6, Quantum theory of Fluorescence

1. Quantum theory of Damping: Density operator

operators,
the reservoir may be taken as any large collection of systems with many degrees
(e.g. phonons, other photon modes, etc), with closely spaced frequencies
k
,

b
k
and

b

k
operators,
the Hamiltonian for the eld-reservoir system is

H = a

a +

b +

k
g
k
(

k
a + a

b
k
),
the Heisenberg equations of motion for the operators are
d
dt
a(t) =
i

[

H, a] = i a(t) i

k
g
k

b
k
(t),
d
dt

b
k
(t) = i
k

b(t) ig
k
a(t),
IPT5340, Fall 06 p.2/47
Damping via oscillator reservoir
the closed equation for the eld operator a(t) is,

b
k
(t) =

b
k
(0)e
i
k
t
ig
k

0
dt

e
i
k
(tt

)
a(t

),
d
dt
a(t) = i a(t)

k
[g
k
[
2

0
dt

e
i
k
(tt

)
a(t

) +

f
a
(t),
where

f
a
(t) = i

k
g
k

b
k
(0)e
i
k
t
,
dene slowly varying operator a

(t) = ae
it
, then
d
dt
a

(t) =

k
[g
k
[
2

0
dt

e
i(
k
)(tt

)
a

(t

) +

F
a
(t),

F
a
(t) = i

k
g
k

b
k
(0)e
i(
k
)t
,
we use the notation a(t) to replace a

(t) in the following formulations,

k
[g
k
[
2

0
dt

e
i(
k
)(tt

)
a(t

) +

F
a
(t),

F
a
(t) = i

k
g
k

b
k
(0)e
i(
k
)t
,
with the Weisskopf-Wigner theory, G(t) =

2
(t),

k
[g
k
[
2

0
dt

e
i(
k
)(tt

)
a(t

0
dt

G(t t

) a(t

)

2
a(t),
where G(t t

) =

k
[g
k
[
2
e
i(
k
)(tt

)
=

D()[g()[
2
e
i(
k
)(tt

)
,
D() is the density of state for the reservoir, and the equation of motion for the
eld interacting with the reservoir is
d
dt
a(t) =

2
a(t) +

F
a
(t),
there are dissipation and uctuation terms,
IPT5340, Fall 06 p.4/47
Commutation relation
for a Markovian process
d
dt
a(t) =

2
a(t) +

F
a
(t),

F
a
(t) = i

k
g
k

b
k
(0)e
i(
k
)t
,
if one dismiss the uctuation term

F
a
(t),
d
dt
a(t) =

2
a(t),
we have the solution a(t) = a(0)e
/2t
,
for the non-interacting eld, the commutation relation at t = 0 is [ a(0), a

(0)] = 1,
but as time evolves to t ,= 0, the commutation relation is not satised,
[ a(t), a

(t)] = exp(t),
the noise operator with appropriate correlation properties helps to maintain the
commutation relation at all time,
IPT5340, Fall 06 p.5/47
Thermal reservoir
the noise operator is dened as,

F
a
(t) = i

k
g
k

b
k
(0)e
i(
k
)t
,
suppose the reservoir is in thermal equilibrium,
< b
k
(0) >
R
=< b

k
(0) >
R
= 0
< b
k
(0)b
k
(0) >
R
= 0 < b

k
(0)b

(0) >
R
= 0
< b

k
(0)b
k
(0) >
R
= n
k

kk

< b
k
(0)b

(0) >
R
= ( n
k
+ 1)
kk

where
n
t
h = n) =

n
n
nn
=
1
exp(
k
/k
B
T) 1
,
IPT5340, Fall 06 p.6/47
Thermal reservoir
the noise operator is dened as,

F
a
(t) = i

k
g
k

b
k
(0)e
i(
k
)t
,
the Langevin noise operators have zero means

F
a
(t))
R
=

a
(t))
R
= 0,
but non-zero variances,

a
(t)

