You are on page 1of 17

UIUC Physics 436 EM Fields & Sources II

Lect. Notes 8

Prof. Steven Errede

LECTURE NOTES 8
A More Sophisticated Treatment of EM Wave Propagation in Conducting Media
In the previous P436 Lecture Notes 7, we discussed the propagation of EM waves in
conducting media (e.g. metals), taking into account Ohms Law J ( r , t ) = C E ( r , t ) and the
continuity equation i J ( r , t ) = free ( r , t ) t in the conductor.
However, this treatment of conducting media neglected certain inertial effects associated with
the free / conduction electrons in the metal conductor we need to consider more carefully the
actual motion of the free / conduction electrons in the conductor, and their response e.g. to the
application of a monochromatic plane EM wave of angular frequency .
Since the free / conduction electrons in a metal are not bound to individual atoms in a
conductor, there are no elastic restoring forces acting on the free / conduction electrons
{i.e. ke = 0 }, as there was in the case of the polarization of bound atomic electrons, e.g. in a
dispersive, linear non-conducting medium. Thus, the differential equation describing the motion
of the free / conduction electrons in a metal is of the general form:
Fe ( r , t ) = me a ( r , t ) = me

dv ( r , t )
+ me v ( r , t ) + 0 r ( t ) = eE ( r , t )
dt

where the driving force eE ( r , t ) = the charge of the electron ( e ) times the electric field
E ( r , t ) of the monochromatic plane EM wave propagating in the conducting metal.
Thus, we obtain a first-order inhomogeneous differential equation of the form:
me

dv ( r , t )
+ me v ( r , t ) = eE ( r , t )
dt

n.b. here again, we neglect the effects of the magnetic Lorentz force term,

FLorentz ( r , t ) = ev ( r , t ) B ( r , t )

eE ( r , t ) because: v = v vD

v prop .

Note also that me = mass of the free /conduction electron in the metal conductor which,
depending on the microscopic/quantum mechanical details of the conductor is {often} not equal
to the mass of a {truly free, isolated} electron, me = 9.11031 kg .

The quantity me is known as the frictional / Joule dissipation constant, associated with
electron energy / momentum losses i.e. free /conduction electron scattering losses in the
conductor. The damping constant {n.b. units of angular frequency (radians/sec)} is associated
with dissipative/ energy loss mechanism(s) of the motion of free / conduction electrons in the
metal free / conduction electrons scattering off of each other, off of atoms, lattice vibrations,
crystal defects and impurities in real metals. Thus, we can also define a relaxation time constant
rlx 1 associated with the free / conduction electrons present in the metal conductor.
Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

Lect. Notes 8

Prof. Steven Errede

The free current density in the metal conductor is J free ( r , t ) = ne ev ( r , t ) {Amps/m2}, where:
ne = free / conduction electron number density in the metal conductor {#/unit volume = #/m3}.
q = e = the electric charge of free / conduction electrons {Coulombs}.
v ( r , t ) = the {drift} velocity ( c ) of free / conduction electrons in metal at ( r , t ) {m/s}.

me

Thus:

dv ( r , t )
+ me v ( r , t ) = eE ( r , t ) multiply through by ne e :
dt

dv ( r , t )
ne eme v ( r , t ) = + ne e 2 E ( r , t )
dt
dv ( r , t )

2
= me ne e
me ( ne ev ( r , t ) ) = + ne e E ( r , t )
dt

ne eme

= me ( ne ev ( r , t ) ) + me ( ne ev ( r , t ) ) = + ne e 2 E ( r , t )
dt

me

dJ free ( r , t )
dt

+ me J free ( r , t ) = ne e 2 E ( r , t ) or:

dJ free ( r , t )
dt

n e2
+ J free ( r , t ) = e E ( r , t )
me

which is also a first-order linear, inhomogeneous differential equation.

