the final result, because we have to remember the boundary condition
lim R( r ) = 0 and from the definition in (11.2.8) we reassure that > 0 ! Therefore an upper
r
limit p has to be introduced to stop the series and convert it into a polynomial of degree p, e.g.
11 Hydrogenic Atom
page: 91
0 q p., which means that the series stops after p or b p +1 = 0 . Hence b p + 2 = 0 and all
following coefficients vanish due to eq. (11.2.17) and for b p 0 this gives:
b p +1 = 2b p
( p + 1)
= 0
( p + 2)( p + 1) l (l + 1)
(11.2.22)
The equation is fulfilled for a) the nominator = 0 and b) the denominator 0, or:
( p + 1) =
p +1 n
a)
(11.2.23)
b)
>0
>0
( p + 2)( p + 1) >
n( n + 1) >
n >
n
l (l + 1)
l (l + 1)
l
l +1
(11.2.24)
As at the start of the derivation of this formula we (c.f. eq. (11.2.18)) we have to assume p+1 < l.
As expected, we found another quantum number n, related to l, which is related to m (cf. eq.
(10.4.23)). From condition a) we can directly calculate the energies associated with the radial part of
the wavefunction, by re-substituting (11.2.2) and (11.2.8):
2E
Ze 2
=
h
40 h 2 n
En =
Z 2 e 4
32 2 20 h 2 n 2
(11.2.25)
Z 2 e 4
8 20 h 2 n 2
for n l + 1 = 1, 2, 3,K
This is exactly the result Bohr found in his treatment of the hydrogen atom and we also identify our
parameter
=
1
rB
(11.2.26)
11 Hydrogenic Atom
page: 92
The only thing left to do, is the exact determination of the wavefunction R(r). We summarise our
equipment for this task:
r n 1 q
Rn , l ( r ) = exp
bq r
nrB q = 0
1
1
, =
,
rB
nrB
for n l + 1 = 1, 2, 3,K
(q + 1)
coefficien ts : bq +1 = 2bq
( q + 2)( q + 1) l (l + 1)
bq +1 (q + 2)( q + 1) l (l + 1)
bq =
2
( q + 1)
and start with the lowest allowed state: n = 1, hence l = 0
r
R1,0 ( r ) = b0 exp
rB
R1,0 () = b0 e
(11.2.27)
r
with =
rB
R12, 0 ( ) d
=0
= b02
=0
b02
d =
=1
2
(11.2.28)
2 e
(11.2.29)
Now the steps are repeated for n = 2, hence we have to distinguish l = 0, 1 or R2, 0 ( r ) and
R2,1 ( r ) . Generally this is determined by:
r
R2 ,l ( r ) = exp
nrB
1
r
r
hence, R2, 0 ( ) = exp
2rB
b
b0 0
2rB
1
2 rB
b
= 0
2 1 0
2 rB
rB
= N 2 ,0 1 e 2
2
(11.2.31)
11 Hydrogenic Atom
page: 93
b) case: n = 2, l = 1
If we would proceed as in the previous case eq. (11.2.17) yields zero in the denominator.
Therefore, we use the formula backwards
bq +1 (q + 2)( q + 1) l (l + 1)
bq =
2
( q + 1)
therefore we determine b0. and use b1 for normalisation:
b
2 2
b0 = 1
= 0
1
2
1
2 rB
hence R2,1 () =
b1 rB e
= N2 ,1 e
rB
(11.2.32)
Note, that the final results in a) and b) contain the associated normalisation constants N 2 , 0 and
N 2 ,1 .
Continuation of this procedure gives the following table of the radial-parts of the wavefunction for the
hydrogen-like atom, which are depicted in Fig. 11.3. (Note: That Fig. 11.3 shows the radial part
of the wavefunction and not a 3D-projection of the real distribution of the electron, since the angular
terms have been neglected so far)
Table 11.1: The radial-parts Rn, l () of the wavefunction for the hydrogen-like atom as the function
of the radius in multiples of the Bohr-radius.
n
Rn, l ()
R1,0 =
2 e
1
R2, 0 =
(2 ) e 2
2
1
R2,1 =
e 2
2
2
R3, 0 =
27 18 + 22 e 3
27
R3,1 =
2
9 3
(6 ) e
2
2 2 3
R3, 2 =
e
27
11 Hydrogenic Atom
page: 94
Fig. 11.3: Left: The radial part of the wavefunction Rnl() for n=1,2 and 3. Right: The
corresponding probability density function Pnl().