F
a
(t

))
R
=

g
k
g
k

b
k
)exp[i(
k
)t i(
k
)t

],
=

k
[g
k
[
2
n
k
exp[i(
k
)(t t

)],
= n
th
(t t

),
where n
th
= n
(
k
)
,
IPT5340, Fall 06 p.7/47
Fluctuation-dissipation theory
for a single-mode eld interacting with thermal reservoir,
d
dt
a(t) =

2
a(t) +

F
a
(t),
the Langevin noise operators have zero means but non-zero variances,

F
a
(t))
R
=

a
(t))
R
= 0,

a
(t)

F

a
(t

))
R
=

F
a
(t)

F
a
(t

))
R
= 0,

a
(t)

F
a
(t

))
R
= n
th
(t t

),

F
a
(t)

F

a
(t

))
R
= ( n
th
+ 1)(t t

),
the damping of the system is determined from the uctuating forces of the
reservoir, in other words, the uctuations induced by the reservoir give rise to the
dissipation in the system,
=
1
n
th

dt

a
(t)

F
a
(t

))
R
,
this is one formulation of the uctuation-dissipation theorem,
IPT5340, Fall 06 p.8/47
Mollows triplet: Resonance Fluorescence Spectrum
-
a
S
(

)
[
A
.
U
.
]
-1 -0.5 0 0.5 1
10
20
30
40
50
60
70
80
90
100

.
.
.
.
.
.
|n-1>
|1>
|2>
|0>
|n+1>

.
.
.
.
.
.
interaction
reservoir

a
system
|n>
Theory: B. R. Mollow, Phys. Rev. 188, 1969 (1969).
Exp: F. Y. Wu, R. E. Grove, and S. Ezekiel, Phys. Rev. Lett. 35, 1426 (1975).
elastic Rayleigh scattering and inelastic Raman scattering
IPT5340, Fall 06 p.9/47
Photon-Atom Interaction in PhCs
Reservoir Theory
-
a
S
(

)
[
A
.
U
.
]
-1 -0.5 0 0.5 1
10
20
30
40
50
60
70
80
90
100

.
.
.
.
.
.
|n-1>
|1>
|2>
|0>
|n+1>

.
.
.
.
.
.
interaction
reservoir

a
system
|n>
?
|1>
|2>

interaction
reservoir

a
system
IPT5340, Fall 06 p.10/47
Hamiltonian of our system: Jaynes-Cummings model
H =

2

z
+

k
a

k
a
k
+

2
(

e
i
L
t
+
+
e
i
L
t
)
+

k
(g
k

+
a
k
+ g

k
a

)
And we want to solve the generalized Bloch equations:

(t) = i

z
(t)e
it
+

d t

G(t t

)
z
(t)

(t

) + n

(t)

+
(t) = i

z
(t)e
it
+

d t

G
c
(t t

)
+
(t

)
z
(t) + n
+
(t)

z
(t) = i(

(t)e
it

+
(t)e
it
) + n
z
(t)
2

d t

[G(t t

)
+
(t)

(t

) + G
c
(t t

)
+
(t

(t)]
IPT5340, Fall 06 p.11/47
Remarks:
1. coupling constant:
g
k
g
k
(

d,

r
0
) = |d|
a

1
2
0

k
V

d E

k
(

r
0
)
2. memory functions:
G()

k
|g
k
|
2
e
i
k
t
()
G
c
()

k
|g
k
|
2
e
i
k
t
()
3. Markovian approximation:
G(t) = G
c
(t) = (t)
IPT5340, Fall 06 p.12/47
Quantum noise operators
n

(t) = i

k
g
k
e
i
k
t

z
(t)a
k
()
n
+
(t) = i

k
g

k
e
i
k
t
a
+
k
()
z
(t)
n
z
(t) = 2i

k
[g

k
e
i
k
t
a
+
k
()