Recall that the solution to an inhomogeneous differential equation is the sum of the solution
to the corresponding homogeneous differential equation, plus a particular solution satisfying the
inhomogeneous differential equation. Thus, we first need to obtain the solution to the
homogeneous differential equation, i.e. when no EM waves are present in the conducting
material, vis. E ( r , t ) = 0 .
a.) If E ( r , t ) = 0 , the general solution to the first-order linear, homogeneous differential
equation:

dJ free ( r , t )
dt

+ J free ( r , t ) = 0

is of the form of a damped exponential (because > 0):

J free ( r , t ) = J o free ( r ) e t = J o ( r ) e t rlx

where: rlx 1

Transient free current(s) will decay to 1 e = e 1 = 0.3679 of their initial t = 0 value ( J o ( r ) )

in a characteristic relaxation time t = rlx 1 .

Lect. Notes 8

Using the continuity equation i J free ( r , t ) = free ( r , t ) t {expressing conservation of free

electric charge} in the above homogeneous equation, with E ( r , t ) = 0 :

n e2
d
i J free ( r , t ) + i J free ( r , t ) = e i E ( r , t )

dt

me
=0

thus:

2 free ( r , t )
t 2

2 free ( r , t )
t

free ( r , t )
t

free ( r , t )
t

=0

=0

free ( r , t )
t

+ free ( r , t ) = 0

free ( r , t ) = o free ( r ) e t = o free ( r ) e t rlx .

But from P436 Lecture Notes 7, p. 1 we also obtained the relation free ( r , t ) = o free ( r ) e C t .
Thus comparing these two equations, we see that: = C = 1 rlx .
b.) For a static non-zero electric field E ( r ) = Eo , then this static problem can have no explicit
time dependence, hence dJ free ( r , t ) dt = 0 and thus the first-order linear, inhomogeneous
equation becomes:
ne e 2
ne e 2
ne e 2
ne e 2
E
r
E
J
r
E
r
constant
=
=
=
=
or:
(
)
(
)
(
)
o

Eo = constant
free

me
me
me
me

J free ( r ) =

The static / DC continuity equation for free charge is: i J free ( r ) =

free ( r )

t
with static solution(s): J free ( r ) = constant = fcn ( r ,{or t}) and free ( r ) fcn ( t ) .

=0

However, Ohms Law (for DC / steady free electric currents) is: J free ( r ) = C E ( r ) where:

C = the static conductivity of the metal (Siemens/m) = 1 C 1/static resistivity ( -m ).

n e2
n e2
Thus, we see here that: J free ( r ) = e E ( r ) = C E ( r ) i.e. the static conductivity C = e
me
me
1 me
=
or equivalently, the static resistivity: C
. {n.b. both purely real quantities}
C ne e 2

m
= e 2
(Ohms)
A ne e A

{again, a purely real quantity}

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

Lect. Notes 8

Prof. Steven Errede

c.) For a harmonically time-dependent {in general, complex} electric field (e.g. associated with
a monochromatic plane EM wave (or an AC current, using Ohms law), with angular frequency
= 2 f > 0 :
E ( r , t ) = Eo ( r ) e

it

and:

dJ free ( r , t )
dt

ne e 2
+ J free ( r , t ) = E ( r , t )
me

We assume that the {in general, complex} free current density solution to this linear
inhomogeneous first-order differential equation is of the general form: J free ( r , t ) = J o free ( r ) e it .
Then the solution to this linear, inhomogeneous first-order differential equation is:
ne e 2
ne e 2
ne e 2
,

=
=
=
E
r
t
E
r
t
E
r
t
since:
(
)
(
)
(
)

C
C

me
me
me

( i + ) J free ( r , t ) =

C
C

C
J free ( r , t ) =
E ( r , t ) =
E ( r , t ) =
E ( r , t ) C ( ) E ( r , t )

i
i
1
i
/
(
)
(
)
(
)

where:
2

( ne e me )
C

n.b. complex quantity

the AC conductivity: C ( )
=

i
i
(
)
(
)

m
1
= C (1 i / ) = e 2 (1 i / ) complex quantity
and the AC resistivity: C ( )
C ( )
ne e

