The Interaction of Light and Matter: and n

The interaction of light and matter is what makes life interesting.

Everything we see is the result of this interaction. Why is light absorbed or
transmitted by a particular medium?
Light causes matter to vibrate. Matter in turn emits light, which interferes
with the original light.
Destructive interference means absorption.
~ 90 out-of-phase interference
changes the phase velocity of light,
or refractive index.
Excited atoms & the forced oscillator

Refractive
index

Absorption
coefficient

Complex Lorentzian: 1/(- i)

Absorption coefficient, ,
and refractive index, n.

n1

Adding complex amplitudes

When two waves add together with the same complex exponentials,
we add the complex amplitudes, ~E0 + E
~ 0'.
Constructive
interference:
1.0

Destructive
interference:
1.0

1.2

1.0

+
0.2

-0.2

Laser

+
-0.2i

=
0.8

time

Quadrature phase: 90
interference:

1-0.2i

time
Absorption

time
Slower phase velocity

Light excites atoms, which emit light that

adds (or subtracts) with the input light.
When light of frequency excites an atom with resonant frequency 0:
Electric field
at atom

E (t )

Electron
cloud

r
xe (t )

Emitted
field

r
E (t )

Incident light
+
=

On resonance (= 0)

Emitted light
Transmitted light

An excited atom vibrates at the frequency of the light that excited it

and re-emits the energy as light of that frequency.
The crucial issue is the relative phase of the incident light and this
re-emitted light. For example, if these two waves are ~180 out of
phase, the beam will be attenuated. We call this absorption.

What do you think is happening here?

The Forced Oscillator

When we apply a periodic force to a natural oscillator (such as a
pendulum, spring, swing, or atom), the result is a forced oscillator.
Examples:
Child on a swing being pushed
Pushed pendulum
Bridge in wind or an earthquake
Electron in a light wave
Nucleus in a light wave

Tacoma Narrows Bridge

collapsing because
oscillatory winds blew at its
resonance frequency.

The forced oscillator is one of the most important problems in

physics. It is the concept of resonance.

The Forced Oscillator:

The relative phase
of the oscillator
motion with
respect to the
input force
depends on the
frequencies.
Let the oscillators
resonant frequency be 0,
and the forcing frequency
be .
We could think of the
forcing function as a light
electric field and the
oscillator as a nucleus of
an atom in a molecule.

Force

Oscillator

Below
resonance
<< 0

Weak
vibration.
In phase.

On
resonance
= 0

Strong
vibration.
90 out of
phase.

Above
resonance
>> 0

Weak
vibration.
180 out
of phase.

The relative phase

of an electron
clouds motion with
respect to input
light depends on
the frequency.
Let the atoms resonant
frequency be 0, and
the light frequency be .
The electron charge is
negative, so theres a
180 phase shift in all
cases (compared to the
previous slides plots).

Electric field Electron

at atom
cloud

Below
resonance
<< 0

Weak
vibration.
180 out
of phase.

On
resonance
= 0

Strong
vibration.
-90 out
of phase.

Above
resonance
>> 0

Weak
vibration.
In phase.

The relative
phase of
emitted light
with respect to
the input light
depends on the
frequency.
The emitted light is
90 phase-shifted
with respect to the
atoms motion.

Electric field Electron Emitted

at atom
cloud
field

Below
resonance
<< 0

Weak
emission.
90 out of
phase.

On
resonance
= 0

Strong
emission.
180 out
of phase.

Above
resonance
>> 0

Weak
emission.
-90 out
of phase.

The Forced Oscillator: Math

Consider an electron on a spring with position xe(t), and driven by
a light wave, E
~ 0 exp(-it):

me d 2 xe / dt 2 me02 xe eE0 exp(it )

%
The solution is:

e / me

E0 exp(i t )
%

xe (t )

2
0

E (t )

xe (t )

So the electron oscillates at the incident light wave frequency ( ),

but with an amplitude that depends on the difference between the
frequencies.

Checking our
solution

me d 2 xe / dt 2 me02 xe eE0 exp( it )

Substitute the solution for xe(t)

into the forced oscillator
equation to see if it works.

me
E
exp(

t
)

0
e 0
2
2
0

e / me

1
2

0
2
2
0

e / me

e / me

xe (t )

2
0

e / me

2
0

me

e
0
2
2
0

E0 exp(it ) eE0 exp(it )

e / me

2
0

E0 exp(i t )

1
2 1
2
0

02 2

1
2
2
0

QED

The problem with this model

xe (t )

e / me

2
0

Exactly on resonance,
when = 0, xe goes
to infinity.

This is unrealistic.

Well need to fix this.

