Essential Graduate Physics EM: Classical Electrodynamics

Chapter 7. Electromagnetic Wave Propagation

This (rather extensive) chapter focuses on the most important effect that follows from the time-
dependent Maxwell equations, namely the electromagnetic waves, at this stage avoiding the issue of
their origin, i.e. of the wave radiation process – which will be the subject of Chapters 8 and 10. We will
start from the simplest, plane waves in uniform and isotropic media, and then proceed to a discussion of
nonuniform systems, bringing up such effects as reflection and refraction. Then we will discuss the so-
called guided waves, propagating along various transmission lines – such as cables, waveguides, and
optical fibers. Finally, the end of the chapter is devoted to final-length fragments of such lines, serving
as resonant cavities, and to the effects of energy dissipation in transmission lines and cavities.

7.1. Plane waves

Let us start by considering a spatial region that does not contain field sources ( = 0, j = 0), and
is filled with a uniform, isotropic, linear medium, which therefore obeys Eqs. (3.46) and (5.110):

D  E, B  H . (7.1)
Moreover, let us assume for a while that these constitutive equations hold for all frequencies of interest.
(Of course, these relations are exactly valid for the very important particular case of free space, where
we may formally use the macroscopic Maxwell equations (6.100), but with  = 0 and  = 0.) As was
already shown in Sec. 6.8, in this case, the Lorenz gauge condition (6.117) allows the Maxwell
equations to be recast into the wave equations (6.118) for the scalar and vector potentials. However, for
most purposes, it is more convenient to use the homogeneous Maxwell equations (6.100) for the electric
and magnetic fields – which are independent of the gauge choice. After an elementary elimination of D
and B using Eqs. (1),1 these equations take a simple, very symmetric form:
Maxwell H E
equations E   0, H   0, (7.2a)
for uniform t t
linear
media   E  0,   H  0. (7.2b)
Now, let us act by the operator  on each of Eqs. (2a), i.e. take their curl, and then use the vector
algebra identity (5.31). The appearing terms E and H vanish due to Eqs. (2b), so the first terms of
Eqs. (2a) turn into the Laplace operators of these vectors (with the minus sign). Now swapping, in the
second terms, the operators /t and , and using Eqs. (2a) again, we get fully similar wave equations
for the electric and magnetic fields:2
EM wave  2 1 2   2 1 2 
equations    2 2 E  0,    2 2 H  0, (7.3)
 v t   v t 

1 Though in a medium, B rather than H is the actual macroscopic magnetic field, mathematically it is a bit more
convenient (just as it was in Sec. 6.2) to use the vector pair {E, H} in the following discussion, because at sharp
media boundaries, it is H that obeys the boundary condition (5.117) similar to that for E – cf. Eq. (3.37).
2 The two vector equations (3) are of course just a shorthand for six similar equations for three Cartesian
components of E and H.

© K. Likharev
Essential Graduate Physics EM: Classical Electrodynamics

where the parameter v is defined as

1
v2  . (7.4) Wave
 velocity

with v2 = 1/00  c2 in free space – see Eq. (6.120) again.

These equations allow, in particular, solutions of the following type;
Plane
E  H  f ( z  vt ), (7.5) wave

where z is the Cartesian coordinate along a certain (arbitrary) direction n, and f is an arbitrary function
of one argument. Note that this solution, first of all, describes a traveling wave – meaning a certain field
pattern moving, without deformation, along the z-axis, with the constant velocity v. Second, according
to Eq. (5), both E and H have the same values at all points of each plane perpendicular to the direction n
 nz of the wave propagation; hence the second name – plane wave.
According to Eqs. (2), the independence of the wave equations (3) for vectors E and H does not
mean that their plane-wave solutions are independent. Indeed, plugging any solution of the type (5) into
Eqs. (2a), we get
nE Field vector
H , i.e. E  Z H  n , (7.6) relation
Z
where
1/ 2
E  Wave
Z   . (7.7) impedance
H  
The vector relationship (6) means, first of all, that at any point of space and at any time instant,
the vectors E and H are perpendicular not only to the propagation vector n (such waves are called
transverse) but also to each other –– see Fig. 1.

0 Fig. 7.1. Field vectors in a plane electromagnetic

H wave propagating along direction n.
E

Second, this equality does not depend on the function f, meaning that the electric and magnetic
fields increase and decrease simultaneously. Finally, the field magnitudes are related by the constant Z
called the wave impedance of the medium. Very soon we will see that this impedance plays a pivotal
role in many problems, in particular at the wave reflection from the interface between two media. Since
the dimensionality of E, in SI units, is V/m, and that of H is A/m, Eq. (7) shows that Z has the
dimensionality of V/A, i.e. ohms ().3 In particular, in free space,

3 In the Gaussian units, E and H have a similar dimensionality (in particular, in a free-space wave, E = H), making
the (very useful) notion of the wave impedance less manifestly exposed – so in some older physics textbooks it is
not mentioned at all!

Chapter 7 Page 2 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

Wave 1/ 2
impedance  
of free Z  Z 0   0   4 10 7 c  377 Ω . (7.8)
space  0 
Next, plugging Eq. (6) into Eqs. (6.113) and (6.114), we get:
Wave’s
energy u  E 2  H 2 , (7.9a)

Wave’s E2
power S  E H  n  nZH 2 , (7.9b)
Z
so, according to Eqs. (4) and (7), the wave’s energy and power densities are universally related as
S  nuv . (7.9c)
In view of the Poynting vector paradox discussed in Sec. 6.8 (see Fig. 6.11), one may wonder
whether the last equality may be interpreted as the actual density of power flow. In contrast to the static
situation shown in Fig. 6.11, which limits the electric and magnetic fields to the vicinity of their sources,
waves may travel far from them. As a result, they can form wave packets of a finite length in free space
– see Fig. 2.

S0 wave packet

Fig. 7.2. Interpreting the Poynting

vector in a plane electromagnetic
wave. (Horizontal lines show
V equal-field planes.)
S0

Let us apply the Poynting theorem (6.111) to the cylinder shown with dashed lines in Fig. 2, with
one lid inside the wave packet, and another lid in the region already passed by the wave. Then,
according to Eq. (6.111), the rate of change of the full field energy E inside the volume is dE /dt = –SA
(where A is the lid area), so S may be indeed interpreted as the power flow (per unit area) from the
volume. Making a reasonable assumption that the finite length of a sufficiently long wave packet does
not affect the physics inside it, we may indeed interpret the S given by Eqs. (9b-c) as the power flow
density inside a plane electromagnetic wave.
As we will see later in this chapter, the free-space value Z0 of the wave impedance, given by Eq.
(8), establishes the scale of Z of virtually all wave transmission lines, so we may use it, together with
Eq. (9), to get a better feeling of how much different are the electric and magnetic field amplitudes in the
waves – on the scale of typical electrostatics and magnetostatics experiments. For example, according to
Eqs. (9), a wave of a modest intensity S = 1 W/m2 (this is what we get from a usual electric bulb a few
meters away from it) has E ~ (SZ0)1/2 ~ 20 V/m, quite comparable with the dc field created by a standard
AA battery right outside it. On the other hand, the wave’s magnetic field H = (S/Z0)1/2  0.05 A/m. For
this particular case, the relation following from Eqs. (1), (4), and (7),

Chapter 7 Page 3 of 70
Essential Graduate Physics EM: Classical Electrodynamics

E E E
B  H       E  ,
1/ 2
(7.10)
Z  /  1/ 2
v
gives B = 0H = E/c ~ 710-8T, i.e. a magnetic field a thousand times lower than the Earth’s field, and
about 7 orders of magnitude lower than the field of a typical permanent magnet. This huge difference
may be interpreted as follows: the scale B ~ E/c of magnetic fields in the waves is “normal” for
electromagnetism, while the permanent magnet fields are abnormally high because they are due to the
ferromagnetic alignment of electron spins, essentially relativistic objects – see the discussion in Sec. 5.5.
The fact that Eq. (5) is valid for an arbitrary function f means, in the standard terminology, that a
medium with frequency-independent  and  supports the propagation of plane waves without either
decay (attenuation) or waveform deformation (dispersion). However, for any real medium but pure
vacuum, this approximation is valid only within limited frequency intervals. We will discuss the effects
of attenuation and dispersion in the next section and will see that all our prior formulas remain valid
even for an arbitrary linear media, provided that we limit them to single-frequency (i.e. sinusoidal,
frequently called monochromatic) waves. Such waves may be most conveniently represented as4
 
f  Re  f  e i kz  t , (7.11)
Mono-
chromatic
  wave

where f is the complex amplitude of the wave, and k is its wave number (the magnitude of the wave
vector k  nk), sometimes called the spatial frequency. The last term is justified by the fact, evident
from Eq. (11), that k is related to the wavelength  exactly as the usual (“temporal”) frequency  is
related to the time period T:
2 2 Spatial and
k ,  . (7.12) temporal
 T frequencies

In the dispersion-free case (5), the compatibility of that relation with Eq. (11) requires the argument (kz
– t)  k[z – (/k)t] to be proportional to (z – vt), so /k = v, i.e.