Thus, we see that the {complex} AC resistance (aka impedance, Z) e.g. of a long wire is:
Z ( ) RAC ( ) =

C ( )
A

= C (1 i / ) = e 2
(1 i / ) n.b. complex quantity
A
ne e A

The AC continuity equation is: i J free ( r , t ) =

free ( r , t )
t

with: i E ( r , t ) =

free ( r , t ) Gauss
Law

( ne e 2 me )

C
E (r,t )
J free ( r , t ) = C ( ) E ( r , t ) =
E ( r , t ) =
(1 i / )
(1 i / )

But:

( ne e 2 me )

C
i E ( r , t )
Thus: i J free ( r , t ) = C ( ) i E ( r , t ) =
i E ( r , t ) =

1
/
1
/
i

(
)
(
)

free ( r , t )

free ( r , t )
free ( r , t )
C
C
=
=
free ( r , t )
or:

1
i

t
1

(
)
(
)

Lect. Notes 8

Prof. Steven Errede

The solution to the above differential equation for the electric charge volume density is of the
form: free ( r , t ) = o ( r ) e it provided that:
free ( r , t )
t

C
C
free ( r , t ) i =
n.b. implies ( ) {complex!}
(1 i / )
(1 i / )

( ne e me )
C
C
1

i.e. that: ( ) =
but: C ( )
=
i (1 i / )
(1 i / ) (1 i / )

( ) =

C ( ) 1
C
=

i
i (1 i / )

a connection between complex AC electric

permittivity () and complex AC
conductivity C() in a metal / conductor.

Technically-speaking, this relation should be: ( ) = o +

empty
space

C ( )
i
metal

because the vacuum pervades all space however, the conducting medium and the vacuum
coexist in the same region of space-time, invoking the superposition principle , the two electric
permittivities (n.b. both scalar quantities) are additive.
However, it can be seen that because the {DC} conductivity of metals is so high
{typically C ~ 107 Siemens/m }, that even for {angular} frequencies in the optical region
{ ~ 1016 radians/sec } the electric permittivity of free space, o is dwarfed by the second term,
i.e. o = 8.85 1012

C ( ) i = C i (1 i / )

neglected/dropped in many textbooks... For technical correctness / completeness sake, we will

include it here. Thus, we see that even for metal conductors:

( )
( ) o
= 1+ C
= 1 + e ( ) .
i
o
For pure copper metal at low frequencies, i.e. 0 , the static conductivity is CCu 6 107
Siemens/m; the number density of free electrons in pure copper metal is neCu
using C = ( ne e2 me ) Cu = ( neCu e2 me CCu )

(n

Cu 2
e

corresponds to a relaxation time for free / conduction electrons in pure copper of

Cu
rlx
1 Cu 2.5 1014 sec .

Lect. Notes 8

Note that in the static limit ( 0):

E ( r , t ) = Eo ( r ) e it Eo ( r ) static / constant

J = CE

J ( r , t ) = J o ( r ) e it J o ( r ) static / constant

i J =

( r , t ) = o ( r ) e it o ( r ) static / constant

C
C = 1
C
i

1
/

(
)

C ( ) =
C ( ) =

C ( )

( ) = o +

ne e 2

me

static / constant C =

= C (1 i / ) C static / constant

C ( )
C
C
= o +
= o +
singular !!!
( / + i )
i
i (1 i / )

In the previous P436 Lecture Notes 7.5 on dispersion in non-conducting, non-magnetic linear
/ homogeneous / isotropic media, we showed / derived the complex electric permittivity (due to
bound atomic electrons) of the dielectric medium to be:
n e2 n

f josc
( )
e

=
= 1 + e ( )
( ) = o (1 + e ( ) ) = o 1 +
with:
(
)

e
o me j =1 12j 2 i j
o

and: n ( ) = ( ) o = K e ( ) = 1 + e ( ) .
If only one / a single resonance exists, then this relation reduces to:

ne e 2
1

2
2

i
o e 1

1 ( ) = o 1 +

with:

ne e2

3 o me

1 j 02
j

ne e2
1
2
2
me 1 i

1 ( ) = o +

0
j

kej
me

Comparing this expression to that for conducting media with free / conduction electrons:
n e2
C
free ( ) = o +
with static conductivity: C e
( / + i )
me

ne e2 1
ne e2
1
1

=
+
2
o

me [ / + i ]
me + i

free ( ) = o +

or:

free ( ) = o

ne e2
1
with 0 0 i.e. a resonance at zero frequency!

