Dirac Notation and the Three Pictures

1 Dirac Notation
Imagine a vector v in two dimensions. You could write it as v = î vx + ĵ vy . Alternatively, you could write it as
v = ê1 v1 + ê2 v2
where for example,
1 1
ê1 = √ (î + ĵ) ; ê2 = √ (î − ĵ)
2 2
are two other mutually perpendicular unit vectors and,
1 1
v1 = √ (vx + vy ) ; v2 = √ (vx − vy )
2 2
These two different ways of writing the same vector are called representations. Since êj · v = vj ,
v = ê1 (ê1 · v) + ê2 (ê2 · v)
These are written in quantum mechanics notation as,
X
|v >= |e1 >< e1 |v > +|e2 >< e2 |v >= |ej >< ej |v >
j

where we identify,

v ↔ |v > ; ej ↔ |ej > ; < ej |v >↔ (ej · v)

In quantum mechanics, the state |ψ > of a system are vectors in an infinite dimensional space. In the x-representation
we may write, Z
X
|ψ >= |x >< x|ψ >≡ dx |x >< x|ψ >
x

where |x > are the basis of position eigenstates. In case of ordinary vectors, we can write the dot product in a
matrix form also. < a|b >≡ a · b = aT b where a, b are column vectors and aT is the transpose of a which is a row
vector. This means that < a|a > ≥ 0. In quantum mechanics, since the vectors are complex, we have to use a†
instead of aT so that a† a ≥ 0 and not aT a ≥ 0. This means that in quantum mechanics, < a|b >≡ a† b so that
< a|b >∗ =< b|a >. Since the values are continuous we have to replace summation with integration. This means
< x|ψ >≡ ψ(x) is the wavefunction of the system. Alternatively we can have another representation where we use
momentum eigenstates |p > as the basis then we get the momentum representation of the same state |ψ >.
Z
|ψ >= dp |p >< p|ψ >

1
This means < p|ψ >≡ φ(p) is the wavefunction in momentum space. Equating the two,
Z Z
dx |x >< x|ψ >= dp |p >< p|ψ > (1)

We assert, 0 0
< x|x >= δ(x − x )
and 0 0
< p|p >= δ(p − p )
0
Eq.(1) is valid for any |ψ >. Choose |ψ >≡ |p >. This means,
Z Z Z
0 0 0 0
dx |x >< x|p >= dp |p >< p|p >= dp |p > δ(p − p ) = |p > (2)

0
Now choose |ψ >≡ |x > in Eq.(1).
Z Z Z
0 0 0 0
dx |x >< x|x >= dx |x > δ(x − x ) = |x >= dp |p >< p|x > (3)

From this we conclude that, Z Z

dx |x >< x| = dp |p >< p| = 1 (4)

We now wish to calculate < x|p >. This is analogous to finding ej · î or êj · ĵ in the two dimensional example. Just
as < x|ψ >≡ ψ(x) is the wavefunction of a particle in a state ψ , < x|p > is the wavefunction of the particle in a
state with momentum p. We know that this wavefunction is nothing but eikx where p = ~k. Hence,
p
< x|p >= C ei ~ x

We have to calculate the proportionality constant by requiring that,

Z ∞
0 0 0
< p|x >< x|p > dx =< p|p >= δ(p − p )
−∞

This means, Z ∞ Z ∞ 0
0 (p −p) 0
2
< p|x >< x|p > dx = |C| ei ~ x
dx = δ(p − p )
−∞ −∞

Changing variables from x → x~ we get,

Z ∞ Z ∞ 0
0 0
−p)y
< p|x >< x|p > dx = ~|C|2 ei(p dy = δ(p − p )
−∞ −∞

Or,
2π~|C|2 = 1
or C = √1 where h = 2π~. Hence,
h
1 p
< x|p >= √ ei ~ x
h

2
 A representation free 1D Schrodinger’s equation would be,

∂ p2
i~ |ψ(t) >= ( + V (x))|ψ(t) >
∂t 2m
In the position representation it is,

∂ p2 ~2 d2
i~ < x|ψ(t) >=< x|( + V (x))|ψ(t) >= (− + V (x)) < x|ψ(t) >
∂t 2m 2m dx2
where we identify < x|ψ(t) >≡ Ψ(x, t) with the wavefunction of the particle. Hence,

∂ ~2 d 2
i~ Ψ(x, t) = (− + V (x))Ψ(x, t) (5)
∂t 2m dx2
In the momentum representation it is,

∂ p2 p2 d
i~ < p|ψ(t) >=< p|( + V (x))|ψ(t) >= ( + V (i~ )) < p|ψ(t) >
∂t 2m 2m dp

where < p|ψ(t) >≡ Φ(p, t) is the wavefunction in momentum space.

∂ p2 d
i~ Φ(p, t) = ( + V (i~ ))Φ(p, t) (6)
∂t 2m dp
Keep in mind that, Z ∞
dk ikx
Ψ(x, t) = e Φ(~k, t) (7)
−∞ 2π

Homework: If V (x) = ax + b, solve Eq.(5) and Eq.(6) and compare with Eq.(7). Assume that Ψ(x, 0) ≡ g(x) is
known.

Homework: Find < φ|lz > where |φ > is the eigenstate of angle and Lz |lz >= lz |lz >. keep in mind that while,
0 0
< φ|φ >= δ(φ − φ )

since angles are continuous, 0

< lz |lz >= δlz ,lz0
0
since the eigenvalues lz , lz are discrete.

Homework: If L2 = 2~2 , find < lx |lz >.

Homework: If S 2 = 43 ~2 , find < sx |sz >.