(t) g
k
e
i
k
t

+
(t)a
+
k
()]
where the mean and the correlation functions of the reservoir before interaction,
< a
k
() >
R
=< a

k
() >
R
= 0
< a
k
()a
k

() >
R
= 0
< a

k
()a

() >
R
= 0
< a

k
()a
k

() >
R
= n
k

kk

< a
k
()a

() >
R
= ( n
k
+ 1)
kk

IPT5340, Fall 06 p.13/47

(t) = i

z
(t)e
it
+

d t

G(t t

)
z
(t)

(t

) + n

(t)

+
(t) = i

z
(t)e
it
+

d t

G
c
(t t

)
+
(t

)
z
(t) + n
+
(t)

z
(t) = i(

(t)e
it

+
(t)e
it
) + n
z
(t)
2

d t

[G(t t

)
+
(t)

(t

) + G
c
(t t

)
+
(t

(t)]
where the Langevin noise operators have zero means but non-zero variances,
n

(t))
R
= n
+
(t))
R
= n
z
(t))
R
= 0
n

(t)n

(t

))
R
= n
+
(t)n
+
(t

))
R
= 0
n

(t)n
+
(t

))
R
=

k
[g
k
[
2
( n
k
+ 1)e
i
k
(tt

z
(t)
z
(t

))
n
+
(t)n

(t

))
R
=

k
[g
k
[
2
n
k
e
i
k
(tt

z
(t)
z
(t

))
n
z
(t)n
z
(t

))
R
= 4

k
[g
k
[
2
[( n
k
+ 1)e
i
k
(tt

+
(t)

(t

)) + n
k
e
i
k
(tt

(t)
+
(t

))
IPT5340, Fall 06 p.14/47
Modeling DOS of PBCs
0
0.5
1
1.5
2
2.5
3
x 10
4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
D
O
S
Frequency
anisotropic model:
k
=
c
+ A|k k
i
0
|
2
D() =

c
A
3
(
c
)
S. Y. Zhu, et al., Phys. Rev. Lett. 84, 2136 (2000).
IPT5340, Fall 06 p.15/47
Memory functions of PBCs
anisotropic model:
k
=
c
+ A|k k
i
0
|
2
D() =

c
A
3
(
c
)
the memory functions under this anisotropic model
also can be derived:

G() =
3/2
i

c
+

G
c
() =
3/2
i

c
+

a
+
where
3/2
=

2
a
d
2
6
0
A
3/2
, and we have used the space
average coupling strength
3
8

d|d E|
2
in the
derivation.
IPT5340, Fall 06 p.16/47
Memory functions of PBCs
/
a
b
s
[
G
(

)
]
a
r
g
[
G
(

)
]
-100 0 100 200 300 400
0.05
0.06
0.07
0.08
0.09
0.1
-1.7
-1.6
-1.5
-1.4
-1.3
-1.2
-1.1
-1
-0.9
-0.8
-0.7
-0.6
-0.5
-0.4

c
Amplitude and phase spectrum of the memory function
with
a
=
c
= 100.
IPT5340, Fall 06 p.17/47
Correlations of noise operators at zero temperature
n

(
1
) n
+
(
2
)
R
= N(
1
)(
1
+
a

c
)(
1

2
)
n
z
(
1
) n
z
(
2
)
R
= N(
1
)[4(
1

2
) +
z
(
1

2
)
R
]
(
1
+
a

c
)
n
z
(
1
) n

(
2
)
R
= 0
n

(
1
) n
z
(
2
)
R
= N(
1
)