E0 exp(i t )

E (t )

xe (t )

The Damped Forced Oscillator

Our solution has infinite amplitude on resonance, which is unrealistic.
We fix this by using a damped forced oscillator: a harmonic oscillator
experiencing a sinusoidal force and viscous drag.
We must add a viscous drag term:

me

dxe
dt

d 2 xe
dxe
me 2 me
me02 xe eE0 exp(i t )
dt
dt
The solution is now:

(e / me )
xe (t )
E (t )
2
2
(0 i )
The electron still oscillates at the light frequency and with a
potential phase shift, but now with a finite amplitude for all .

Why we included the damping factor,

Atoms spontaneously decay to the ground state after a time.
Also, the vibration of a medium is the sum of the vibrations of all
the atoms in the medium, and collisions cause the sum to cancel.
Collisions

Collisions dephase
the vibrations, causing
cancellation of the
total medium
vibration, typically
exponentially.

(1)
Atom #1 xe (t )
(2)
Atom #2 xe (t )
(3)
Atom #3 xe (t )

(We can use the same

argument for the
emitted light, too.)

Sum

time

The atoms response is approximately

a Complex Lorentzian.
Consider:

xe (t )

e / me
e / me

02 2 i (0 )(0 ) i

Assuming 0, this becomes:

e / me

2 (0 ) i

e / me
1
2 (0 ) i / 2

In terms of the variables 0 and = , the function

1/( i), is called a Complex Lorentzian. Its real and imaginary
parts are:

1
1 i

2
2
2
i i i 2

Complex Lorentzian

Real

Imaginary

1
1 i

2 2 2 2
i
i i

1

Imaginary
(even)
component

Real (odd)
component

Damped forced oscillator solution for

light-driven atoms
The forced-oscillator response is sinusoidal, with a frequencydependent strength that's approximately a complex Lorentzian:
Here, e < 0.

e
1
1
E
(
t
)

E (t )

(0 i / 2)
2 m e (0 i / 2)

xe (t )

When << 0, the electron vibrates

180 out of phase with the light wave:

1
xe (t )
E (t ) E (t )
(0 )

When = 0, the electron vibrates

-90 out of phase with the light wave:

xe (t )

When >> 0, the electron vibrates

in phase with the light wave:

1
E (t ) iE (t )

(i / 2)

1
E (t ) E (t )

( )

xe (t )

The relative phase

of an electron
clouds motion with
respect to input
light depends on
the frequency.

Recall that the

atoms resonant
frequency is 0,
and the light
frequency is .

Electric field Electron

at atom
cloud

Below
resonance
<< 0

Weak
vibration.
180 out
of phase.

On
resonance
= 0

Strong
vibration.
-90 out
of phase.

Above
resonance
>> 0

Weak
vibration.
In phase.

Okay, so weve determined what the

light wave does to the atom.

Now, what does the atom do to the

light wave?

Re-emitted light from an excited atom

interferes with original light beam
The re-emitted light may
interfere constructively,
destructively, or, more
generally, somewhat out of
phase with the original light
wave.
We model this process by
considering the total electric
field,

E (z,t) = E

original

(z,t) + E

z
Incident light
+
Emitted light
=
Transmitted light

re-emitted

(z,t)

Maxwell's Equations will allow us to solve for the total field, E (z,t).
The input field will be the initial condition.

The Inhomogeneous Wave Equation

r

The induced polarization,P , is due to the medium:

E
1 E
P
2
0
2
2
z
c t
t 2
2

where:
r

r
P (t ) N e xe (t )

and e is the electron charge, and

N = the electron number density.

P (t )

e
1

E (t )
2 m e (0 i / 2)

xe (t )

For our
vibrating
electrons:

e
1

E0 exp[i (kz t )]
2 m e (0 i / 2)

P ( z, t ) Ne

P0

E(z,t)

The electric-field amplitude depends on z.

The effect of the medium will be to change the field complex amplitude with
distance. And because the polarization depends on E, its amplitude, P0, will also.
Constant in time

E ( z , t ) E0 ( z ) exp i kz t

and P ( z , t ) P0 ( z ) exp
i kz t

Specifically, the envelopes, E0(z) and P0(z), are assumed to vary

slowly; the fast variations will all be in the complex exponential.
The time derivatives are easy (as before, they just multiply by a
factor of 2) because the envelopes are independent of t:

1 2E 2
2 2 2 E
c t
c

2 P
0 2 0 2 P
t

The Slowly Varying Envelope Approximation

But the 2E/ z2 derivative is trickier.

E ( z , t ) E0 ( z ) exp i kz t
The z-derivatives:

E ( z , t ) E0

ik E0 ( z ) exp i kz t
z
z

E0
2 E ( z , t ) 2 E0
E0

ik
exp

i
kz

ikE
i kz t

0 ik exp

2
2

z
z
z

E0
2 E ( z , t ) 2 E0
2

2
ik

k
E0 exp i kz t

2
2
z
z
z

Because variations of the envelope, E0(z), in space will be slow,

well neglect the 2nd z-derivative.