    ,
1/ 2 Dispersion
k (7.13) relation
v
so in that particular case, the dispersion relation (k) is linear.
Now note that Eq. (6) does not mean that the vectors E and H retain their direction in space.
(The wave in which they do, is called linearly polarized.5) Indeed, nothing in the Maxwell equations
prevents, for example, a joint rotation of this vector pair around the fixed vector n, while still keeping all
these three vectors perpendicular to each other at any instant – see Fig. 1. However, an arbitrary rotation
law or even an arbitrary constant frequency of such rotation would violate the single-frequency
(monochromatic) character of the elementary sinusoidal wave (11). To understand what is the most
general type of polarization the wave may have without violating that condition, let us represent two

4 As we have already seen in the previous chapter (see also CM Sec. 5.1), such complex-exponential
representation of sinusoidally changing variables is more convenient for mathematical manipulation than by using
sine and cosine functions, especially because in all linear relations, the operator Re may be omitted (implied) until
the very end of the calculation. Note, however, that this is not valid for the quadratic forms such as Eqs. (9).
5 The possibility of different polarizations of electromagnetic waves was discovered (for light) in 1699 by Rasmus
Bartholin, a.k.a. Erasmus Bartholinus.

Chapter 7 Page 4 of 70
Essential Graduate Physics EM: Classical Electrodynamics

Cartesian components of one of these vectors (say, E) along any two fixed axes x and y, perpendicular to
each other and the z-axis (i.e. to the vector n), in the same form as used in Eq. (11):
 
E x  Re  E x e i kz  t  ,
 
E y  Re  E y e i kz  t  . (7.14)
   
To keep the wave monochromatic, the complex amplitudes E x and E y have to be constant in time;
however, they may have different magnitudes and an arbitrary phase shift between them.
In the simplest case when the arguments of these complex amplitudes are equal,
E  x , y  E  x , y e i . (7.15)

the real field components have the same phase:

E x  E x cos(kz  t   ), E y  E y cos(kz  t   ), (7.16)

so their ratio is constant in time – see Fig. 3a. This means that the wave is linearly polarized, with the
polarization plane defined by the relation
tan   Ey / Ex . (7.17)

Ey (a) Ey (b) Ey (c)

E y E
E y
  (t )  (t )
0 E x Ex 0 E Ex 0 E x Ex

Fig. 7.3. Time evolution of the instantaneous electric field vector in monochromatic waves with:
(a) a linear polarization, (b) a circular polarization, and (c) an elliptical polarization.

Another simple case is when the moduli of the complex amplitudes E x and E y are equal, but
their phases are shifted by +/2 or –/2:
i    / 2 
E  x  E  e i , E  y  E  e . (7.18)
In this case
 
E x  E coskz  t   , E y  E cos kz  t       E sin kz  t    . (7.19)
 2
This means that on the wave’s plane (normal to n), the end of the vector E moves, with the wave’s
frequency , either clockwise or counterclockwise around a circle – see Fig. 3b:
 (t )   (t   ) . (7.20)

Chapter 7 Page 5 of 70
Essential Graduate Physics EM: Classical Electrodynamics

Such waves are called circularly polarized. In the dominant convention, the wave is called right-
polarized (RP) if it is described by the lower sign in Eqs. (18)-(20), i.e. if the vector  of the angular
frequency of the field vector’s rotation coincides with the wave propagation’s direction n, and left-
polarized (LP) in the opposite case. These particular solutions of the Maxwell equations are very
convenient for quantum electrodynamics, because single electromagnetic field quanta with a certain
(positive or negative) spin direction may be considered elementary excitations of the corresponding
circularly polarized wave.6 (This fact does not exclude, from the quantization scheme, waves of other
polarizations, because any monochromatic wave may be presented as a linear combination of two
opposite circularly polarized waves – just as Eqs. (14) represent it as a linear combination of two
linearly polarized waves.)
Finally, in the general case of arbitrary complex amplitudes E x and E y, the field vector’s end
moves along an ellipse (Fig. 3c); such a wave is called elliptically polarized. The elongation
(“eccentricity”) and orientation of the ellipse are completely described by one complex number, the ratio
Ex/Ey, i.e. by two real numbers – for example, E x/E y and  = arg(E x/E y).7

7.2. Attenuation and dispersion

Let me start the discussion of the dispersion and attenuation effects by considering a particular
case of the time evolution of the electric polarization P(t) of a dilute, non-polar medium, with negligible
interaction between its elementary dipoles p(t). As was discussed in Sec. 3.3, in this case, the local
electric field acting on each elementary dipole, may be taken equal to the macroscopic field E(t). Then,
the dipole moment p(t) may be caused not only by the values of the field E at the same moment of time
(t), but also by those at the earlier moments t’ < t. Due to the linear superposition principle, the
macroscopic polarization P(t) = np(t) should be a sum (practically, an integral) of the values of E(t’) at
all moments t’  t, weighed by some function of t and t’: 8
t Temporal
P (t )   E (t' )G(t , t' )dt' .

(7.21) Green’s
function

6 This issue is closely related to that of the wave’s angular momentum; it will be more convenient for me to
discuss it later in this chapter (in Sec. 7).
7 Note that the same information may be expressed via four so-called Stokes parameters s , s , s , and s , which
0 1 2 3
are popular in practical optics, because they may be used for the description of not only completely coherent
waves that were discussed here but also of party coherent or even fully incoherent waves – including the natural
light emitted by thermal sources such as our Sun. (In contrast to the coherent waves (14), whose complex
amplitudes are deterministic numbers, the amplitudes of incoherent waves should be treated as random variables.)
For more on the Stokes parameters, as well as many other optics topics I will not have time to cover, I can
recommend the classical text by M. Born et al., Principles of Optics, 7th ed., Cambridge U. Press, 1999.
8 In an isotropic media, the vectors E, P, and hence D =  E + P, are all parallel, and for notation simplicity, I will
0
drop the vector sign in the following formulas of this section. I am also assuming that P at any point r is only
dependent on the electric field at the same point, and hence drop the factor exp{ikz}, the same for all variables.
This last assumption is valid if the wavelength  is much larger than the elementary dipole’s size a. In most
systems of interest, the scale of a is atomic (~10-10m), so this approximation is valid up to extremely high
frequencies,  ~ c/a ~ 1018 s-1, corresponding to hard X-rays.

Chapter 7 Page 6 of 70
Essential Graduate Physics EM: Classical Electrodynamics

The condition t’  t, which is implied by this relation, expresses a keystone principle of all
science, the causal relation between a cause (in our case, the electric field E(t’) applied to each dipole)
and its effect (the polarization P(t) it creates). The function G(t, t’) is called the temporal Green’s
function for the electric polarization.9 To reveal its physical sense, let us consider the case when the
applied field E(t) is a very short pulse at the moment t0 < t, which may be well approximated with
Dirac’s delta function:
E (t )   (t  t" ) . (7.22)

Then Eq. (21) yields just P(t) = G(t, t”), so the Green’s function G(t, t’) is just the polarization at
moment t, created by a unit -functional pulse of the applied field at moment t’ (Fig. 4).

E (t ) E (t )   (t  t' )
P(t )
P(t )  G (t , t' ) Fig. 7.4. An example of the temporal
0 Green’s function for the electric
t' t polarization (schematically).

What are the general properties of the temporal Green’s function? First, for systems without
infinite internal “memory”, G should tend to zero at t – t’  , although the type of this approach (e.g.,
whether the function G oscillates approaching zero, as in Fig. 4, or not) depends on the medium’s
properties. Second, if the parameters of the medium do not change in time, the polarization response to
an electric field pulse should be dependent not on its absolute timing, but only on the time difference 
 t – t’ between the pulse and observation instants, when Eq. (21) is reduced to
t 
P (t ) 

 E (t' )G(t  t' )dt'   E (t   )G( )d .
0
(7.23)

-it
For a sinusoidal waveform, E(t) = Re [Ee ], this equation yields

i (t  )    
P (t )  Re  E e G ( )d  Re  E  G ( )e i d  e it  . (7.24)
0  0  
The expression in the last parentheses is of course nothing else than the complex amplitude P of the
polarization. This means that though even if the static linear relation (3.43), P = e0E, is invalid for an
arbitrary time-dependent process, we may still keep its Fourier analog,

1
P   e   0 E , with  e     G( )e
i
d , (7.25)
0 0

for each sinusoidal component of the process, using it as the definition of the frequency-dependent
electric susceptibility e(). Similarly, the frequency-dependent electric permittivity may be defined
using the Fourier analog of Eq. (3.46):

9The idea of these functions is very similar to that of the spatial Green’s functions (see Sec. 2.10), but with a new
twist, due to the causality principle. A discussion of the temporal Green’s functions in application to classical
mechanics (which to some extent overlaps with our current discussion) may be found in CM Sec. 5.1.