(
1

2
)
R
(
1
+
a

c
)
n
z
(
1
) n
+
(
2
)
R
= N(
1
)
+
(
1

2
)
R
(
1
+
a

c
)
n
+
(
1
) n
z
(
2
)
R
= 0
with N() 4
3/2

a
+
c

a
+
Quantum noises of the photonic bandgap reservoir are
not only color noises but also exhibit bandgap behaviour.
IPT5340, Fall 06 p.18/47
Liouville operator expansion

ij
(t) = e
i/(tt

ij
(t

) =

n=0
[i(t t

)]
n
n!
L
n

ij
(t

)
For zero-th order Liouville operator expansion, we get

(t) = i

z
(t)e
it

dt

G(t t

(t

) + n

(t)

+
(t) = i

z
(t)e
it

dt

G
c
(t t

)
+
(t

) + n
+
(t)

z
(t) = i(

(t)e
it

+
(t)e
it
)

dt

[G(t t

) + G
c
(t t

)](1 +
z
(t

)) + n
z
(t)
valid for the case of
atom with longer lifetime and under weak pumping
IPT5340, Fall 06 p.19/47
Solve the optical Bloch equations
Since it is now a linear problem, by using Fourier transform the modied optical Bloch
equations become:
/()

.() =

.
0
()
where
/() =

i( + ) +

G() 0 i

2
0 i( ) +

G
c
() i

2
i i i +

G() +

G
c
()

.() =

( + )

+
( )

z
()

.
0
() =

( + )
n
+
( )
2[

G() +

G
c
()]() + n
z
()

where n

(), n
+
(), n
z
(),

G(), and

G
c
() are Fourier transforms of n

(t), n
+
(t),
n
z
(t), G(t), and G
c
(t), respectively.
IPT5340, Fall 06 p.20/47
Solve the optical Bloch equations
The solutions are

( + ) =
(2g h +
2
) n

() +
2
n
+
() + ig n
z
() i2g[

G() +

G
c
()]()

2
(f + g) + 2f g h

+
( ) =

2
n

() + (2f h +
2
) n
+
() if n
z
() + i2f[

G() +

G
c
()]()

2
(f + g) + 2f g h

z
() =
2ig n

() 2if n
+
() + 2f g n
z
() 4f g[

G() +

G
c
()]()

2
(f + g) + 2f g h
where
f() = i i +

G()
g() = i + i +

G
c
()
h() = i +

G() +

G
c
()
IPT5340, Fall 06 p.21/47
Fluorescence spectrum
Because the two-time correlation function of the atomic dipole is proportional to
the rst order correlation function g
(1)
(), we can obtain the uorescence
spectrum by taking the Fourier transform of the rst order correlation function:
S() =

d g
(1)
()e
i

+
()

())
R
It should be noted that here we cannot directly apply the quantum regression
theorem since it is invalid for non-Markovian process.
IPT5340, Fall 06 p.22/47
Mollows triplet
As a check, we rst use our formulation to calculate the free space case.
At free space, one can assume the memory functions are delta functions since

k
[g
k
[
2
e
i
k
t
= (t) with being the decay rate of the excited atom.
The noise correlation functions at zero temperature are also delta-function
correlated (i.e., white noises).
Therefore, the uorescence spectrum at steady state is given by:

+
()

())
R
=

2
(

2
4
+
2
)

2
2
+
2
+

2
4
( + )
+

4
(

2
2
+
2
+ ( + )
2
)
2(

2
2
+
2
+

2
4
)[
2
(

2
2
+
2
+

2
4
2( + )
2
)
2
+ ( + )
2
(
2
+
2
+
5
4

2
( + )
2
)
2
]
IPT5340, Fall 06 p.23/47
Mollows triplet
In the limit of strong on-resonance pumping ( , = 0), Eq.(1) can be
reduced to:

+
()

())
R
= 22

2
4
2
() +
3
16

( + )
2
+
9
16

2
+
1
4

2
+
1
4

2
+
3
16

( )
2
+
9
16

Then, the resonance uorescence spectrum exhibits the Mollow triplets for white
noise: three Lorentzian proles with peaks in the ratio 1 : 3 : 1, and widths of
3
2
,
, and
3
2
.
-
a
S
(