SVEA continued
Substituting the derivatives into the inhomogeneous wave equation:

2 E 1 2 E
2 P
2 2 0 2
2
z
c t
t

E0
2
2
2
2
ik

k
E

E
exp
i
kz

P0 exp

i kz t

0
0
0

z
c

Now, use k = /c. And canceling the

complex exponentials leaves:

2ik

E0
0 2 P0
z

E0 0 2
0 ( k 2 / 0 0 )
ik

P0
P0
P0
z
2ik
2ik
2 0

Why the re-emitted light is 90 out of phase

with the polarization
E0
ik

P0
z 2 0
Usually, P0 = P0 (E0), and hence P0(z), too.
But consider for the moment P0 ~ constant.
Converting to finite differences, the re-emitted field is just E0, and it
will be:

ik
E0
P0 z
2 0

Note the i, which means that the re-emitted

field has a 90 phase shift with respect to the
polarization and hence a -90 phase shift with
respect to the electron cloud motion.

The re-emitted
wave leads the
electron cloud
motion by 90
in phase.
This phase shift
adds to the
potential phase shift
of the electron
cloud motion with
respect to the input
light.

Electric field Electron Emitted

at atom
cloud
field

Below
resonance
<< 0

Weak
emission.
90 out of
phase.

On
resonance
= 0

Strong
emission.
180 out
of phase.

Above
resonance
>> 0

Weak
emission.
-90 out
of phase.

Solving for the slowly varying envelope

Define , the susceptibility:

e
1
P0 Ne

E0 0 E0
2 m e (0 i / 2)
Substituting for P0 ,

k
z
2

The solution: E0 ( z ) E0 (0) exp i

Define new quantities
for the real and
imaginary parts of :
so that:

E0
k
i E0
z
2
k

E0 (0) exp
i
i
Im

Re

E0
ik

P0 becomes:
z 2 0

k Im

1
(n 1) Re
2

E0 ( z ) E0 (0) exp [ / 2 i(n 1) k ] z

where is the absorption coefficient and n is the refractive index.

The complete electric field in a medium:

The electromagnetic wave in the medium becomes (combining the
slowly varying envelope with the complex exponential):

E ( z , t ) E0 (0) exp [ / 2 i (n 1)k ]z exp[i (kz t )]

Simplifying:

E0(z)

E ( z , t ) E0 (0) exp[( / 2) z ] exp[i (nkz t )]

Absorption
attenuates the field

Refractive index
changes the k-vector

To summarize, in a medium:

E0 ( z ) E0 (0) exp[( / 2) z ]

k nk and / n

A light wave in a medium

Vacuum (or air)

n=1

Medium
n

n=2
Absorption depth = 1/

k0

nk0
Wavelength decreases

E ( z, t ) E0 (0) exp[i (k0 z t )]

E0 (0) exp[( / 2) z ] exp[i( nk0 z t )]

The speed of light, the wavelength (and k), and the amplitude change,
but the frequency, , doesnt change.

Absorption Coefficient and the Irradiance

The irradiance is proportional to the (average) square of the field.
Since E(z)

exp(-z/2), the irradiance is then:

I(z) = I(0) exp(-z) = I(0) exp(-z)

We often separate
out the factor of
from
where is the
absorption
cross-section.

where I(0) is the irradiance at z = 0, and I(z) is the irradiance at z.

Thus, due to absorption, a beams irradiance exponentially
decreases as it propagates through a medium.
The 1/e distance, 1/, is a rough measure of the distance light can
propagate into a medium (the penetration depth).

Refractive index and Absorption coefficient

n comes from the real part of :

Ne 2 / me
k
k
(n 1) k
Re
Re

2 0
2 0
2 (0 i / 2)

comes from the imaginary part of :

Ne 2 / me
k
k
/2
Im
Im

2 0
2 0
2 (0 i / 2)
Simplifying:

(0 )

2
2
(

/
2)
0

Ne 2
/2
N

2 0 c0 me (0 ) 2 ( / 2) 2
Ne 2
n 1
4 0 me

These
results are
valid for
small values
of these
quantities.

Refractive index and Absorption coefficient

Ne2
/2
N
2 0c0 me (0 ) 2 ( / 2) 2

0
Ne 2
n 1
4 0 me (0 ) 2 ( / 2) 2
Recall that we
approximated the
actual response,
which was:

Actual (no
approximation)

Refractive
index

n 1

Absorption
coefficient

0
Actual n 1 (no
approximation)

1
2
0 2 i
But the Lorentzian
is good enough for
most purposes.

The difference effects are called precursors, and theyre interesting.