Chapter 7 Page 7 of 70
Essential Graduate Physics EM: Classical Electrodynamics

Complex
D    E . (7.26a) electric
permittivity

Then, according to Eq. (3.47), the complex permittivity is related to the temporal Green’s function by
the usual Fourier transform:

P
 ( )   0     0   G ( )e i d . (7.26b)
E 0

This relation shows that the function () may be complex,

 
 ( )  ' ( )  i" ( ), with '     0   G ( ) cos  d , " ( )   G ( ) sin  d , (7.27)
0 0

and that its real part ’() is always an even function of frequency, while the imaginary part ”() is an
odd function of . Note that though the particular causal relationship (21) between P(t) and E(t) is
conditioned by the elementary dipole independence, the frequency-dependent complex electric
permittivity () may be introduced, in a similar way, if any two linear combinations of these variables
are related by a similar formula.
Absolutely similar arguments show that magnetic properties of a linear, isotropic medium may
be characterized by a frequency-dependent, complex permeability (). Now rewriting Eqs. (1) for the
complex amplitudes of the fields at a particular frequency, we may readily repeat all calculations of Sec.
1, and verify that all its results are valid for monochromatic waves even for a dispersive (but necessarily
linear!) medium. In particular, Eqs. (7) and (13) now become
1/ 2
  ( )  Complex
Z ( )   k ( )    ( )  ( ) ,
1/ 2
 , (7.28) Z and k
  ( ) 
so the wave impedance and the wave number may be both complex functions of frequency.10
This fact has important consequences for electromagnetic wave propagation. First, plugging the
representation of the complex wave number as the sum of its real and imaginary parts, k()  k’() +
ik”(), into Eq. (11):

f  Re f  e
i [ k ( ) z t ]
 e k" ( ) z

Re f  e
i [ k' ( ) z t ]
, (7.29)

we see that k”() describes the rate of wave attenuation in the medium at frequency .11 Second, if the
waveform is not sinusoidal (and hence should be represented as a sum of several/many sinusoidal
components), the frequency dependence of k’() provides for wave dispersion, i.e. the waveform
deformation at the propagation, because the propagation velocity (4) of component waves is now
different.12

10 The first unambiguous observations of dispersion (for the case of light refraction) were described by Sir Isaac
Newton in his Optics (1704) – even though this genius has never recognized the wave nature of light!
11 It may be tempting to attribute this effect to wave absorption, i.e. the dissipation of the wave’s energy, but we
will see very soon that wave attenuation may be due to different effects as well.
12 The reader is probably familiar with the most noticeable effect of the dispersion: the difference between the

group velocity vgr  d/dk’ giving the speed of the envelope of a wave packet with a narrow frequency spectrum,
and the phase velocity vph  /k’ of the component waves. The second-order dispersion effect, proportional to
d2/d2k’, leads to the deformation (gradual broadening) of the envelope itself. Following tradition, these effects

Chapter 7 Page 8 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

As an example of such a dispersive medium, let us consider a simple but very representative
Lorentz oscillator model.13 In dilute atomic or molecular systems (e.g., gases), electrons respond to the
external electric field especially strongly when its frequency  is close to certain frequencies j
corresponding to the spectrum of quantum interstate transitions of a single atom/molecule. A
phenomenological description of this behavior may be obtained from a classical model of several
externally driven harmonic oscillators, generally with non-zero damping. For a single oscillator, driven
by the electric field’s force F(t) = qE(t), we can write the 2nd Newton law as

 
m x  2 0 x   02 x  qE (t ) , (7.30)

where 0 is the own frequency of the oscillator, and 0 is its damping coefficient. For the electric field
of a monochromatic wave, E(t) = Re [Eexp{–it}], we may look for a particular, forced-oscillation
solution of this equation in a similar form x(t) = Re [xexp{–it}].14 Plugging this solution into Eq. (30),
we readily find the complex amplitude of these oscillations:
q E
x  . (7.31)
 
m  0   2  2i 0
2

Using this result to calculate the complex amplitude of the dipole moment as p = qx, and then the
electric polarization P = np of a dilute medium with n independent oscillators for unit volume, for its
frequency-dependent permittivity (26) we get
Lorentz q2 1
 ( )   0  n . (7.32)
oscillator
model 
m  j    2i 0
2 2

This result may be readily generalized to the case when the system has several types of
oscillators with different masses and frequencies:
fj
 ( )   0  nq 2 
j 
m j  2j   2   2i j , (7.33)

where fj  nj/n is the fraction of oscillators with frequency j, so the sum of all fj equals 1. Figure 5
shows a typical behavior of the real and imaginary parts of the complex dielectric constant, described by
Eq. (33), as functions of frequency. The oscillator resonances’ effect is clearly visible, and dominates
the media response at   j, especially in the case of low damping, j << j. Note that in the low-
damping limit, the imaginary part of the dielectric constant ”, and hence the wave attenuation k”, are
negligibly small at all frequencies besides small vicinities of j, where the derivative d’()/d is

are discussed in more detail in the quantum-mechanics part of this series (QM Sec. 2.2), because they are a crucial
factor of Schrödinger’s wave mechanics. (See also a brief discussion in CM Sec. 6.3.)
13 This example is focused on the frequency dependence of  rather than , because electromagnetic waves
interact with “usual” media via their electric field much more than via the magnetic field. Indeed, according to Eq.
(7), the magnetic field of the wave is of the order of E/c, so the magnetic component of the Lorentz force (5.10),
acting on a non-relativistic particle, Fm ~ quB ~ (u/c)qE, is much smaller than that of its electric component, Fe =
qE, and may be neglected. However, as will be discussed in Sec. 6, forgetting about the possible dispersion of
() may result in missing some remarkable opportunities for manipulating the waves.
14 If this point and Eq. (30) are not absolutely clear, please see CM Sec. 5.1 for a more detailed discussion.

Chapter 7 Page 9 of 70
Essential Graduate Physics EM: Classical Electrodynamics

negative.15 Thus, for a system of weakly-damped oscillators, Eq. (33) may be well approximated by a
sum of singularities (“poles”):
q2 fj
 ( )   0  n
2
 m  
j  
, for  j     j   j   j ' . (7.34)
j j j

 ()

'
 (0)
" 0
0  Fig. 7.5. Typical frequency
dependence of the real and imaginary
parts of the complex electric
1 2 3 permittivity, according to the
generalized Lorentz oscillator model.

This result is especially important because according to quantum mechanics,16 Eq. (34) (with all
mj equal) is also valid for a set of non-interacting, similar quantum systems (whose dynamics may be
completely different from that of a harmonic oscillator!), provided that j are replaced with frequencies
of possible quantum interstate transitions, and coefficients fj are replaced with the so-called oscillator
strengths of the transitions – which obey the same sum rule, j fj = 1.
At   0, the imaginary part of the complex permittivity (33) also vanishes (for any j), while
its real part approaches its electrostatic (“dc”) value
nj
 ( 0)   0  q 2  . (7.35)
j m j j
2

Note that according to Eq. (30), the denominator of the fraction in Eq. (35) is just the effective spring
constant j = mjj2 of the jth oscillator, so the oscillator masses mj as such are actually (and quite
naturally) not involved in the static dielectric response.
In the opposite limit of very high frequencies,  >> j, j, the permittivity also becomes real and
may be represented as
  p2  q2 nj
 ( )   0 1  2
 
,

where  p2 
0
m . (7.36) () in
plasma
  j j

15 In optics, such behavior is called anomalous dispersion.

16 See, e.g., QM Chapters 5-6.

Chapter 7 Page 10 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

This result is very important because it is also valid at all frequencies if all j and j vanish, for example
for gases of free charged particles, in particular for plasmas – ionized atomic gases, provided that the
ion collision effects are negligible. (This is why the parameter p defined by Eq. (36) is called the
plasma frequency.) Typically, the plasma as a whole is neutral, i.e. the density n of positive atomic ions
is equal to that of the free electrons. Since the ratio nj/mj for electrons is much higher than that for ions,
the general formula (36) for the plasma frequency is usually well approximated by the following simple
expression:
ne 2
 p2  . (7.37)
 0 me
This expression has a simple physical sense: the effective spring constant ef  mep2 = ne2/0
describes the Coulomb force that appears when the electron subsystem of the plasma is shifted, as a
whole, from its positive-ion subsystem, thus violating the electroneutrality.17 Hence, there is no surprise
that the function () given by Eq. (36) vanishes at  = p: at this resonance frequency, the polarization
electric field E may oscillate, i.e. have a non-zero amplitude E = D/(), even in the absence of
external forces induced by external (stand-alone) charges, i.e. in the absence of the field D these charges
induce – see Eq. (3.32).
The behavior of electromagnetic waves in a medium that obeys Eq. (36), is very remarkable. If
the wave frequency  is above p, the dielectric constant () and hence the wave number (28) are
positive and real, and waves propagate without attenuation, following the dispersion relation,
Plasma
dispersion
relation
k ( )    ( )  0 
1/ 2

c

1 2
   p2 
1/ 2
, (7.38)
whose plot is shown in Fig. 6.

2
 1
p
1
  ck
Fig. 7.6. The plasma dispersion law (solid
line) in comparison with the linear dispersion
0 in the free space (dashed line).
0 1 2 3
k /( p / c )

At   p the wave number k tends to zero. Beyond that point (i.e. at  < p), we still can use
Eq. (38), but it is instrumental to rewrite it in the mathematically equivalent form

17 Indeed, let us consider such a small shift x, perpendicular to the plane surface of a broad, plane slab filled with
plasma. The uncompensated ion charges, with equal and opposite surface densities  = enx, that appear at the
slab surfaces, create inside it, according to Eq. (2.3), a uniform electric field with Ex = enx/0. This field exerts
the force –eEx = –(ne2/0)x  –efx on each electron, pulling it back to its equilibrium position.