)
[
A
.
U
.
]
-1 -0.5 0 0.5 1
10
20
30
40
50
60
70
80
90
100

.
.
.
.
.
.
|n-1>
|1>
|2>
|0>
|n+1>

.
.
.
.
.
.
interaction
reservoir

a
system
|n>
IPT5340, Fall 06 p.24/47
Resonance uorescence spectra near the band-edge
0
50
100
150
200
S
(

)
[
A
.
U
.
]
0
0.5
1
(
a
-
c
)
/

-0.4
-0.2
0
0.2
0.4
(

a
)
/
= 0.25

c
= 100
DOS
w Wc
Wa
IPT5340, Fall 06 p.25/47
Quadrature spectra
Dene quadrature eld operator as:

(t) = e
i

E
(+)
(t) + e
i

E
()
(t)
= 0 (

2
) are the in-phase (out-of-phase) quadrature elds.
Then the corresponding spectra with normally order
variance is:
S

() <

E

(),

E

() >

1
4
[<

()

() > e
2i
+ <
+
()

() >
+ <
+
()

() > + <
+
()
+
() > e
2i
]
IPT5340, Fall 06 p.26/47
Quadrature spectra in free space
= 5.0 = 2.0
= 1.0 = 0.5
Theory: D. F. Walls and P. Zoller, Phys. Rev. Lett. 47, 709 (1981).
Theory: L. Mandel, Phys. Rev. Lett. 49, 136 (1982).
IPT5340, Fall 06 p.27/47
Observation of squeezing uorescence spectra
with
147
Yb atoms
Exp: Z. H. Lu, S. Bali, and J. E. Thomas, Phys. Rev. Lett. 81, 3635 (1998).
IPT5340, Fall 06 p.28/47
Fluorescence quadrature spectra near the band-edge
(-
a
)/
Q
u
a
d
r
a
t
u
r
e
S
p
e
c
t
r
u
m
-1 -0.5 0 0.5
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
= 0.25

c
= 100
-0.1
0
0.1
0.2
0.3
S
1
(

)
[
A
.
U
.
]
0.2
0.4
0.6
0.8
1
1.2
1.4
(
a
-
c
)
/

-1
-0.5
0
0.5
1
(
-
a
)/
= 0.25

c
=100
DOS
w Wc
Wa
IPT5340, Fall 06 p.29/47
Resonance uorescence quadrature spectra near the band-edge
1. Suppression and enhancement of the relative uorescence
peak amplitudes varied at different wavelength offsets.
2. Squeezing occurs in the out-of-phase quadrature for free
space when
2
< 4
2
.
3. Squeezing occurs in the in-phase quadrature for PhCs when

2
> 4
2
.
4. Resonance uorescence squeezing spectra come from the
interference between two sidebands of Mollows triplet.
R.-K. Lee and Y. Lai, J. Opt. B, 6, S715 (2004).
IPT5340, Fall 06 p.30/47
Density operator method
Heisenberg-Langevin equation is a directly correspondence to classical
description of stochastic system,
in many cases, Heisenberg-Langevin equation is nonlinear,
in general it is extremely difcult to deal with the Heisenberg-Langevin equation,
another method is to develop a Schrdinger or interaction picture analysis,
in this way we want to use a linear deterministic differential equation for the reduced
system density operator,
naturally, as the quantum system is open, there is statistical as well as quantum
uncertainty and a true wave function description is no longer possible,
Ref:
Ch. 7 in Mesoscopic Quantum Optics, by Y. Yamamoto and A. Imamoglu.
Quantum Noise, by C. W. Gardiner.
IPT5340, Fall 06 p.31/47
Master equation
we consider a system S interacting with a reservoir R via the interaction
Hamiltonian

V ,
the combined density operator is denoted by (t),
assume that at an initial time t = 0, the two systems are uncorrelated,
(t = 0) =
S
(0)
R
(0),
in the interaction picture, the dynamics of (t) is
d
dt
(t) =
1
i
[