3
2 Schrodinger, Heisenberg and Interaction Picture
In classical physics, a quantity A(r, p) is time independent if

d
A(r(t), p(t)) = 0
dt
where r(t) is the trajectory of the particle and p(t) = mṙ(t). But in quantum physics, there is no trajectory but we
would still like to make sense of the statement that a quantity is time independent. One way it to find a number
associated with this quantity and ask whether this is time independent. In general in quantum physics A(r, p) is
an operator since both r and p are operators. So we can ask if its expectation value with respect to some state is
time independent or not? It is desirable to have a stronger condition namely this,
d
< ψ(t)|A(r, p)|χ(t) >= 0
dt
where |ψ(t) > and |χ(t) > are two different states obeying Schrodinger’s equation (if ψ = χ this would be the
expectation value).
∂ p2
i~ |ψ(t) >= ( + V (r))|ψ(t) >
∂t 2m
∂ p2
|χ(t) >= (
i~ + V (r))|χ(t) >
∂t 2m
As we don’t expect conserved quantities when the potential depends on time we have chosen a time independent
p2
potential V (r). The conjugate to the first one is (set H = 2m + V (r) )

∂
−i~ < ψ(t)| =< ψ(t)| H
∂t
Think of it like this. For a simple 2 × 2 Hamiltonian,
 
a(t)
|ψ(t) >=
b(t)
 
Haa Hab
H=
Hba Hbb
∂
|ψ(t) >= H|ψ(t) >
i~ (8)
∂t
means,     
∂ a(t) Haa Hab a(t)
i~ = (9)
∂t b(t) Hba Hbb b(t)
On the other hand,
∂
− i~ < ψ(t)| =< ψ(t)|H (10)
∂t
means,  
∂ Haa Hab
a∗ (t) b∗ (t) a∗ (t) b∗ (t)



− i~ = (11)
∂t Hba Hbb

4
Homework: Verify that Eq.(8) and Eq.(10) mean the same thing by verifying that Eq.(9) and Eq.(11) are the
same things.

Coming back to the question, the rate of change we are looking for is,
d d d
i~ < ψ(t)|A(r, p)|χ(t) >= (i~ < ψ(t)|)A(r, p)|χ(t) > + < ψ(t)|A(r, p)(i~ |χ(t) >)
dt dt dt
= −(< ψ(t)|H)A(r, p)|χ(t) > + < ψ(t)|A(r, p)(H|χ(t) >) =< ψ(t)|[A(r, p), H]|χ(t) >
Thus,
d
i~ < ψ(t)|A(r, p)|χ(t) >=< ψ(t)|[A(r, p), H]|χ(t) >
dt
Since this is true for any ψ and χ, A is conserved if,

[A(r, p), H] = 0

Homework: Show that L = r × p is conserved for a particle acted on by a central force.

Homework: Show that

i
|χ(t) >= e− ~ tH |χ(0) >
i
< ψ(t)| =< ψ(0)|e ~ tH
obeys Schrodinger’s equation. This means,

i i
< ψ(t)|A(r, p)|χ(t) >=< ψ(0)|e ~ tH A(r, p)e− ~ tH |χ(0) > (12)

We want to simplify the operator,

i i
e ~ tH A(r, p)e− ~ tH
further. We define time dependent operators as follows.
i i
r(t) = e ~ tH r e− ~ tH

Note that r(t) is NOT THE TRAJECTORY of the particle. r(t) is still an operator just as r is an operator.
Similarly we define,
i i
p(t) = e ~ tH p e− ~ tH
Here too p(t) is an operator.

Homework: Even though r(t) is not the trajectory show that,

d
p(t) = m r(t)
dt
even in quantum mechanics.

5
 i i
Homework: Show that e ~ tH f (r) e− ~ tH = f (r(t)) where f is any function. Hint Taylor expand f in powers of
r.

This means
i i
e ~ tH A(r, p)e− ~ tH = A(r(t), p(t)) (13)
Hence,
< ψ(t)|A(r, p)|χ(t) >=< ψ(0)|A(r(t), p(t))|χ(0) > (14)
The left side of Eq.(14) is called “Schrodinger’s picture” and the right side is called “Heisenberg’s picture”.
In Schrodinger’s picture the states are time dependent and the operators are time independent, and vice versa
for Heienberg’s picture. There is a third picture called interaction picture which is used while doing perturbation
theory. Imagine that an external time dependent perturbation is applied.

∂ p2
i~ |ψ(t) >= ( + V (r) + Hext (t))|ψ(t) > (15)
∂t 2m
∂ p2
i~ |χ(t) >= ( + V (r) + Hext (t))|χ(t) > (16)
∂t 2m
Define,
i p2 i
|χ(t) >int = e ~ t( 2m +V (r)) |χ(t) >≡ e ~ tH |χ(t) >
or,
i p2
|χ(t) >= e− ~ t( 2m +V (r)) |χ(t) >int
Substituting into Eq.(16) we get,
∂
i~ |χ(t) >int = Hext,ev (t)|χ(t) >int (17)
∂t
∂
i~ |ψ(t) >int = Hext,ev (t)|ψ(t) >int (18)
∂t
where,
i i
Hext,ev (t) = e ~ tH Hext (t)e− ~ tH
Hence,
i i
< ψ(t)|A(r, p)|χ(t) >=< ψint (t)|e ~ tH A(r, p)e− ~ tH |χint (t) >=< ψint (t)|A(r(t), p(t))|χint (t) > (19)

The term < ψint (t)|A(r(t), p(t))|χint (t) > is called the Interaction picture. In this picture the states evolve ac-
cording to the time dependent external Hamiltonian but the operators evolve according to the the time independent
Hamiltonian.