Chapter 7 Page 11 of 70
Essential Graduate Physics EM: Classical Electrodynamics

k ( )  
i 2
p   2 
1/ 2

i
, where  
c
. (7.39)
c   2
p 2 
1/ 2

At  < p, the so-defined parameter  is real, and Eq. (29) shows that the electromagnetic field
exponentially decreases with distance:

f  Re f  e i kz t   exp  Re f  e it .

 z
(7.40)
 
Does this mean that the wave is being absorbed in the plasma? Answering this question is a good
pretext to calculate the time average of the Poynting vector S = EH of a monochromatic
electromagnetic wave in an arbitrary dispersive (but still linear and isotropic) medium. First, let us spell
out the real fields’ time dependences:

  12 E e it  c.c.,
E (t )  Re E e it   
H (t )  Re H  e it   1  E it

2  Z ( )
e

 c.c. . (7.41)

Now, a straightforward calculation yields18
2 1/ 2
E E*  1 1  E E* 1 E   ( ) 
S  E t H t    Z ( )  *   2 Re Z ( )  2 Re   ( )  . (7.42)
4  Z ( )   
Let us apply this important general formula to our simple model of plasma at  < p. In this
case, the magnetic permeability equals μ0, i.e. μ() = μ0 is positive and real, while () is real and
negative, so 1/Z() = [()/ μ()]1/2 is purely imaginary, and the average Poynting vector (42) vanishes.
This means that the energy, on average, does not flow along the z-axis. So, the waves with  < p are
not absorbed in plasma. (Indeed, the Lorentz model with j = 0 does not describe any energy dissipation
mechanism.) Instead, as we will see in the next section, the waves are rather reflected from the plasma’s
boundary, more exactly from its surface layer of a thickness ~.
Note also that in the limit  << p, Eq. (39) yields
1/ 2 1/ 2
c  c 2  0 me   me 
   
   
 . (7.43)
 p  ne 2    0 ne
2

But this is just a particular case (for q = e, m = me, and  = 0) of Eq. (6.44) that was derived in Sec. 6.4
for the depth of the magnetic field’s penetration into a lossless (collision-free) conductor in the
quasistatic approximation. This fact shows again that, as was already discussed in Sec. 6.7, this

18 For an arbitrary plane wave, the total average power flow may be calculated as an integral of Eq. (42) over all
frequencies. By the way, combining this integral and the Poynting theorem (6.111), is it straightforward to prove
the following interesting expression for the average electromagnetic energy density of a narrow ( << ) wave
packet propagating in an arbitrary dispersive (but linear and isotropic) medium:
1  d '   d '   
u  
2 packet  d
E E* 
d
H  H * d .


Chapter 7 Page 12 of 70
Essential Graduate Physics EM: Classical Electrodynamics

approximation (in which the displacement currents are neglected) gives an adequate description of the
time-dependent phenomena at  << p, i.e. at  << c/ = 1/k = /2.19
There are two most important examples of natural plasmas. For the Earth’s ionosphere, i.e. the
upper part of its atmosphere, which is almost completely ionized by the ultra-violet and X-ray
components of the Sun’s radiation, the maximum value of n, reached about 300 km over the Earth’s
surface, is between 1010 and 1012 m-3 (depending on the time of the day and the Sun’s activity phase), so
the maximum plasma frequency (37) is between 1 and 10 MHz. This is much higher than the particles’
typical reciprocal collision time -1, so Eq. (38) gives a good description of wave dispersion in this
plasma. The effect of reflection of electromagnetic waves with  < p from the ionosphere enables long-
range (over-the-globe) radio communications and broadcasting at the so-called short waves, with cyclic
frequencies of the order of 10 MHz:20 they may propagate in the flat channel formed by the Earth’s
surface and the ionosphere, being reflected repeatedly by these parallel “walls”. Unfortunately, due to
the random variations of the Sun’s activity, and hence of p, this natural radio communication channel is
not too reliable, and in our age of transworld optical-fiber cables (see Sec. 7 below), its practical
importance has diminished.
Another important example of plasmas is free electrons in metals and other conductors. For a
typical metal, n is of the order of 1023 cm-3  1029 m-3, so Eq. (37) yields p ~ 1016 s-1. This value of p
is somewhat higher than the mid-optical frequencies ( ~ 31015 s-1), explaining why planar, clean
metallic surfaces, such as the aluminum and silver films used in mirrors, are so shiny: at these
frequencies, their complex permittivity () is almost exactly real and negative, leading to light
reflection, with very little absorption.
The simple model (36), which neglects electron scattering, becomes inadequate at lower
frequencies,  ~ 1. A good phenomenological way of extending the model to the account of scattering
is to take, in Eq. (33), the lowest frequency j equal zero (to describe the free electrons), while keeping
the damping coefficient 0 of this mode larger than zero, to account for the energy dissipation due to
their scattering. Then Eq. (33) is reduced to
n0 q 2 1 n0 q 2 1
 ef ( )   opt ( )    opt ( )  i , (7.44)
m    2i 0
2
2 0 m 1  i / 2 0
where the response opt() at high (in practice, optical) frequencies is still given by Eq. (33), but now
with j > 0. The result (44) allows for a simple interpretation. To show that, let us incorporate into our
calculations the Ohmic conduction of the medium, generalizing Eq. (4.7) as j = ()E to account for
the possible frequency dependence of the Ohmic conductivity. Plugging this relation into the Fourier
image of the relevant macroscopic Maxwell equation, H = j – iD  j – i()E, we get

  H       i  E  . (7.45)

19 One more convenience of the simple model of a collision-free plasma, which has led us to Eq. (36), is that it
may be readily generalized to the case of an additional strong dc magnetic field B0 (much higher than that of the
wave) applied in the direction n of wave propagation. It is straightforward (and hence left for the reader) to show
that such plasma exhibits the Faraday effect of the polarization plane’s rotation, and hence gives an example of an
anisotropic media that violates the Lorentz reciprocity relation (6.121).
20 These frequencies are an order of magnitude lower than those used for TV and FM-radio broadcasting.

Chapter 7 Page 13 of 70
Essential Graduate Physics EM: Classical Electrodynamics

This relation shows that for a monochromatic wave, the addition of the Ohmic current density j to the
displacement current density is equivalent to the addition of () to –i(), i.e. to the following
change of the ac electric permittivity:21
  
     ef ( )   opt    i . (7.46)

Now the comparison of Eqs. (44) and (46) shows that they coincide if we take
n0 q 2 1 1 Generalized
 ( )    (0) , (7.47) Drude
m0 1  i 1  i formula

where the dc conductivity (0) is described by the Drude formula (4.13), and the phenomenologically
introduced coefficient 0 is associated with 1/2. Eq. (47), which is frequently called the generalized (or
“ac”, or “rf”) Drude formula,22 gives a very reasonable (semi-quantitative) description of the ac
conductivity of many metals almost up to optical frequencies.
Now returning to our discussion of the generalized Lorentz model (33), we see that the
frequency dependences of the real (’) and imaginary (”) parts of the complex permittivity it yields are
not quite independent. For example, let us have one more look at the resonance peaks in Fig. 5. Each
time the real part drops with frequency, d’/d < 0, its imaginary part ” has a positive peak. Ralph
Kronig (in 1926) and Hendrik (“Hans”) Kramers (in 1927) independently showed that this is not an
occasional coincidence pertinent only to this particular model. Moreover, the full knowledge of the
function ’() enables the calculation of the function ”(), and vice versa. The mathematical reason
for this fact is that both these functions are always related to a single real function G() – see Eqs. (27).
To derive these relations, let us consider Eq. (26b) on the complex frequency plane,     ’
+ i”:
 
f ω    (ω)   0   G ( )e iω
d   G ( )e i' e " d . (7.48)
0 0

For all stable physical systems, G() has to be finite for all important values of the real integration
variable ( > 0), and tend to zero at   0 and   . (Indeed, according to Eq. (23), a non-zero G(0)
would mean an instantaneous response of the medium to the external force, while G()  0 would mean
that it has an infinitely long memory.) Because of that, and thanks to the factor e–”, the expression
under the integral in Eq. (48) tends to zero at      in all upper half-plane (”  0). As a result, we
may claim that the complex function f () given by this relation, is analytical in that half-plane. This fact
allows us to apply to it the general Cauchy integral formula23
1 dΩ
f (ω)  
2i C
f (Ω)
Ωω
, (7.49)

21 Alternatively, according to Eq. (45), it is possible (and in the field of infrared spectroscopy, conventional) to
attribute the ac response of a medium, at all frequencies, to its effective complex conductivity: ef ()  () –
i()  –ief ().
22 It may be also derived from the Boltzmann kinetic equation in the so-called relaxation-time approximation
(RTA) – see, e.g., SM Sec. 6.2.
23 See, e.g., MA Eq. (15.2).

Chapter 7 Page 14 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

where   ’ + i” is also a complex variable. Let us take the integration contour C of the form shown
in Fig. 7, with the radius R of the larger semicircle tending to infinity, and the radius r of the smaller
semicircle (around the singular point  = ) tending to zero. Due to the exponential decay of f() at
  , the contribution to the right-hand side of Eq. (49) from the larger semicircle vanishes,24 while
the contribution from the small semicircle, where  =  + rexp{i}, with –    0, is
f ( ) ir exp{i}d f ( )
0 0
1 dΩ 1
lim r 0 
2 i Ω  r exp{i )
f ( Ω)
Ω 
 
2 i  r exp{i}

2   d  f ( ). (7.50)
2

Ω"
Ω R

C  Fig. 7.7. Deriving the Kramers-

Kronig dispersion relations.
0 Ω   r  0 '

As a result, for our contour C, Eq. (49) yields

  r   
1   d 1
f ( )  lim r 0      f ()
2i     r 
 f ( ) .,
  2
(7.51)
 
where   ’ on the real axis (where ” = 0). Such an integral, excluding a symmetric infinitesimal
vicinity of a pole singularity, is called the principal value of the (formally, diverging) integral from –
to +, and is denoted by the letter P before it.25 Using this notation, subtracting f()/2 from both parts of
Eq. (51), and multiplying them by 2, we get

1 dΩ
f ( )  P  f (Ω) . (7.52)
 i  Ω 
Now plugging into this complex equality the polarization-related difference f()  () – 0 in the form
[’() – 0] + i[”()], and requiring both real and imaginary components of the two sides of Eq. (52) to
be equal separately, we get the famous Kramers-Kronig dispersion relations
Kramers-  
1 dΩ 1 dΩ
Kronig ' ( )   0  P  " (Ω) , " ( )   P  ' (Ω)   0  . (7.53)
dispersion
relations
  Ω   
Ω 
We may use the already mentioned fact that ’() is always an even function, while ”() an odd
function of frequency, to rewrite these relations in the following equivalent form,

24Strictly speaking, this also requires f() to decrease faster than -1 at the real axis (at ” = 0), but due to the
inertia of charged particles, this requirement is fulfilled for all realistic models of dispersion – see, e.g., Eq. (36).
25 I am typesetting this symbol in a Roman (upright) font, to avoid any possibility of confusion with the medium’s
polarization.