2
2
me 0 i

Lect. Notes 8

Prof. Steven Errede

Note that the sign in the above formula {relative to that for bound atomic electrons} arises because
of fact that free / conduction electrons in metal are not bound (i.e. phase shift of 180o @ f = 0 Hz).
Then for dispersive conducting media containing free electrons, Maxwells equations become:
J free = C E =

1) i D = free i E = free

5) Ohms Law:

2) i B = 0

6) Continuity Equation:

3) E =

i J free =

C
1 i
free
t

B
t

D
E
= o J free + o
4) B = o J free + o

t
t

Assume dispersive conducting

medium is non-magnetic

( o )

Again, we use the curl operator on equations 3) and 4) to obtain wave equation(s) for EM waves
propagating in a dispersive metal / conductor, e.g. for equation 3):

E =
i E

2 E
B = o free o 2
t
t
t

2
2 E = E E

o C
o
t
t 2

and:

n2 1
=
c2 v2

conductors

( o )

Thus, for times t > 1012 sec = 1 ps we obtain: 2 E

C
1 2 E
E
E
= o C
= o
2
2
t
v t
(1 i ) t

Using equation 4) above, an identical wave equation is obtained for the magnetic field B ( r , t ) :
2 B

C
1 2B
B
B
= o C
= o
2
2
t
v t
(1 i ) t

Monochromatic plane wave solutions to above wave equation(s) for EM waves propagating in
dispersive conducting media with free electrons will (again) be of the form
i ( kz t )
e.g. E ( r , t ) = Eo e
for monochromatic plane EM waves propagating in the + z direction,
again with B ( r , t ) =

k
i ( kz t )
.
k E ( r , t ) = Bo e

Note that here again the wavenumber k is complex and frequency-dependent:

k ( ) k ( ) + i ( ) with k ( ) = e k ( ) = 2 ( ) and ( ) = m k ( ) {related to

absorption/dissipation of the EM wave as it propagates in the dispersive conductive medium.}

Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

Lect. Notes 8

We can (again) associate complex wavenumber k ( ) with complex index of refraction n ( ) :

c
k ( ) = n ( ) or: n ( ) = k ( ) where the complex index of refraction:
c

c
c
n ( ) n ( ) + i ( ) thus: n ( ) = e ( n ( ) ) = k ( ) and: ( ) = m ( n ( ) ) = ( ) .

i ( kz t )
Plugging in the expression E ( r , t ) = Eo e
for monochromatic plane EM waves propagating
in a dispersive, non-magnetic conducting medium into the above wave equation, we obtain the
so-called characteristic equation for this situation {Exercise - explicitly work this out yourselves!}:
2

n e2
n e2
io C
1

k ( ) = +
with: c 2 =
and: C = e or: = e
o o
c (1 i )
me
me C
2

i C o
c
Then: n 2 ( ) = k 2 ( ) = (Complex Index of Refraction)2 n 2 ( ) = 1 +
(1 i )

n e2
1 where: = e
me C

and:

( C o ) .

In this regime, the characteristic/mean/average time between successive collisions experienced

by the free / conduction electrons in the dispersive, non-magnetic conducting medium coll is
short compared to the oscillation period of EM waves, = 1 f = 2 , i.e. coll

io C
io C

= io C
For 0 (but not 0 !): k ( ) = +
k 2 ( )
1
c (1 i )

keeping only
terms linear in

Now note that:

1+ i
i = 1 =
{Please work this out/derive this yourselves!}
2

Thus for 0 ( i.e.

