EIGENFUNCTIONS OF POSITION AND MOMENTUM; UNIT

OPERATORS

Link to: physicspages home page.

To leave a comment or report an error, please use the auxiliary blog.
References: Griffiths, David J. (2005), Introduction to Quantum Mechan-
ics, 2nd Edition; Pearson Education, Section 3.6.
Tom Lancaster and Stephen J. Blundell, Quantum Field Theory for the
Gifted Amateur, (Oxford University Press, 2014), Section 3.5.
There are a few results that will be used frequently in quantum theory
that I think it’s worth collecting together and explaining in full.
First, we’ll revisit the eigenfunctions of the position and momentum op-
erators. In the earlier post, we showed that the eigenfunctions of the position
operator are delta functions and we wrote

(1) |x0 i = δ (x − x0 )

Strictly speaking this equation gives the position space representation of

the eigenfunction. More precisely, we should just say that |x0 i is an eigen-
function of the position operator x̂ and leave it at that. In order to write it
as a ’proper’ function (that is, a function we can use in calculations such as
integrals), we need to specify the space we’re using and then write |x0 i in
that space, as we did above for position space.
For momentum, we’ve seen that the eigenfunctions are

1
(2) |p0 i = √ eip0 x/h̄
2π h̄
The normalizations of both the position and momentum eigenfunctions
give us more delta functions:

ˆ
(3) hx1 |x0 i = δ (x − x1 ) δ (x − x0 ) dx
(4) = δ (x1 − x0 )
ˆ
1
(5) hp1 |p0 i = ei(p0 −p1 )x/h̄ dx
2π h̄
(6) = δ (p0 − p1 )
1
 EIGENFUNCTIONS OF POSITION AND MOMENTUM; UNIT OPERATORS 2

Given a complete basis set of states, we can define a set of projection

operators each of which projects a function onto the basis vector that defines
the projection operator. A projection operator has the form

(7) P̂ ≡ |αihα|

so that applying it to a state |ψi gives

(8) P̂ |ψi = hα |ψ i |αi

Note that this is a completely general expression; we can choose any ba-
sis states |αi (they could be the eigenstates of position or momentum, or
the discrete set of states for some system such as the states of the infinite
square well or harmonic oscillator) and the projection operator gives the
component of |ψi ’along’ that basis vector. In practice, to do calculations
we usually express |ψi in position or momentum space (or in matrix form
if it’s a spin state) but in this formula, |ψi is just an abstract symbol repre-
senting some arbitrary state.
For a complete set of discrete basis states we can define the unit operator

(9) 1 ≡ ∑ |αi hα|

α

or for a continuous set of basis states

ˆ
(10) 1≡ dα |αi hα|

This works because it’s just like expressing a 3-d vector as a sum of its
components in some basis, such as rectangular coordinates

(11) v = vx x̂ + vy ŷ + vz ẑ
Since the basis consisting of the states |αi is complete, we can write any
other state in terms of that basis set. We’re using the projection operator for
each basis state to project out the new state onto each of the basis states in
turn, then adding up the result:

(12) |ψi = ∑ |αi hα |ψ i

α

or
 EIGENFUNCTIONS OF POSITION AND MOMENTUM; UNIT OPERATORS 3

ˆ
(13) |ψi = dα |αi hα |ψ i

Example 1. Armed with these results, it’s worth looking at Example 3.6 in
Lancaster & Blundell in a bit more detail. In that example, they extend the
creation-annihilation operator representation to cases where the momentum
(and hence the energy) states merge into a continuum. In that case, the
commutation relation for the operators becomes
h i
(14) ap , aq = δ (3) (p − q)
†

L&B are dealing with a 3-d particle in a box with periodic boundary
conditions (rather than requiring the wave function to be zero outside the
box), so the momentum eigenstate |pi is just a plane wave so that its position
space representation is

1
(15) hx |p i = √ eip·x
V

where V is the volume of the box.