H
I
(t), (t)],
since the number of degrees of freedom of the reservoir is very large, it is
impossible to keep track of its quantum evolution,
we can only focus on the system with a reduced density operator, by tracing over
the reservoir degrees of freedom,
d
dt

S
(t) =
1
i
Tr
R
([

H
I
(t), (t)]),
IPT5340, Fall 06 p.32/47
Master equation
we can only focus on the system with a reduced density operator,
d
dt

S
(t) =
1
i
Tr
R
([

H
I
(t), (t)]),
where
d
dt
(t) =
1
i
[

H
I
(t), (t)],
without any approximation, the master equation for the reduced density operator is,
d
dt

S
(t) = (
1
i
)
2

t
0
dt

Tr
R
([

H
I
(t), [

H
I
(t

), (t

)]]) +
1
i
Tr
R
([

H
I
(t), (0)]),
since Tr
R
([

H
I
(t), (0)]) vanish for all the interaction Hamiltonians of interest in
quantum optics, we have the master equation,
d
dt

S
(t) = (
1
i
)
2

t
0
dt

Tr
R
([

H
I
(t), [

H
I
(t

), (t

)]]),
IPT5340, Fall 06 p.33/47
Born-Markov approximation
the master equation for the reduced density operator, without any approximation,
d
dt

S
(t) = (
1
i
)
2

t
0
dt

Tr
R
([

H
I
(t), [

H
I
(t

), (t

)]]),
four key approximations used in the following,
1. Rotating-wave approximation,
2. Born approximation,
(t

) =
S
(t

)
R
(t

),
3. the initial radiation eld density operator commutes with the free Hamiltonian
and the reservoir is not affected by the interaction with the system,

R
(t) = Tr
S
[ (t)] =
R
(0),
4. Markov approximation,

S
(t

)
S
(t),
IPT5340, Fall 06 p.34/47
Master equation
the master equation for the reduced density operator, without any approximation,
d
dt

S
(t) = (
1
i
)
2

t
0
dt

Tr
R
([

H
I
(t), [

H
I
(t

), (t

)]]),
with Born-Markov approximation, the master equation becomes
d
dt

S
(t) = (
1
i
)
2

t
0
dt

Tr
R
([

H
I
(t), [

H
I
(t

),
S
(t)
R
(0)]]),
atom damping by eld reservoirs,
eld damping by eld reservoirs,
eld damping by atomic reservoirs,
IPT5340, Fall 06 p.35/47
Atom damping by eld reservoirs
consider a two-level atom damped by a eld reservoir in free space,
the interaction Hamiltonian is

H
I
=

k
(g
k

k
e
i(
k
)t
+ H. C),
assume the reservoir density operator is a multimode thermal eld,

R
=

n
exp(

k
n
k
B
T
)
1 exp(

k
n
k
B
T
)
[n)
kk
n[,
the equation of motion for the reduced density operator Tr
R
[ (t)]
a
(t) is,
d
dt

a
(t) = (
1
i
)
2

t
0
dt

Tr
R
([

H
I
(t), [

H
I
(t

),
a
(t)
R
(0)]]),
=
1
2
n
th
[

+

a

+

a

] + (n
th
+ 1)[
+

a

+
]) + H. C,
IPT5340, Fall 06 p.36/47
Field damping by eld reservoirs
consider a single-mode eld in a cavity with a nite leakage rate,
assume the reservoir density operator is a multimode thermal eld,