1) : k 2 ( ) io C k ( ) = k ( ) + i ( ) io C =

o C
2

(1 + i )

From which we see immediately that for monochromatic plane EM waves propagating in a
dispersive, non-magnetic conducting medium, that k ( ) ( ) =

o C

and thus here

2

{again} we see that the B -field lags the E -field by: k = B E = tan 1 tan 1 (1) = = 45o .
4
k

Lect. Notes 8

Prof. Steven Errede

c
1) since: n ( ) = k 2 ( )

Similarly, for 0 ( i.e.

2
2

i
1
c 2
c
2
= + C . Using:

n
k
i
Then: ( )
( )
o C
= i o C
2
o

o o

n ( ) = n ( ) + i ( ) =

i C

1+ i
i =

C
C
.
(1 + i ) which again implies that n ( ) ( ) =
2 o
2 o

Here again, we can define a low-frequency skin depth: sc ( )

( )

o C

same as before,
but with o

which is the characteristic distance scale that the E and B fields fall to 1 e = e 1 = 0.3679 of
their original {z = 0} values.

( ) 2 ( ) = 2 sc ( )

and corresponding low-frequency extinction coefficient: ( ) 2 ( )

2 o C

C
2 o

The characteristic distance scales over which the {time-averaged} EM wave energy density
u ( z ) and intensity I ( z ) fall to 1 e = e 1 = 0.3679 of their initial (z = 0) values, respectively are:
uEM ( z ) = uEM ( 0 ) e

I ( z ) = I ( 0) e

and

z
c

6 107 Siemens /m and thus

CCu o

6.5 1018 / sec , then for low frequencies, e.g. f = 2 1012 Hz we see that pure
copper metal is in the low-frequency regime for vacuum wavelengths of:
c 3 108
o =
f
1012

2
= 2
3 10 m = 300 m cf w/ Cu ( f ) =
k( f )
4

o ( 2 f ) CCu
2

106 m = 1 m

In pure copper metal, the skin depth sc ( ) associated with monochromatic EM waves of
frequency f ~ 1012 Hz = 1 THz = 106 MHz is:

sccu ( f = 1 THz , o = 300 m )

o C

2
4 10 2 1012 6 107
7

6.5 108 m = 65 nm

{ (10

12

sc

} { (10

Hz ) = 65 nm

12

Cu

Hz ) = 1 m

{o = 300 m}

Lect. Notes 8

Prof. Steven Errede

In pure copper metal, the skin depth sc ( ) associated with monochromatic EM waves of
frequency f ~ 1010 Hz = 10 GHz = 104 MHz (e.g. cell phones) {which corresponds to a vacuum
wavelength of o = c f = 3 108 1010 = 3 102 m 3 cm } is:

sccu (1010 Hz )

( )

o C

2
4 10 2 1010 6 107
7

6.5 107 m = 650 nm

The wavelength of EM waves with frequency f = 1010 Hz propagating in the pure copper metal is:

Cu ( f ) =

2
= 2
k( f )

o ( 2 f ) CCu
2

= 105 m = 10 m

{ (10

10

sc

} { (10

Hz ) = 650 nm

10

Cu

Hz ) = 10 m

{o = 3.0 cm}

ne e 2
1 where: =
me C

and:

The square of the complex index of refraction for a metal is: n 2 ( ) = 1 +

( C o ) .

i C o
(1 i )

At high frequencies, the free / conduction electrons in a metal will undergo a great many
oscillations before scattering i.e. the period of oscillations = 1 f = 2 is short compared to
the mean time between scatterings, essentially the relaxation time, relax = 1 .
In the high-frequency regime, the volume charge densities of free electrons and positive
ions in metal together can be thought of as a plasma whose charge density oscillates
longitudinally { i.e. in the direction of propagation of the EM wave) at the {natural} angular
frequency of known as the plasma frequency, defined as:
The plasma frequency in a metal/conductor: P
For pure copper: PCu = 1.644 1016 rad / sec

ne e 2

2
o me

f PCu =

PCu
2.616 1015 Hz
2

c
f

Cu
P

= 1.147 107 m = 114.7 nm

Operative in the optical UV portion of the EM spectrum typical for many metals !!!