Dr. P. Blmler, School of Physical Sciences, University of Canterbury, UK
11 Hydrogenic Atom
page: 95
The radial part of the eigen-functions of the hydrogenic atom are another ortho-normal class of
polynomials (like the Hermite and Legendre polynomials), called associated Laguerre polynomials:
They are solutions of the so-called associated Laguerre equation, in which eq. (11.2.1) can be
transformed.
xf ( x) + (k + 1 x ) f ( x) + j f ( x ) = 0
solutions : f ( x) = L kj ( x)
associated Laguerre polynomial s :
L jk ( x ) = (1) k
(11.2.33)
j+ k
j + k xi
(1) i
dx k i = 0
i i!
dk
/ n
2l + 1
Ln l 1 ()
n
2 ( n l 1)!
n 2n(n + l )!
N nl =
and =
r
rB
Ze 2
40 h 2
3
2 ( n l 1)! (2l + 1)( l m )!
n 2n(n + l )! 4 ( l + m )!
Z 2 e 4
8 20 h 2 n 2
with n = 1, 2, 3,... and
l n 1 and
l m l
(11.3.1)
11 Hydrogenic Atom
page: 96
As expected the three degrees of freedom in motion generate three quantisation conditions,
described by the three quantum numbers n, m and l. Different nomenclature is in use for them:
n
n = 1, 2, 3, 4,....
n = K, L, M, N, ....
spectroscopic notation:
m
l m l
spectroscopic notation:
The following table is illustrating the mutual dependence of the quantum numbers.
Tab. 11.2: Summary of the interdependence of the quantum numbers for the hydrogenic atom, for
n = 1, 2 and 3 including nomenclature and degeneracy. (Neglecting the spin of the
electron, which is done in the next lecture)
n
1
2
3
l
m
0
0
0
0
name
sp.n.*
1s
2s
1s
2s
-1
1
0
2p
2p
2p
0
0
-1
3s
2p
D(n)**
1
4
*spectroscopic name including m-notation.
** D(n) degeneracy of state.
3s
1
0
-2
-1
3p
3p
3p
2
0
3d
3d
3d
3p
3d
3d
3d
The energy is determined by n only (cf. eq. (11.2.25)). Hence we have 2l+1 degenerate states
(span of m), where the span of l is defined by n-1. Therefore the degeneracy of a state n is given by:
n 1
D( n) =
(11.3.2)
l =0
Combining the equations in (11.3.1) gives the complete wavefunctions for the hydrogenic atom as
listed in Tab. 11.3:
11 Hydrogenic Atom
page: 97
Tab. 11.3: The complete wavefunctions, nlm ( , , ) , of the hydrogenic atom for
n = 1, 2 and 3.
Name
nlm ( , , )
1s
1
e
2s
1
(2 )e / 2
4 2
2p
-1
1
8
2p
2p
e / 2 sin e i
1
e / 2 cos
4 2
1
8
e / 2 sin e i
3s
1
27 18 + 2 2 e / 3
81 3
3p
-1
1
(6 ) e / 3 sin e i
81
3p
2
(6 ) e / 3 cos
81
3p
1
(6 ) e / 3 sin ei
81
3d
-2
1
2 e / 3 sin 2 e 2i
162
3d
-1
1
2 e / 3 sin cos e i
81
3d
1
81 6
2 e / 3 3 cos 2 1
3d
1
2 e / 3 sin cos ei
81
3d
1
2 e / 3 sin 2 e 2i
162
11 Hydrogenic Atom
page: 98
For the calculation of the normalisation constant N nlm it is important to realise, that the integration is
(preferably) done in spherical co-ordinates (cf. Tab. E2.1) and the integrand can be further
simplified according to eq. (10.4.25), where we have learned that the -term behaves constant.
2
N nlm
2
N nlm
*( x, y, z)(x, y, z)dxdydz
( E 2. 1)
0 0 0
2
* ( r , , ) ( r , , ) r 2 sin dr dd
(10 . 4. 25 )
(11.3.3)
r = 0 =0 =0
2
N nlm
2
[rR(r )()]
sin drd = 1
r =0 =0
This procedure can be easily understood, when we have a look at the s-orbitals, which according to
l = 0 have spherical symmetry. Hence, we can replace the radial probability distribution function (e.g.
in Fig. 11.3/right) by an integration over spherical shells of thickness dr. Hence, we have to substitute
the probing volume (d) of the radial probability distribution function, Pn 0 () d = Rn20 () d ,
with the volume of the spherical shells (42 d) to get the projection of the three-dimensional
~
distribution function, Pn0 () (we call it the radial distribution function), of the electron on the
radius or
~
Pn 0 () d = 42 Rn20 () d
(11.3.4)
This principle is illustrated in Fig. 11.4 using the 1s-orbital as an example. From the plot of the
radial distribution function for the 1s, 2s and 3s-orbital (see Fig. 11.5) the Bohrs idea of shells, in
which the electron is located can be anticipated, but we also realise, that the nth shell has only the
maximum probability. The electron has a finite probability to be found in lower shells. (Closer
examination also uncovers that -except of 1s- that the maximum is not at integer multiples of rB) We
also realise that there is a small but non-zero probability to find the electron inside the nucleus,
assuming the nucleus has a finite radius. This is of great importance for nuclear reaction e.g. +decay via electron capture.