Chapter 7 Page 15 of 70
Essential Graduate Physics EM: Classical Electrodynamics

 
2 Ω dΩ 2 dΩ
' ( )   0  P  " (Ω) , " ( )   P  ' (Ω)   0  , (7.54)
 0 Ω2   2  0 Ω 2
2

which is more convenient for most applications, because it involves only physical (positive) frequencies.
Though the Kramers-Kronig relations are “global” in frequency, in certain cases they allow an
approximate calculation of dispersion from experimental data for absorption, collected even within a
limited (“local”) frequency range. Most importantly, if a medium has a sharp absorption peak at some
frequency j, we may describe it as
" ( )  c (   j )  a more smooth function of  , (7.55)

and the first of Eqs. (54) immediately gives

Dispersion
2c j near an
' ( )   0   another smooth function of  , (7.56) absorption
  2j   2 line

thus predicting the anomalous dispersion near such a point. This calculation shows that such behavior
observed in the Lorentz oscillator model (see Fig. 5) is by no means occasional or model-specific.
Let me emphasize again that the Kramers-Kronig relations (53)-(54) are much more general than
the Lorentz model (33), and require only a causal linear relation (21) between the polarization P(t) with
the electric field E(t’).26 Hence, these relations are also valid for the complex functions relating Fourier
images of any cause/effect-related pair of variables. In particular, at a measurement of any linear
response r(t) of any experimental sample to any external field f(t’), whatever the nature of this response
and physics behind it, we may be confident that there is a causal relationship between the variables r and
f, so the corresponding complex function ()  r/f does obey the Kramers-Kronig relations.
However, it is still important to remember that a linear relationship between the Fourier amplitudes of
two variables does not necessarily imply a causal relationship between them.27

7.3. Reflection
The most important new effect arising in nonuniform media is wave reflection. Let us start its
discussion from the simplest case of a plane electromagnetic wave that is normally incident on a sharp
interface between two uniform, linear, isotropic media.
Moreover, let us first consider an even simpler sub-case when one of the two media (say, that
located at z > 0, see Fig. 8) cannot sustain any electric field at all – as implied, in particular, by the
macroscopic model of a perfect conductor – see Eq. (2.1):

26 Actually, in mathematics, the relations even somewhat more general than Eqs. (53) and valid for an arbitrary
analytic function of complex argument (the Sokhotski-Plemelj theorem) have been known at least from 1868.
27 For example, the function ()  E / P , in the Lorentz oscillator model, does not obey the Kramers-Kronig
 
relations. This is evident not only physically, from the fact that E(t) is not a causal function of P(t), but even
mathematically. Indeed, Green’s function describing a causal relationship has to tend to zero at small time delays
  t – t’, so its Fourier image has to tend to zero at    . This is certainly true for the function f() given by
Eq. (32), but not for the reciprocal function ()  1/f()  (2 – 02) – 2i, which diverges at large
frequencies.

Chapter 7 Page 16 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

E z 0  0. (7.57)

This condition is evidently incompatible with the single traveling wave (5). However, this solution may
be readily generalized using the fact that the dispersion-free 1D wave equation,
 2 1 2 
 2  2 2  E  0 , (7.58)
 z v t 
supports waves propagating, with the same speed, in any of two opposite directions. As a result, the
following linear superposition of two such waves,

E z 0  f ( z  vt )  f ( z  vt ) , (7.59)

satisfies both the equation and the boundary condition (57), for an arbitrary function f. The second term
on the right-hand side of Eq. (59) may be interpreted as a result of total reflection of the incident wave
(described by its first term) – in this particular case, with the change of the electric field’s sign. This
means, in particular, that within the macroscopic model, a conductor acts as a perfect mirror. By the
way, since the vector n of the reflected wave is opposite to that of the incident one (see the arrows in
Fig. 8), Eq. (6) shows that the magnetic field of the wave does not change its sign at the reflection:
1
H z 0   f ( z  vt )  f ( z  vt ) . (7.60)
Z

n incident

0 z
n reflected Fig. 7.8. A snapshot of the electric field at the
reflection of a sinusoidal wave from a perfect
conductor: a realistic pattern (red lines) and its
macroscopic, ideal-mirror approximation (blue
lines). Dashed lines show the snapshots after a
half-period time delay (t = ).
 

The blue lines in Fig. 8 show the resulting pattern (59) for the simplest, monochromatic wave:

 
Wave’s
total
reflection
E z 0  Re E e i (kz t )  E e i ( kz t ) . (7.61a)

Depending on convenience in a particular context, this pattern may be legitimately represented and
interpreted either as the linear superposition (61a) of two traveling waves or as a single standing wave:
i t   / 2  
E z 0    
 2 Im E e it sin kz  2 Re iE e it sin kz  2 Re  E e  sin kz , (7.61b)

in which the electric and magnetic field oscillate with the phase shifts by /2 both in time and space:

Chapter 7 Page 17 of 70
Essential Graduate Physics EM: Classical Electrodynamics

E E  E 
H  Re   e i ( kz t )   e i (  kz t )   2 Re  e it  cos kz .
z 0 (7.62)
Z Z   Z 
As a result of this shift, the time average of the Poynting vector’s magnitude,

S ( z , t )  EH 
1
Z
 
Re E2 e 2it sin 2kz , (7.63)

equals zero, showing that at the total reflection, there is no average power flow. (This is natural because
the perfect mirror can neither transmit the wave nor absorb it.) However, Eq. (63) shows that the
standing wave features local oscillations of energy, transferring it periodically between the
concentrations of the electric and magnetic fields, separated by the distance z = /2k = /4.
In the case of the sinusoidal waves, the reflection effects may be readily explored even for the
more general case of dispersive and/or lossy (but still linear) media in which () and (), and hence
the wave vector k() and the wave impedance Z(), defined by Eqs. (28), are certain complex functions
of frequency. The “only” new factor we have to account for is that in this case, the reflection may not be
total, so inside the second medium we have to use the traveling-wave solution as well. This factor may
be taken care of by looking for the solution to our boundary problem in the form

E z0   
 Re E  e ik  z  R e  ik  z e  i  t ,  E z0  ik z

 Re E  Te  e  i  t , (7.64)
Wave’s
partial
reflection

and hence, according to Eq. (6),

H
 E
 
 Re   e ik z  R e ik z e it ,
z 0   E ik z 
H z 0  Re   Te  e it  . (7.65)
 Z  ( )   Z  ( ) 
(The indices + and – correspond to the media located at z > 0 and z < 0, respectively.) Please note the
following important features of Eqs. (64)-(65):
(i) They satisfy the Maxwell equations in both media. (Historically, the fact that at z > 0, these
solutions do not include any components proportional to exp{ik–z}, looked surprising and was called the
wave extinction paradox.)
(ii) Due to the problem’s linearity, we could (and did :-) take the complex amplitudes of the
reflected and transmitted wave proportional to that (E) of the incident wave, while scaling them with
dimensionless, generally complex coefficients R and T. As a comparison of Eqs. (64)-(65) with Eqs.
(61)-(62) shows, the total reflection from an ideal mirror corresponds to R = –1 and T = 0.
(iii) Since in our current problem, the incident wave arrives from one side only (from z = –),
there is no need to include a term proportional to exp{–ik+z} into Eqs. (64)-(65) – even though this term
is also a legitimate solution of our wave equation. However, we would need to add such a term if the
medium at z > 0 had been nonuniform (e.g., had at least one more interface or any other inhomogeneity),
because the wave reflected from that additional inhomogeneity would be incident on our interface
(located at z = 0) from the right.
(iv) Eqs. (64)-(65) may be used even for the description of the cases when waves cannot
propagate to z  0, for example, a conductor or a plasma with p > . Indeed, the exponential drop of
the field amplitude at z > 0 in such cases is automatically described by the imaginary part of the wave
number k+ – see Eq. (29).

Chapter 7 Page 18 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

In order to calculate the coefficients R and T, we need to use boundary conditions at z = 0. Since
in our current case of the normal incidence, the reflection does not change the transverse character of the
partial waves, both vectors E and H remain tangential to the interface plane (in our notation, z = 0).
Reviewing the arguments that have led us, in statics, to the boundary conditions (3.37) and (5.117) for
these components, we see that they remain valid for the time-dependent situation as well,28 so for our
current case of normal incidence, we may write:
E z  0 E z  0 , H z  0 H z  0 . (7.66)

Plugging Eqs. (64)-(65) into these conditions, we readily get two equations for the coefficients R and T:
1
1 R  T, 1  R   1 T . (7.67)
Z Z
Solving this simple system of linear equations, we get29
Reflection
Z  Z 2Z 
and
transmission:
R , T . (7.68)
sharp interface
Z  Z Z  Z
These formulas are very important, and much more general than one might think because they
are applicable for virtually any 1D waves – electromagnetic or not, provided that the impedance Z is
defined properly.30 Since in the general case the wave impedances Z defined by Eq. (28) with the
corresponding indices, are complex functions of frequency, Eqs. (68) show that R and T may have
imaginary parts as well. This fact has important consequences at z < 0, where the reflected wave,
proportional to R, combines (“interferes”) with the incident wave. Indeed, with R = R ei (where  
arg R is a real phase shift), the expression in the parentheses in the first of Eqs. (64) becomes

e ik z  R e ik z  1  R  R e ik z  R e i e ik z

  (7.69)
 1  R e ik z  2 R e i / 2 sink   z    , where    .
2k 
This means that the field may be represented as a sum of a traveling wave and a standing wave, with an
amplitude proportional to  R , shifted by the distance – toward the interface, relative to the ideal-mirror
pattern (61b) – see Fig. 8. This effect is frequently used for the experimental measurements of an
unknown impedance Z+ of some medium, provided than Z – is known – most often, it is the free space,
where Z– = Z0. For that, a small antenna (the probe), not disturbing the fields’ distribution too much, is
placed into the wave field, and the amplitude of the ac voltage induced in it by the wave is measured
with a detector (e.g., a semiconductor diode with a nearly-quadratic I-V curve), as a function of z (Fig.