10

Lect. Notes 8

Prof. Steven Errede

We can rewrite the (complex index of refraction)2 in terms of the plasma frequency :
n 2 ( ) = 1 +

used:
i ( C o )
n e2
ne e 2
1
1
1
= 1+ i e
=

(1 i )
o me (1 i )
o me i (1 i ) i 1
i = 1 i
= P2

Thus: n 2 ( ) = 1

P2

+ i )

where: P

ne e 2
ne e 2

and:

o me
me C

o ne e 2
P2
=
=

C o me ( C o )

Now: n 2 = ( n + i )i( n + i ) = n 2 + 2in + 2 with Re ( n 2 ) = n 2 2 and Im ( n 2 ) = 2n

Equating the real and imaginary portions of n 2 = ( n 2 2 ) + i ( 2n ) use the following trick:
2
2
2
1
1
2 i
2 i ( i ) (1 i ) 1 i
= 2
=
i
=
= 2 2
= 2 2
2
2 + i ( 2 + i ) ( 2 i ) 4 + 2 2
( + 2 )
( + 2 ) +
2
P2 (1 i )
P2 ( ) P
= 1
= 1 2
+i 2
= ( n 2 2 ) + i ( 2n )
n ( ) = 1 2
2
2
2
2
+ i
+
+ +

P2

= n 2 2

n 2 ( ) 2 ( ) = 1
Now define: x ( )

P2
2

+ 2)

P2
2

+ 2)

= 2 n

( ) P2
=
2
n

and
( ) ( ) 2 2
+
and x ( ) ( ) .

Then: n 2 ( ) 2 ( ) = 1 x ( ) and: 2n ( ) ( ) = ( ) x ( ) = x ( ) x ( )
Thus: =

x x
2n

n2 2 = 1 x n2

x2 x2
4n 2

= (1 x ) or: n 4 (1 x ) n 2 x x = 0
2

Let x n , then: x (1 x ) x x x = 0 which is a quadratic equation of the form:
2
2

ax 2 + bx + c = 0 with: a = 1, b = (1 x ) and c = x x .
2
must select +ve root on physical grounds!
b b 2 4ac (1 x )
=
The solution is: x =
2a

2
(1 x ) + x2 x2

11

Lect. Notes 8

n ( )
2

1 ( ) )
(
=
1+

( ) x ( )
1 + x

1 x ( )

(1 ( ) ) 1 +

( ) x ( )
1 + x

1 x ( )

Where: n ( ) =

And:

( ) =

x ( ) x ( )
2n ( )

where: P

Typical plasma frequencies for metals are P

ne e 2
o me

x ( )

P2
2

+ 2)

n e2
and: e
me C

P2
=

( C o )

Typical -factors for metals are = 1 relax 1013 1014 rad / sec .
Hence, in a typical metal/conductor we see that: P = P relax ~ 102 -103
Thus, at high frequencies (

P
Then: x ( ) 2
2
( + )

1 and: x ( )

And:

And:

( ) =

1 x ( )

1

1 P 1 P 1 {Using 1 2 1 12 2 for
2

x ( ) x ( )
2n ( )

i.e. the imaginary part ( )

( )

1}

P
2

1

1 P 2 2
2 1
2

1 at high frequencies.

P the complex index of refraction n ( ) is predominantly real,

real part n ( ) at high frequencies.

Metals are transparent to -rays and x-rays for

12

1 at high frequencies.

P
P
2

x x
P

2
2
1x

P
P
1
1

Then: n ( )

):

P2

1 i.e. P

P !!!

Lect. Notes 8

Prof. Steven Errede

Some {alkali} metals are transparent / begin to be transparent in the UV region of the EM spectrum!