Finally we want to visualise the atomic orbitals, which is a difficult task, because we have a 3Dwavefunction which has a different values at the co-ordinates ,, . Therefore, we have to depict it
either as a cloud of varying density, or we have to pick a certain threshold of probability to which we
assign a surface (isosurface). Both ways of visualisation are shown in Tab. 11.4, where the different
signs of the lobes have been colour-coded (positive = red, negative = green).
For an on-line visualisation, you can download Orbital-Viewer (a program from D. Manthey) from
the webpage (http://wwwnmr.ukc.ac.uk/staff/pb/teach/PH502/Aos/ ao.html) of the
course.
11 Hydrogenic Atom
page: 99
P()
~
P ()
d
1
r=rB
Fig. 11.4: Illustration for the construction of the radial probability density function for the 1s-orbital:
Upper: If an infinitesimal small and constant sampling volume d is scanned along the
radius, we will receive the radial-part of the electron density P(). (If additionally the
detector would be moved on a circle, for each the same value would be recorded. The
1s-orbital is isotropic.) Lower: Now the electron density is measured over a spherical
~
shell of thickness d, to determine the radial probability distribution function P () , which
has a maximum at the Bohr radius ( = 1).
16
14
2.5
~
P10( )
2
1.5
1
2
0.5
0
12
10
8
~
P20( )
10
15
20
25
30
10
15
20
25
30
P~30 ()
6
4
2
0
10
15
20
25
30
Fig. 11.5: The radial distribution function (not-normalised) for the first three s-states (l = 0).
Compare with the according states in Fig. 11.3.
11 Hydrogenic Atom
page: 100
Tab. 11.4: Atomic orbitals for hydrogenic atoms for n = 1, 2 and 3. On the left of each cell the orbital is shown as a probability plot using 10000 points. On
the right (for ns-orbitals in the centre) an isosurface (corresponding to a probability of 10-4) is shown. For the ns-orbitals additionally, an octant of
this isosurface has been cut away to show the inner shells. Red colour corresponds to positive lobes, green to negative.
0/s
0/s
-2
-1
1/p
0/s
1/p
2/d
m
0
+1
+2
11 Hydrogenic Atom
page: 101
Fig. 11.6: Energy levels for hydrogen 1H including the l 5. Energy is measured in units of
h 2 /( 2rB2 ) .
8 2
(E2.22)
x
2
1
/ 2
[
]
=
e
sin cos
=
e / 2 2px
8 2
4 2
rB
Analogously we get:
(E2.22)
y
1
1
(2 p 2 p ) = 1 e / 2 sin sin =
e / 2 2p y
4 2
4 2
rB
2 p =
(E2. 22 )
z
1 2 / 2
1
e
cos =
e / 2 2p z
8
4 2
rB
Hence, we receive three equivalent 2p-functions, which can be correlated to the three Cartesian coordinates. This is useful for geometrical considerations (e.g. chemical complexes).
Similarly we proceed with the other wavefunctions (see Tab. 11.5)
Dr. P. Blmler, School of Physical Sciences, University of Canterbury, UK
11 Hydrogenic Atom
page: 102
combination
hybrid wavefunction
1s
1s
1
e
2s
2s
1
(2 )e / 2
4 2
2pz
2p
1
e / 2 cos
4 2
2px
1
(2p + 2p )
2
1
e / 2 sin cos
4 2
2py
m1
1
(2p 2p )
2
1
e / 2 sin sin
4 2
3s
3s
1
27 18 + 2 2 e / 3
81 3
3pz
3p
2
(6 ) e / 3 cos
81
3px
1
(3p + 3p )
2
2
(6 ) e / 3 sin cos
81
3py
m1
1
(3p 3p)
2
2
(6 ) e / 3 sin sin
81
3d z 2
3d
3dxz
1
(3d + 3d )
2
2 2 / 3
e
sin cos cos
81
3dyz
m1
i
(3d 3d )
2
2 2 / 3
e
sin cos sin
81
3d x 2 y 2 3
1
(3d + 3d )
2
3dxy
1
81 6
m2
i
(3d 3d )
2
1
81 2
1
81 2
2 e / 3 3 cos 2 1
2 e / 3 sin 2 cos 2
2 e / 3 sin 2 sin 2
The hybrid orbitals do not look different from those in Tab. 11.4. They are just transferred back to
Cartesian co-ordinates (with slightly different normalisation constants).
Thanks!
page: 103
Acknowledgements:
I would like to thank Dr. C. Isenberg to provide access to his notes for a first layout of the course.
They have turned out to be very helpful and guiding for a new lecturer. I also want to thank Mr. Ian
Campbell for proof-reading the first draft of the manuscript.