28 For example, the first of Eqs. (66) may be obtained by integrating the full (time-dependent) Maxwell equation
E + B/t = 0 over a narrow and long rectangular contour with dimensions l and d (d << l) stretched along the
interface. At the application of the Stokes theorem to this integral, the first term gives El, while the contribution
of the second term is proportional to the product ld, so its contribution at d/l  0 is negligible. The proof of the
second boundary condition is similar – as was already discussed in Sec. 6.2.
29 Please note that only the media impedances (rather than their wave velocities) are important for reflection in

this case! Unfortunately, this fact is not clearly emphasized in some textbooks that discuss only the case  = 0,
when Z = (0/)1/2 and v = 1/(0)1/2 are proportional to each other.
30 See, e.g., the discussion of elastic waves of mechanical deformations in CM Secs. 6.3, 6.4, 7.7, and 7.8.

Chapter 7 Page 19 of 70
Essential Graduate Physics EM: Classical Electrodynamics

9). From the results of such a measurement, it is straightforward to find both R  and -, and hence
restore the complex R, and then use Eq. (68) to calculate both the modulus and the argument of Z+.
(Before computers became ubiquitous, a specially lined paper called the Smith chart, had been
frequently used for performing this recalculation graphically; even nowadays, it is still used for result
presentation.)

V  E 2 ( z, t )
z

Fig. 7.9. Measurement of the complex

impedance of a medium (schematically).

Now let us discuss what these results give for waves incident from the free space (Z–() = Z0 =
const, k– = k0 = /c) onto the surfaces of two particular important media.
(i) For a collision-free plasma (with negligible magnetization) we may use Eq. (36) with () =
0, to represent the impedance (28) in either of two equivalent forms:
 
Z  Z0  iZ 0 . (7.70)
 2
 p
2 1/ 2
 2
p 2 
1/ 2

The first of these forms is more convenient in the case  > p, when the wave vector k+ and the wave
impedance Z+ of the plasma are real, so a part of the incident wave does propagate into it. Plugging this
expression into the latter of Eqs. (68), we see that T is real as well:
2
T . (7.71)
     p2 
2 1/ 2

Note that according to this formula, and somewhat counter-intuitively, T > 1 for any frequency (above
p), inviting the question: how can the transmitted wave be more intensive than the incident one that has
induced it? To answer this question, we need to compare the powers (rather than the electric field
amplitudes) of these two waves, i.e. their average Poynting vectors (42):

S incident 
E
2

S 
TE
2


E
2

4  2   p2 1/ 2

    
, . (7.72)

2
2Z 0 2Z  2Z 0 2
 2 1/ 2
p

The ratio of these two values31 is always below 1 (and tends to zero at   p), so only a fraction of the
incident wave power may be transmitted. Hence the result T > 1 may be interpreted as follows: an
interface between two media may be an impedance transformer: it can never transmit more power than
the incident wave provides, i.e. can only decrease the product S = EH, but since the ratio Z = E/H
changes at the interface, the amplitude of one of the fields may increase at the transmission.
Now let us proceed to case  < p when the waves cannot propagate in the plasma. In this case,
the second of the expressions (70) is more convenient, because it immediately shows that Z+ is purely

31 Thisratio is sometimes also called the “wave transmission coefficient”, but to avoid its confusion with the T
defined by Eq. (64), it is better to call it the power transmission coefficient.

Chapter 7 Page 20 of 70
Essential Graduate Physics EM: Classical Electrodynamics

imaginary, while Z- = Z0 is purely real. This means that (Z+ – Z-) = (Z+ + Z-)*, i.e. according to the first of
Eqs. (68),  R  = 1, so the reflection is total, i.e. no incident power (on average) is transferred into the
plasma – as was already discussed in Sec. 2. However, the complex R has a finite argument,

  arg R  2 arg( Z   Z 0 )  2 tan 1 , (7.73)
 2
p 2 
1/ 2

and hence provides a finite spatial shift (69) of the standing wave toward the plasma surface:
  c 
   tan 1 . (7.74)
2k 0   2
p 2 
1/ 2

On the other hand, we already know from Eq. (40) that the solution at z > 0 is exponential, with
the decay length  described by Eq. (39). Calculating, from the coefficient T, the exact coefficient
before this exponent, it is straightforward to verify that the electric and magnetic fields are indeed
continuous at the interface, completing the pattern shown with red lines in Fig. 8. This wave penetration
into a fully reflecting material may be experimentally observed, for example, by thinning its sample.
Even without solving this problem exactly, it is evident that if the sample’s thickness d becomes
comparable to , a part of the exponential “tail” of the field reaches its second interface, and induces a
propagating wave. This is a classical-electromagnetic analog of the quantum-mechanical tunneling
through a potential barrier.32
Note that at low frequencies, both – and  tend to the same frequency-independent value,
1/ 2 1/ 2
c c 2  0 me   me  
,        , at  0, (7.75)
 p  ne 2 
   0 ne
2 
 p
which is just the field penetration depth (6.44) calculated for a perfect conductor model (assuming m =
me and  = 0) in the quasistatic limit. This is natural, because the condition  << p may be recast as 0
 2c/ >> 2c/p  2, i.e. as the quasistatic approximation’s validity condition.
(ii) Now let us consider electromagnetic wave’s reflection from an Ohmic, non-magnetic
conductor. In the simplest low-frequency limit, when  is much less than 1, the conductor may be
described by a frequency-independent conductivity . 33 According to Eq. (46), in this case we can take
1/ 2
 0 
Z    . (7.76)
  ( )  i /  
 opt 
With this substitution, Eqs. (68) immediately give us all the results of interest. In particular, in the most
important quasistatic limit when s  (2/0)1/2 << 0  2c/, i.e. / >> 0 ~ opt, the conductor’
impedance is low:
1/ 2 1/ 2
   2 s Z
Z   0      Z0 , i.e.   1. (7.77)
 i   i  0 Z0

32 See, e.g., QM Sec. 2.3.

33In a typical metal,  ~ 10-13s, so this approximation works well up to  ~ 1013 s-1, i.e. up to the far-infrared
frequencies.

Chapter 7 Page 21 of 70
Essential Graduate Physics EM: Classical Electrodynamics

This impedance is complex, and hence some fraction f of the incident wave is absorbed by the
conductor. The fraction may be found as the ratio of the dissipated power (either calculated, as was done
above, from Eqs. (68), or just taken from Eq. (6.36), with the magnetic field amplitude  H  = 2 E /Z0
– see Eq. (62)) to the incident wave’s power given by the first of Eqs. (72). The result,
2 s 
f   4 s  1 . (7.78)
c 0
is used for crude estimates of the energy dissipation in metallic-wall waveguides and resonators. It
shows that to keep the energy losses low, the characteristic size of such systems (which gives a scale of
the free-space wavelengths 0 at which they are used) should be much larger than s. A more detailed
theory of these structures, and of the energy loss in them, will be discussed later in this chapter.

7.4. Refraction
Now let us consider the effects arising at a plane interface between two uniform media when the
wave’s incidence angle  (Fig. 10) is arbitrary rather than equal to zero as in our previous analysis, for
the simplest case of fully transparent media, with real () and (). (For the sake of notation
simplicity, in most formulas below, the argument of these functions will be dropped, i.e. just implied.)

z
r k

  ,  0 k  sin r
  ,   k  sin  k ' sin  ' x

 ' Fig. 7.10. Plane wave’s reflection, transmission, and

refraction at a plane interface. The plane of the drawing is
k k ' selected to contain all three wave vectors: k+, k-, and k’-.