They begin with a one-particle state

 0 D  E
(16) p p = â†p 0 â†p0 0

D   E
 †
(17) = 0 âp âp0  0

We can now use the relation 14 to get

âp â†p0 = δ (3) p − p0 + â†p0 âp

(18)

and since âp |0i = 0 we get

 0 D 
 E
 (3) 0 
(19) p p
 = 0 δ p−p 0

= δ (3) p − p0


(20)
Now we want to get the position space version of the state |pi. From 1
(generalized to 3-d) we see that
 EIGENFUNCTIONS OF POSITION AND MOMENTUM; UNIT OPERATORS 4

(21) φp (x) ≡ hx |p i
ˆ
d 3 x0 x x0 x0 |p

 

(22) =
ˆ
d 3 x0 δ (3) x − x0 x0 |p



(23) =

so if we can write |pi as a function of x0 then the expression hx |pi merely

picks out the precise position x that we’re interested in. Using a set of
momentum basis states |qi we can transform |xi using the unit operator 13:

ˆ
(24) |xi = d 3 q |qi hq |xi
ˆ
(25) = d 3 q |qi hx |qi∗
ˆ
(26) = d 3 qφq∗ (x) |qi

Therefore

ˆ
(27) hx| = d 3 qφq (x) hq|
ˆ
(28) hx |pi = d 3 qφq (x) hq |p i
ˆ
(29) = d 3 qφq (x) δ (3) (q − p)
(30) = φp (x)

Example 2. We can apply the same arguments to a 2-particle state. Start

with

0 0  D   E
(31) p q |pq = 0 âp0 âq0 â†q â†p  0


From commutation relations 14 we get

 EIGENFUNCTIONS OF POSITION AND MOMENTUM; UNIT OPERATORS 5

 
âp0 âq0 â†q â†p = âp0 δ (3) q − q0 + â†q âq0 â†p


(32)

 
= δ (3) q − q0 δ (3) p − p0 + â†p âp0 + âp0 â†q âq0 â†p


(33)

 
= δ (3) q − q0 δ (3) p − p0 + â†p âp0 +


(34)
  
δ (3) q − p0 + â†q âp0 δ (3) p − q0 + â†p âq0



(35)

Applying âp |0i = 0 we get

0 0
p q |pq = δ (3) q − q0 δ (3) p − p0 + δ (3) q − p0 δ (3) p − q0





(36)

To convert to position coordinates, this time we have two independent

positions, one for each particle, which we’ll call x and y, so a position state
is |xyi. Since the particles are independent, we can represent the compound
state as the product of two single-particle states:

 0 0  0  0
(37) p q = p q

where each single-particle state has its own coordinates, independent of

the other state. The inner product of two such states is

(38) hrs |pq i = hr |p i hs |q i

since each particle’s coordinates are independent of the other particle.

Using the more familiar wave function representation, we could have
something like this:

ˆ ˆ
(39) hrs |pq i = d 3 yφr∗ (x) φs∗ (y) ψp (x) ψq (y)
3
d x
ˆ ˆ
(40) = d xφr (x) ψp (x) d 3 yφs∗ (y) ψq (y)
3 ∗

(41) = hr |p i hs |q i

In that case we can write

 EIGENFUNCTIONS OF POSITION AND MOMENTUM; UNIT OPERATORS 6

ˆ ˆ
1 3 0
d 3 q0 p0 q0 q0 |y p0 |x
  


(42) |xyi = √ d p
2! ˆ ˆ
1 3 0
d 3 q0 φp∗0 (x) φq∗0 (y) p0 q0
  
(43) = √ d p
2! ˆ ˆ
1 3 0
d 3 q0 φp0 (x) φq0 (y) p0  q0 



(44) hxy| = √ d p
2! ˆ ˆ
1 3 0
d 3 q0 φp0 (x) φq0 (y) p0 q0 


(45) = √ d p
2!
The √12! is there because the double integral extends over all values of
both p0 and q0 so it counts the state |p0 i |q0 i twice, once as |p0 i |q0 i and
once as |q0 i |p0 i. It’s a square root because we’re dealing with a raw wave
function and it’s the square modulus of this that must be normalized.
With this, we get, using 36
ˆ ˆ
1
d p d 3 q0 φp0 (x) φq0 (y) p0 q0 |pq
3 0


(46) hxy |pqi = √
2!
1
(47) = √ [φp (x) φq (y) + φq (x) φp (y)]
2
This is the symmetrized wave function for two identical bosons. Follow-
ing through the same argument using anticommutators for fermions gives
the fermion result

1
(48) hxy |pq i f ermion = √ [φp (x) φq (y) − φq (x) φp (y)]
2