R
=

n
exp(

k
n
k
B
T
)
1 exp(

k
n
k
B
T
)
[n)
kk
n[,
the equation of motion for the reduced density operator Tr
R
[ (t)]
f
(t) is,
d
dt

f
(t) = (
1
i
)
2

t
0
dt

Tr
R
([

H
I
(t), [

H
I
(t

),
f
(t)
R
(0)]]),
=

t
t
0
dt

k
g
2
k
n
th
[ a a

f
(t

) a

f
(t

) a]e
i(
k
)(tt

)
+(n
th
+ 1)[ a

a
f
(t

) a
f
(t

) a

])e
i(
k
)(tt

)
+ H. C,
IPT5340, Fall 06 p.37/47
Field damping by eld reservoirs
the equation of motion for the reduced density operator Tr
R
[ (t)]
f
(t) is,
d
dt

f
(t) =

t
t
0
dt

k
g
2
k
n
th
[ a a

f
(t

) a

f
(t

) a]e
i(
k
)(tt

)
+(n
th
+ 1)[ a

a
f
(t

) a
f
(t

) a

])e
i(
k
)(tt

)
+ H. C,
again, by replacing sum
k
g
2
k
with the integral

d
k
D(
k
)g(
k
)
2
, and

t
t
0
dt

k
g
2
k
e
i(
k
)(tt

)
=

t
t
0
dt

d
k
D(
k
)g(
k
)
2
e
i(
k
)(tt

)
,

d
k
D(
k
)g(
k
)
2
(
k
),
D()g()
2

1
2
(

Q
e
),
where /Q
e
is the cavity photon decay rate due to leakage (output coupling) via a
partially reecting mirror,
IPT5340, Fall 06 p.38/47
Field damping by eld reservoirs
the equation of motion for the reduced density operator Tr
R
[ (t)]
f
(t) is,
d
dt

f
(t) =
1
2
(

Q
e
)n
th
[ a a

f
(t

) a

f
(t

) a] + (n
th
+ 1)[ a

a
f
(t

) a
f
(t

) a

]),
+H. C,
compared to the case of atom damping by eld reservoirs,
d
dt

a
(t) =
1
2
n
th
[

+

a

+

a

] + (n
th
+ 1)[
+

a

+
]) + H. C,
IPT5340, Fall 06 p.39/47
Field damping by atomic reservoirs
consider the damping of an optical cavity mode by a two-level atomic beam
reservoir,
this is the reverse problem of a laser,
the statistics of the atomic reservoir is determined by the Boltzmann distribution,

R=atom
(t = 0) =

aa
0
0
bb

=
aa
[a)a[ +
bb
[b)b[,
where

aa
=
1
1 + exp(
0
/k
B
t)
, and
aa
=
exp(
0
/k
B
t)
1 + exp(
0
/k
B
t)
,
assume there is no quantum coherence between the upper and lower states,

ab
=
ba
= 0,
IPT5340, Fall 06 p.40/47
Field damping by atomic reservoirs
the interaction Hamiltonian for a single atom is

H
I
= g(

+
+
a) = g

0 a
a

,
at t = 0, the atom-eld density operator is,
(t) =
f
(t)
R
=

aa

f
(t) 0
0
bb

f
(t)

,
the terms for the commutator are
[

H
I
, [

H
I
, (t)]] = g
2

a a

aa

f
(t) a
bb

f
(t) a

0
0 a

a
bb

f
(t) a

ee

f
(t) a

+H. C,
IPT5340, Fall 06 p.41/47
Field damping by atomic reservoirs
the equation of motion for the reduced density operator Tr
R
[ (t)]
f
(t) is,
d
dt

f
(t) = (
1
i
)
2

t
0
dt

Tr
R
([

H
I
(t), [

H
I
(t

),
f
(t)
R
(0)]]),
assume that r atoms are injected into the cavity per second,
and they spend an average time of seconds inside the cavity, i.e.