In contrast with the case of normal incidence, here the wave vectors k–, k’–, and k+ of the three
components (incident, reflected, and transmitted) waves may have different directions. (Such change of
the transmitted wave’s direction is called refraction.) Hence let us start our analysis by writing a general
expression for a single plane, monochromatic wave for the case when its wave vector k has all three
Cartesian components, rather than one. An evident generalization of Eq. (11) for this case is

 i k x  k y y  k z z )  t  
f (r, t )  Re  f  e x


  Re f  e


i (k r t )
. (7.79)

This expression enables a ready analysis of “kinematic” relations, which are independent of the
media impedances. Indeed, it is sufficient to notice that to satisfy any linear, homogeneous boundary
conditions at the interface (z = 0), all partial plane waves must have the same temporal and spatial
dependence on this plane. Hence if we select the x-z plane so that the vector k– lies in it, then (k–)y = 0,
and k+ and k’– cannot have any y-component either, i.e. all three wave vectors lie in the same plane –

Chapter 7 Page 22 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

that is selected as the plane of the drawing in Fig. 10. Moreover, because of the same reason, their x-
components should be equal:
k  sin   k ' sin  '  k  sin r . (7.80)
From here we immediately get two well-known laws: of reflection
Reflection
angle '  , (7.81)
and of refraction:34
sin r k 
Snell  . (7.82)
law sin  k 
In this form, the laws are valid for plane waves of any nature. In optics, the Snell law (82) is frequently
represented in the form
sin r n
 , (7.83)
sin  n
where n is the index of refraction (also called the “refractive index”) of the corresponding medium,
defined as its wave number normalized to that of the free space (at the particular wave’s frequency):
1/ 2
k   
Index
n        . (7.84)
of refraction
k0   0 0 
Perhaps the most famous corollary of the Snell law is that if a wave propagates from a medium
with a higher index of refraction to that with a lower one (i.e. if n- > n+ in Fig. 10), for example from
water to air, there is always a certain critical value c of the incidence angle,
1/ 2
n   
Critical
 c  sin 1
 sin 1     , (7.85)
   
angle
n 
at which the refraction angle r (see Fig. 10 again) reaches /2. At a larger , i.e. within the range c < 
< /2, the boundary conditions (80) cannot be satisfied by a refracted wave with a real wave vector, so
the wave experiences the so-called total internal reflection. This effect is very important for practice
because it means that dielectric surfaces may be used as optical mirrors, in particular in optical fibers –
to be discussed in more detail in Sec. 7 below. This is very fortunate for telecommunication technology
because light’s reflection from metals is rather imperfect. Indeed, according to Eq. (78), in the optical
range (0 ~ 0.5 m, i.e.  ~ 1015 s-1), even the best conductors (with  ~ 6108 S/m and hence the
normal skin depth s ~ 1.5 nm) provide power loss of at least a few percent at each reflection.
Note, however, that even within the range c <  < /2, the field at z > 0 is not identically equal
to zero: it penetrates into the lower-n media by a distance of the order of 0, exponentially decaying
inside it, just as it does at the normal incidence – see Fig. 8. However, at   0 the penetrating field still
propagates, with the wave number (80), along the interface. Such a field, exponentially dropping in one
direction but still propagating as a wave in another direction, is commonly called the evanescent wave.

34 The latter relation is traditionally called the Snell law, after a 17th-century astronomer Willebrord Snellius, but
it has been traced all the way back to a circa 984 work by Abu Saad al-Ala ibn Sahl. (Claudius Ptolemy who
performed pioneering experiments on light refraction in the 2nd century AD, was just one step from this result.)

Chapter 7 Page 23 of 70
Essential Graduate Physics EM: Classical Electrodynamics

One more remark: just as at the normal incidence, the field’s penetration into another medium
causes a phase shift of the reflected wave – see, e.g., Eq. (69) and its discussion. A new feature of this
phase shift, arising at   0, is that it also has a component parallel to the interface – the so-called Goos-
Hänchen effect. In geometric optics, this effect leads to an image shift (relative to its position in a
perfect mirror) with components both normal and parallel to the interface.
Now let us carry out an analysis of “dynamic” relations that determine amplitudes of the
refracted and reflected waves. For this, we need to write explicitly the boundary conditions at the
interface (i.e. the plane z = 0). Since now the electric and/or magnetic fields may have components
normal to the plane, in addition to the continuity of their tangential components, which were repeatedly
discussed above,
E x , y z  0  E x , y z  0 , H x , y z  0  H x , y z  0 , (7.86)

we also need relations for the normal components. As it follows from the homogeneous macroscopic
Maxwell equations (6.99b), they are also the same as in statics, i.e. Dn = const, and Bn = const, for our
coordinate choice (Fig. 10) giving

  Ez z  0    Ez z  0 ,  H z z  0   H z z  0 . (7.87)

The expressions of these components via the amplitudes E, RE, and TE of the incident,
reflected, and transmitted waves depend on the incident wave’s polarization. For example, for a linearly-
polarized wave with the electric field vector normal to the plane of incidence, i.e. parallel to the
interface plane, the reflected and refracted waves are similarly polarized – see Fig. 11a.

z (a) z (b)
H k
k
 H
r r 
E
  ,    ,  E

  ,  0 x   ,  0 x

k- H'- k- E'-
   
H-
E E'-  H-  H- '⊙

'
- k - E- k '-

Fig. 7.11. Reflection and refraction at two different linear polarizations of the incident wave.

As a result, all Ez are equal to zero (so the first of Eqs. (87) is inconsequential), while the
tangential components of the electric field are equal to their full amplitudes, just as at the normal
incidence, so we still can use Eqs. (64) expressing these components via the coefficients R and T.
However, at   0 the magnetic fields have not only tangential components
E  E 
Hx z  0  Re   (1  R) cos  e it , Hx z  0  Re   T cos r e it  , (7.88)
 Z   Z 
but also normal components (see Fig. 11a):

Chapter 7 Page 24 of 70
Essential Graduate Physics EM: Classical Electrodynamics

E  E 
Hz z  0  Re   (1  R) sin  e it , Hz z  0  Re   T sin r e it  . (7.89)
 Z   Z 
Plugging these expressions into the boundary conditions expressed by Eqs. (86) (in this case, for
the y-components only) and the second of Eqs. (87), we get three equations for two unknown
coefficients R and T. However, two of these equations duplicate each other because of the Snell law, and
we get just two independent equations,
1
1 R  T, 1  R  cos  1 T cos r , (7.90)
Z Z
which are a very natural generalization of Eqs. (67), with the replacements Z-  Z–cosr, Z+  Z+cos.
As a result, we can immediately use Eq. (68) to write the solution of the system (90):35
Z  cos   Z  cos r 2 Z  cos 
R , T . (7.91a)
Z  cos   Z  cos r Z  cos   Z  cos r
If we want to express these coefficients via the angle of incidence alone, we should use the Snell
law (82) to eliminate the angle r, getting frequently quoted bulkier expressions:

R

Z  cos   Z  1  (k  / k  ) 2 sin 2  1/ 2

, T
2Z  cos 
. (7.91b)
Z  cos   Z  1  (k 2
 / k  ) sin
2

1/ 2

Z  cos   Z  1  (k  / k  ) 2 sin 2  
1/ 2

However, conceptually it is preferable to use the kinematic relation (82) and the dynamic relations (91a)
separately, because Eq. (91b) obscures the very important physical fact that the ratio of k, i.e. of the
wave velocities of the two media, is only involved in the Snell law, while Eqs. (91b) explicitly include
only the wave impedances – just as in the case of normal incidence.
In the opposite case of the linear polarization of the electric field within the plane of incidence
(Fig. 11b), it is the magnetic field that does not have a normal component, so it is now the second of
Eqs. (87) that does not participate in the solution. However, now the electric fields in the two media
have not only tangential components,

Ex z  0 
 Re E (1  R) cos  e it ,  Ex z  0  
 Re E T cos r e it , (7.92)

but also normal components (Fig. 11b):

Ez z  0  E (1  R) sin  , Ez z  0   E T sin r. (7.93)

As a result, instead of Eqs. (90), the reflection and transmission coefficients are related as
1
(1  R) cos   T cos r , 1  R   1 T . (7.94)
Z Z
Again, the solution of this system may be immediately written using the analogy with Eq. (67):

35Note that we may calculate the reflection and transmission coefficients R’ and T’ for the wave traveling in the
opposite direction just by making the following parameter swaps: Z+  Z- and   r, and that the resulting
coefficients satisfy the following Stokes relations: R’ = –R, and R2 + TT’ = 1, for any Z.

Chapter 7 Page 25 of 70
Essential Graduate Physics EM: Classical Electrodynamics

Z  cos r  Z  cos  2 Z  cos 

R , T , (7.95a)
Z  cos r  Z  cos  Z  cos r  Z  cos 
or, alternatively, using the Snell law, in a more bulky form:

R

Z  1  (k  / k  ) 2 sin 2  
1/ 2
 Z  cos 
, T
2Z  cos 
. (7.95b)
Z 1  (k 
2
/ k  ) sin 2

1/ 2
 Z  cos  Z 1  (k 
2
/ k  ) sin 2

1/ 2
 Z  cos 
For the particular case + = – = 0, when Z+/Z– = (–/+)1/2 = k–/k+ = n–/n+ (which is
approximately correct for traditional optical media), Eqs. (91b) and (95b) are called the Fresnel
formulas.36 Most textbooks are quick to point out that there is a major difference between them: while
for the electric field polarization within the plane of incidence (Fig. 11b), the reflected wave’s amplitude
(proportional to the coefficient R) turns to zero37 at a special value of  (called the Brewster angle):38
n
 B  tan 1  , (7.96)
n
while there is no such angle in the opposite case (shown in Fig. 11a). However, note that this statement,
as well as Eq. (96), is true only for the case + = –. In the general case of different  and , Eqs. (91)
and (95) show that the reflected wave vanishes at  = B with

            /   , for E  n z (Fig.11a),
tan 2  B   (7.97) Brewster
            /   , for H  n z (Fig.11b). angle

Note the natural    symmetry of these relations, resulting from the E  H symmetry for
these two polarization cases (Fig. 11). These formulas also show that for any set of parameters of the
two media (with ,  > 0), tan2B is positive (and hence a real Brewster angle B exists) only for one of
these two polarizations. In particular, if the interface is due to the change of  alone (i.e. if + = -), the
first of Eqs. (97) is reduced to the simple form (96) again, while for the polarization shown in Fig. 11b,
there is no Brewster angle, i.e. the reflected wave has a non-zero amplitude for any .
Such an account of both media parameters,  and , on an equal footing is necessary to describe
several interesting effects. The first of them is the so-called negative refraction.39 As was shown in Sec.