P =

Metal
7
3

Alkali Metals

Li

c 2 c
=
f P P

ne e 2
o me

155.0 nm

23
11

Na

209.0 nm

39
19

287.0 nm

63
29

Cu

114.7 nm

85
37

Rb

322.0 nm UV

Above the plasma frequency ( > P ) : k ( ) = k ( ) + i ( ) and ( )

k ( ) ,

with: k ( ) = n ( ) and thus: k ( ) = n ( ) and: ( ) = ( ) .
c
c
c
Thus, since ( )

k ( ) in the high-frequency (

P ) regime, we see that:

( )
0 for
k ( )

k = B = tan 1

and thus we see that B is {nearly} in-phase with E for EM waves propagating in
metals/conductors with frequency above the plasma frequency ( > P ) .
Above the plasma frequency, we also have

Fe ( r , t ) = me a ( r , t ) = me
Because E ( r , t ) = Eo e

i kz t

, such that:

dv ( r , t )
+ me v ( r , t ) = eE ( r , t ) me a ( r , t ) = eE ( r , t )
dt

, and a ( r , t ) = r ( t ) = 2 r ( t ) t 2 we see that the free electrons

in a metal oscillate in-phase with the driving E of the monochromatic EM wave for > P .
Professor Steven Errede, Department of Physics, University of Illinois at Urbana-Champaign, Illinois

13

Lect. Notes 8

Prof. Steven Errede

Above the plasma frequency P , a metals complex index of refraction is real (i.e. absorption is
small), and the conductor becomes increasingly transparent as is increased beyond P .
Note also that n ( ) < 1 for > P {i.e. in the anomalous dispersion region}!!!
2

n ( ) 1 P for ( > P )

( ) P for ( > P )

and:

Since n 2 ( ) = ( ) o for non-magnetic conductors, ( ) is {also} predominantly real

for non-magnetic conductors at high frequencies ( > P ) , i.e. ( ) o = K e ( ) 1 for > p ;
c 1
=
the skin depth for > p is: sc ( ) =
( ) ( )
1

Below the plasma frequency P , the index of refraction is complex, i.e. ( ) is large and
therefore non-negligible; Metals are thus opaque for < P .

: k ( ) ( )

( )
and k = tan 1

k ( )

tan 1 (1)

= 45o B lags E by 45o for

The free / conduction electrons in a dispersive conductor also lag E -field by same phase lag.
For < P , since the incident EM wave is not transmitted, much of the EM wave is reflected.

In certain situations, such as EM waves propagating in the Earths ionosphere, or e.g. in a

tenuous electronic plasma in the laboratory, the electrons are {truly} free, hence damping is
negligible ( 0 ) in such situations.
Then for situations with negligible damping, i.e. 0 , the above relations simply somewhat:
2
( )
ne e 2
P
2
2
n ( ) =
= K e ( ) = (1 + e ( ) ) e ( n ( ) ) n ( ) 1

with:
P

o
o me

2

The above relation holds {i.e. is valid} over a wide range of frequencies, including < .
For situations with 0 , the imaginary part of complex n ( ) ,

( ) = m ( n ( ) )

0 because 0 .

14

1 P
c

Lect. Notes 8

which is sometimes expressed as: 2 = P2 + ck 2 dispersion relation for k()

n.b. In dispersive conducting media, is not small, so the above formulae apply only for

1 P for < P implies that k is purely imaginary!!!
In a tenuous plasma: k =
c

EM waves with < P incident on such a plasma are reflected for < P , because the EM
fields inside tenuous plasma are exponentially attenuated by the factor e kz .
It is precisely this fact that enables e.g. short-wave radio communication around the globe
the short-wave radio waves reflect off of the plasma in the earths ionosphere! {see handout}
For tenuous electron plasmas in the laboratory:

ne

P =

ne e 2
o me

( )

P c at = 0.

Skin depth in a metal (at = 0):

sc ( = 0 ) =

1
c
=
0.5 cm 5 103 cm for static (or low-frequency) EM fields
( = 0 ) P

Expulsion of EM fields within a plasma is well-known phenomena, e.g. in controlled thermonuclear processes, and can also be exploited e.g. in attempts at confinement of a hot plasma.