0
dt

rt

=
1
2
r
2
,
then,
d
dt

f
(t) =
1
2
R
e
[ a a

f
a

f
a]
1
2
R
g
[ a

a
f
a
f
(t) a

] + H.C,
where
R
e
= r
aa
g
2

2
, and R
g
= r
bb
g
2

2
,
IPT5340, Fall 06 p.42/47
Field damping by atomic reservoirs
the equation of motion for the reduced density operator Tr
R
[ (t)]
f
(t) is,
d
dt

f
(t) =
1
2
R
e
[ a a

f
a

f
a]
1
2
R
g
[ a

a
f
a
f
(t) a

] + H.C,
where
R
e
= r
aa
g
2

2
, and R
g
= r
gg
g
2

2
,
R
e
is the rate coefcient for photon emission by atoms per second,
R
g
is the rate coefcient for photon absorption by atoms per second,
the cavity photon decay rate

Q
0
and the thermal equilibrium photon number n
th
are dened by

Q
0
R
g
R
e
, and R
e
(1+n
th
) = R
g
n
th
n
th
=
R
e
R
g
R
e
=
1
exp(/k
b
T) 1
,
the later one condition gives the thermal equilibrium photon number n
th
,
R
e
(1 + n
th
) is the sum of spontaneous and stimulated emission rate per second,
R
g
n
th
is (stimulated) absorption rate per second,
IPT5340, Fall 06 p.43/47
Field damping by atomic reservoirs
the equation of motion for the reduced density operator Tr
R
[ (t)]
f
(t) is,
d
dt

f
(t) =
1
2
R
e
[ a a

f
a

f
a]
1
2
R
g
[ a

a
f
a
f
(t) a

] + H.C,
=
1
2

Q
0
n
th
[ a a

f
a

f
a] + (n
th
+ 1)[ a

a
f
a
f
(t) a

] + H.C,
where we use that R
e
=

Q
0
n
th
and R
g
=

Q
0
(n
th
+ 1),
the diagonal elements of the reduced density matrix are
d
dt

n,n
(t) =

Q
0
[n
th
(n + 1) (n
th
+ 1)n]
n,n
n
th

n1,n1
(n
th
+ 1)(n + 1)
n+1,n+1
,
= [R
e
(n + 1) R
g
n]
nn
+ R
e
n
n1,n1
+ R
g
(n + 1)
n+1,n+1
,
IPT5340, Fall 06 p.44/47
Detailed balance
equilibrium is obtain when the net ow between all pairs of level vanishes,
R
g
n
n,n
= R
e
n
n1,n1
, or
n,n
=
n
th
n
th
+ 1

n1,n1
,
is condition is referred to as detailed balance,
the solution for detailed balance is,

n,n
= [1 exp(/k
B
T)]exp(n/k
B
T),
with n
th
=
1
exp(/k
b
T)1
,
detailed balance in this case gives the thermal (Bose-Einstein) distribution with an
average photon number,
n) =

n,n
n =
1
exp(/k
B
T) 1
= n
th
IPT5340, Fall 06 p.45/47
Detailed balance
detailed balance in this case gives the thermal (Bose-Einstein) distribution with an
average photon number,
n) =

n,n
n =
1
exp(/k
B
T) 1
= n
th
this is the result we use for the thermal radiation eld,
although the led
f
may initially be in a pure state, the process of tracing over the
(unobserved) atomic states leads to a eld in a mixed state,
f
=

n

n,n
[n)n[,
the effect of the atomic beam is to bring the eld to the same temperature as that of
atoms,
IPT5340, Fall 06 p.46/47
Reservoir, decoherence, and measurement
the reservoir theory lies in the process of tracing over the reservoir coordinates,
which induces dissipation and decoherence of the system,
at the same time, this is an irreversible dynamics for the system,
this process corresponds to the lack of measurement as to whether the atom is in
the upper level or in the lower level after interaction with the eld,
if the initial and nal states of the atom are know, i.e if the information concerning
the atomic beam is not discarded, the eld remains in a pure state,
the primary difference between the reservoir and quantum measurement theories
is whether information stored in the environment (reservoir) that interacts with
system is discarded or read out.
IPT5340, Fall 06 p.47/47