Lecture 22:
Coherent States
Phy851 Fall 2009
Summary memorize
Properties of the QM SHO:
P2 1 h
H= + m 2 X 2 =
2m 2 m
1 X 1 X
A= + i P A = i P
2 h 2 h
X=
(A + A ) P = i h A A ( )
2 2
1
H = h A A +
2
H n = h (n + 1 / 2) n n
n =
( A ) 0
A n = n n 1 A n = n + 1 n + 1 n!
x2
1/ 2
n (x) = [ 2 n n! ] H n ( x / ) e 2 2
2 x n 1
n (x) = n1 (x) n2 (x)
n n
x2 x2
1/ 2 1/ 2 x
0 (x) = [ ] e 2 2 1 (x) = 2 [ ] 2 e
2 2
h
X = n + 1 / 2 P = n +1/ 2
What are the `most classical states
of the SHO?
In HW6.4, we saw that for a minimum
uncertainty wavepacket with:
osc h
x = osc =
2 Mosc
The uncertainties in position and
momentum would remain constant.
The interesting thing was that this was
true independent of x0 and p0, the initial
expectation values of X and P.
We know that other than the case x0=0
and p0=0, the mean position and
momentum oscillate like a classical
particle
This means that for just the right initial
width, the wave-packet moves around
like a classical particle, but DOESNT
SPREAD at all.
Coherent States
Coherent states, or as they are sometimes
called Glauber Coherent States are the
eigenstates of the annihilation operator
A = =1
Here can be any complex number
i.e. there is a different coherent state for every
possible choice of
(Roy Glauber, Nobel Prize for Quantum Optics
Theory 2005)
These states are not really any more
coherent then other pure states,
they do maintain their coherence in the
presence of dissipation somewhat more
efficiently
In QM the term coherence is over-used and
often abused, so do not think that it always
has a precise meaning
Glauber Coherent States are very important:
They are the most classical states of the
harmonic oscillator
They describe the quantum state of a laser
Replace the number of quanta with the number
of photons in the laser mode
They describe superfluids and super-conductors
Series Solution
Let us expand the coherent state onto energy
eigenstates (i.e. number states)
= cn n
n =0
Plug into eigenvalue equation:
A =
A cn n = cn n
n =0 n =0
c
n =0
n n n 1 = cn n
n =0
Hit from left with m|:
m c n n n 1 = cn n
n =0 n =0
c
n =0
n n m n 1 = cn m n
n =0
cm +1 m + 1 = cm
Continued
cm +1 = cm cm = cm 1
m +1 m
Start from: c 0 = ! ( )
The constant N() will be used at the end for
normalization
few iterations:
Try a
c1 = c0 = ! ( )
1 1
2
c2 = c1 = ! ( )
2 2 1
3
c3 = c2 = ! ( )
3 3 2 1
4
c4 = c2 = ! ( )
4 4 3 2 1
So clearly by induction we have:
n
cn = ! ( )
n!
Normalization Constant
n
cn = ! ( )
n!
So we have:
n
= ! ( ) n
n= 0 n!
For normalization we require:
1=
m
2 n
= ! ( ) mn
n= 0 m!n!
m= 0
2n
2
= ! ( )
n= 0
n!
2
2
= ! ( ) e
Which gives us:
2
2
n
=e 2
n
! ( ) = e 2
n =0 n!
Orthogonality
Let us compute the inner-product of two
coherent states:
2 2
+ m
n
=e 2
mn
n =0 m!n!
m =0
2 2 n
+
( )
=e 2
n!
n= 0
2 2
+
+
=e 2
Note that:
2
( )
( )
e =e
( +
2 2
+a + )
=e
2
=
So coherent states are NOT orthogonal
Does this contradict our earlier results
regarding the orthogonality of eigenstates?
Expectation Values of Position
Operator
Lets look at the shape of the coherent
state wavepacket
Let ( x) = x
X = dx ( x) x ( x)
Better to avoid these integrals, instead
lets try using A and A :
X =
2
( A + A )
Recall the definition of |:
A = A =
X =
2
( A + A )
=
2
( +
)
=
2
( + )
X = 2 Re{ }
Expectation Value of Momentum Operator
We can follow the same procedure for the
momentum:
h
P = i ( A A )
2
2h
=
2i
( A A
)
2h
=
2i
(
)
2h
P = Im{}
X = 2 Re{ }
Not surprisingly, this gives:
1 1
= X + i P
2 h
Variance in Position
Now let us compute the spread in x:
2 2
X2 =
2
( A+ A )
2 2
=
2
( A + AA
+ A
A + A A )
Put all of the A s on the right and the A s on
the left:
This is called Normal Ordering
2 2
=
2
( A + 2A
A + 1+ A A )
2 2
=
2
(
+ 2 + 1+ 2
)
2
=
2
(( 2
+ ) +1 )
2
X = ( + ) X2 = X
2
+
2 2
2 2 Exactly the same variance as
X = X X = the ground state |n=0
2
Momentum Variance
Similarly, we have:
h2 2
P 2
= 2 (A A )
2
h2
= 2 ( AA AA A A + A A )
2
Normal ordering gives:
h2
P 2
= 2 ( AA 2A A 1+ A A )
2
h2
= 2 ( AA 2A A 1+ A A )
2
h2
( 2
= 2 2 2 + 1
2
)
h2
( 2
= 2 ( ) 1
2 )
2h
P =
2i
(
)
2 h2 2
P = 2
P P
2 h
P = P + 2 =
2 2
Minimum Uncertainty States
Let us check what Heisenberg Uncertainty
Relation says about coherent states:
2 2
X = X X =
2
2 h
P = P2 P =
2
h
XP =
2 2
h
XP =
2
So we see that all coherent states (meaning
no matter what complex value takes on)
are Minimum Uncertainty States
This is one of the reasons we say they are
most classical
Time Evolution
We can easily determine the time evolution of
the coherent states, since we have already
expanded onto the Energy Eigenstates:
Let
(t = 0) = 0
Thus we have:
2
0n
(0) = e 2
n
n =0 n!
2
0n i ( n +1/ 2 )t
(t ) = e 2
e n
n =0 n!
2
i t / 2
0n i n t
=e e2
e n
n =0 n!
2
i t n
( e )
0
Let
=e it / 2
e 2
n!
n
n= 0
(t ) = 0 e i t
By this we mean it remains in a
(t ) = (t ) coherent state, but the value of
the parameter changes in
time
Why most classical?
What we have learned:
Coherent states remain coherent states as time
evolves, but the parameter changes in time
as
(t ) = 0 e i t
This means they remain a minimum
uncertainty state at all time
The momentum and position variances are the
same as the n=0 Energy eigenstate
Recall that:
X = 2 Re{ }
2h
P = Im{}
So we can see that:
x0 = (t ) X (t )
1 x
0 = 0 + i p0 p0 = (t ) P (t )
2 h
We already know that <X> and <P> behave as
classical particle in the Harmonic Oscillator, for
any initial state.
p0
x(t ) = x0 cos(t ) + sin(t ) p (t ) = p0 cos(t ) x0 sin(t )
Conclusions
The Coherent State wavefunction looks
exactly like ground state, but shifted in
momentum and position. It then moves as a
classical particle, while keeping its shape
fixed.
Note: the coherent state is also called a
Displaced Ground State