Electromagnetic Wave Propagation in a Linear, Homogeneous,

Isotropic Dispersive Conductive Medium

In a conducting medium, metals in particular, the optical / EM properties of many metals are
dominated by the free electrons in the metal, resulting in high reflectance R ~ 85-95%, nearly
independent of the frequency / wavelength of the incident light, at least in the visible light region
of the EM spectrum (~ 350 nm < < 780 nm). These free electrons in such metals as aluminum,
osmium, and even iron have a silvery-gray appearance {due to visible light photons scattering off
of the free electrons in the metal} essentially these metals are colorless because of their
wavelength-independent reflectance, but remember / realize that the physical color of an object
illuminated e.g. by white light (= flat distribution in frequency) is reflected light hence the
illuminated object is a poor absorber of light at that wavelength indeed, the object preferentially
absorbs light at other frequencies / wavelengths, and not so much at the frequency / wavelength
associated with the color we perceive it to be, from reflected light see figure below:

15

Lect. Notes 8

Prof. Steven Errede

Some conducting metals, such as gold and copper, as well as various alloys e.g. bronze and
brass (zinc, tin and copper) do not have a silvery-gray appearance, but have a yellow / orange /
red tinge to them. This is because these metals preferentially absorb (i.e. transmit) in the
green-blue region and thus reflect more-so in the yellow / orange / red region of the visible light
spectrum, giving these metals their characteristic hues / colors.
Indeed, e.g. gold-metalized sun glasses having a very thin layer of gold deposited on them
have transmittance T(f) which is peaked in the green portion of the visible spectrum. Goggles
coated with a thin gold layer are used e.g. by people who work around high-temperature furnaces
because while visible light from the furnace (which is peaked in the green) is transmitted,
infrared light (i.e. heat) is strongly absorbed by gold!
Because free / conduction electrons in a metal have no resonances (except the one at = 0)
because there are no restoring forces acting on them. In such metals as gold and copper, bound
electrons e.g. in the so-called valence bands of the metal also play a non-negligible role e.g. in
the optical properties of the metal i.e. in the visible light portion of the EM spectrum!

Since (i.e. we assume) the response of the free electrons in the conduction band is separable
/ independent of the response of the bound electrons in the valence band(s) to incident EM waves
(valid as long as the amplitudes of E / B are not too large). Thus we {again} use the principle of
linear superposition e.g. for the complex electric permittivity: tot ( ) = bound ( ) + free ( )
bound

osc
b 2
n
n
bound
f

n
e
j

with:
osc
e
f
=1
where: bound ( ) = o 1 +

j

2
2

j =1
o e j =1 1 j i j

3
e
n.b. nb = # density (= #/m ) of bound atomic electrons in the dispersive conducting medium

16

Lect. Notes 8

n f e2
P2
P2
1

and where: free ( ) = o e 2

o
o
2
2
( 2 + i 0 )
o me 0 i 0
+ + i 0
where: 0 = 0 and nef = # density (#/m3) of free conduction-band electrons in the dispersive
conducting medium, and the plasma frequency of the dispersive conducting medium is:

nef e 2
P
o me
bound

osc
b 2
n
f
ne e
P2
j

= 1 + tot
2
Then: tot ( ) = o 1 +
e ( ))

2
2
o(
o me j =1 1 j i j + i 0

tot ( ) = o (1 + etot ( ) )

= etot ( )

refraction n ( ) = n ( ) + i ( ) is related to the complex permittivity ( ) = ( ) + i ( ) by:

bound

osc
b 2
n
f

(
)
ne e
P2
j
tot ( )
2

2
= (1 + etot ( ) )
n ( ) =
=
= 1 +

2
2

o o o me j =1 1 j i j + i 0

with: n 2 ( ) = ( n + i )( n + i ) = ( n 2 2 ) + i ( 2n )
Complex n ( ) = n ( ) + i ( ) is again related to complex k ( ) = k ( ) + i ( ) by:

c
c
c
k ( ) = n ( ) or: n ( ) = k ( ) with: n ( ) = k ( ) and: ( ) = ( )
c