36 Named after Augustin-Jean Fresnel (1788-1827), one of the wave optics pioneers, who is credited, among
many other contributions (see, in particular, discussions in Ch. 8), for the concept of light as a purely transverse
wave.
37 This effect is used in practice to obtain linearly polarized light, with the electric field vector perpendicular to
the plane of incidence, from the natural light with its random polarization. An even more widespread application
of the effect is a partial reduction of undesirable glare from wet pavement (for the water/air interface, n+/n–  1.33,
giving B  50) by covering glasses and car headlights with thin vertically-polarizing layers.
38 A very simple interpretation of Eq. (96) is based on the fact that, together with the Snell law (82), it gives r + 

= /2. As a result, the vector E+ is parallel to the vector k’–, and hence the oscillating electric dipoles of the
medium do not have the Cartesian component that could induce the transverse electric field E’– of the potential
reflected wave.
39 Despite some important background theoretical work by A. Schuster (1904), L. Mandelstam (1945), D.
Sivikhin (1957), and especially V. Veselago (1966-67), the negative refractivity effects became a subject of
intensive scientific research and engineering development only in the 2000s.

Chapter 7 Page 26 of 70
Essential Graduate Physics EM: Classical Electrodynamics

2, in a medium with electric-field-driven resonances, the function () may be almost real and negative,
at least within limited frequency intervals – see, in particular, Eq. (34) and Fig. 5. As has already been
discussed, if, at these frequencies, the function () is real and positive, then k2() = 2()() < 0,
and k may be represented as i/ with a real , meaning the exponential field decay into the medium.
However, let us consider the case when both () < 0 and () < 0 at a certain frequency. (This is
possible in a medium with both E-driven and H-driven resonances, at a proper choice of their resonance
frequencies.) Since in this case k2() = 2()() > 0, the wave vector is real, Eq. (79) describes a
traveling wave, and one could think that there is nothing new in this case. Not so!
First of all, for a sinusoidal plane wave (79), the operator  is equivalent to the multiplication by
ik. As the Maxwell equations (2a) show, this means that at a fixed direction of vectors E and k, the
simultaneous reversal of signs of  and  means the reversal of the direction of the vector H. Namely, if
both  and  are positive, these equations are satisfied with mutually orthogonal vectors {E, H, k}
forming the usual, right-hand system (see Fig. 1 and Fig. 12a), the name stemming from the popular
“right-hand rule” used to determine the vector product’s direction. However, if both  and  are
negative, the vectors form a left-hand system – see Fig. 12b. (Due to this fact, the media with  < 0 and
 < 0 are frequently called the left-handed materials, LHM for short.) According to the basic relation
(6.114), which does not involve media parameters, this means that for a plane wave in a left-hand
material, the Poynting vector S = EH, i.e. the energy flow, is directed opposite to the wave vector k.

(a) (b)
S S
k
H H Fig. 7.12. Directions of the main
vectors of a plane wave inside a
medium with (a) positive and (b)
E k negative values of  and .
E

This fact may look strange but is in no contradiction with any fundamental principle. Let me
remind you that, according to the definition of the vector k, its direction shows the direction of the phase
velocity vph = /k of a sinusoidal (and hence infinitely long) wave, which cannot be used, for example,
for signaling. Such signaling (by sending wave packets – see Fig. 13) is possible only with the group
velocity vgr = d/dk. This velocity in left-hand materials is always directed (as in the right-hand
materials) along the vector S, i.e. along the wave’s energy flow.

f ( z, t )
v ph v gr

Fig. 7.13. An example of a wave

packet moving along axis z with a
z negative phase velocity, but positive
group velocity. Blue lines show a
packet’s snapshot a short time interval
after the first snapshot (red lines).

Chapter 7 Page 27 of 70
Essential Graduate Physics EM: Classical Electrodynamics

Perhaps the most fascinating effect possible with left-hand materials is the wave refraction at
their interfaces with the usual, right-handed materials – first predicted by V. Veselago in 1960. Consider
the example shown in Fig. 14a. In the incident wave, arriving from a usual material, the directions of
the vectors k– and S– coincide, and so they are in the reflected wave with vectors k’– and S’–. This
means that the electric and magnetic fields in the interface plane (z = 0) are, at our choice of the
coordinate axes, proportional to exp{ikxx}, with a positive component kx = k–cos . To satisfy any linear
boundary conditions, the refracted wave, propagating into the left-handed material, has to match that
dependence, i.e. have a positive x-component of its wave vector k+. But in this medium, this vector has
to be antiparallel to the vector S, which in turn should be directed out of the interface, because it
represents the power flow from the interface into the material’s bulk. These conditions cannot be
reconciled by the refracted wave propagating along the usual Snell-law direction (shown with the
dashed line in Fig. 13a), but are all satisfied at refraction in the direction given by Snell’s angle with the
opposite sign. (Hence the term “negative refraction”).40

z (a) (b)
k
r r image
S b
   0,    0 bd a
kx x d 2d
   0,    0 a

S  
'
S' ad
k k  object
Fig. 7.14. Negative refraction: (a) waves at the interface between media with positive and negative values
of , and (b) the hypothetical perfect lens: a parallel plate made of a material with  = –0 and  = –0.

In order to understand how unusual the results of the negative refraction may be, let us consider
a parallel slab of thickness d, made of a hypothetical left-handed material with exactly selected values 
= –0, and  = –0 (see Fig. 14b). For such a material, placed in free space, the refraction angle r = –,
so the rays from a point source, located in free space at a distance a < d from the slab’s surface,
propagate as shown on that panel, i.e. all meet again at the distance a beyond the surface, and then
continue to propagate to the second surface of the slab. Repeating our discussion for this surface, we see
that a point’s image is also formed beyond the slab, at distance 2a + 2b = 2a + 2(d – a) = 2d from the
object.
Superficially, this system looks like a usual lens, but the well-known lens formula, which relates
a and b with the focal length f, is not satisfied. (In particular, a parallel beam is not focused into a point
at any finite distance.) As an additional difference from the usual lens, the system shown in Fig. 14b
does not reflect any part of the incident light. Indeed, it is straightforward to check that for all the above
formulas for R and T to be valid, the sign of the wave impedance Z in left-handed materials has to be
kept positive. Thus, for our particular choice of parameters ( = –0,  = –0), Eqs. (91a) and (95a) are

40 In some publications inspired by this fact, the left-hand materials are prescribed a negative index of refraction
n. However, this prescription should be treated with care. For example, it complies with the first form of Eq. (84),
but not its second form, and the sign of n, in contrast to that of the wave vector k, is a matter of convention.

Chapter 7 Page 28 of 70
Essential Graduate Physics EM: Classical Electrodynamics

valid with Z+ = Z- = Z0 and cos r = cos  = 1, giving R = 0 for any linear polarization, and hence for any
other wave polarization – circular, elliptic, natural, etc.
The perfect lens suggestion has triggered a wave of efforts to implement left-hand materials
experimentally. (Attempts to find such materials in nature have failed so far.) Most progress in this
direction has been achieved using the so-called metamaterials, which are essentially quasi-periodic
arrays of specially designed electromagnetic resonators, ideally with high density n >> -3. For example,
Fig. 15 shows the metamaterial that was used for the first demonstration of negative refractivity in the
microwave region – for ~10-GHz waves.41 It combines straight strips of a metallic film, working as
lumped resonators with a large electric dipole moment (and hence strongly coupled to the wave’s
electric field E), and several almost-closed film loops (so-called split rings), working as lumped
resonators with large magnetic dipole moments, strongly coupled to the field H. The negative
refractivity is achieved by designing the resonance frequencies close to each other. More recently,
metamaterials with negative refractivity were demonstrated in the optical range as well,42 although to
the best of my knowledge, their relatively large absorption still prevents practical applications.

Fig. 7.15. An artificial left-hand material

providing negative refraction at microwave
frequencies ~10 GHz. The original by Jeffrey
D. Wilson (in the public domain) is available
at https://en.wikipedia.org/wiki/Metamaterial.

This progress has stimulated the development of other potential uses of metamaterials (not
necessarily the left-handed ones), in particular, designs of nonuniform systems with handcrafted
distributions (r, ) and (r, ) that may provide electromagnetic wave propagation along the desired
paths, e.g., around a certain region of space, making it virtually invisible for an external observer – so
far, within very limited frequency ranges.43
As was mentioned in Sec. 5.5, another way to reach negative values of () is to place a
ferromagnetic material into such an external dc magnetic field that the frequency r of the ferromagnetic
resonance is somewhat lower than . If thin layers of such a material (e.g., nickel) are interleaved with
layers of a non-magnetic good conductor (such as copper), the average value of () of the resulting
metamaterial may be positive but substantially below 0. According to Eq. (6.33), the skin-depth s of
such a material may be larger than that of the good conductor alone, enforcing a more uniform
distribution of the ac current flowing along the layers, and hence making the energy losses lower than in
the good conductor alone. This effect may be useful, in particular, for electronic circuit interconnects.44

41 R. Shelby et al., Science 292, 77 (2001); J. Wilson and Z. Schwartz, Appl. Phys. Lett. 86, 021113 (2005).
42 See, e.g., J. Valentine et al., Nature 455, 376 (2008).
43 For a review of such “invisibility cloaks”, see, e.g., B. Wood, Comptes Rendus Physique 10, 379 (2009).
44 See, for example, N. Sato et al., J. Appl. Phys. 111, 07A501 (2012), and references therein.

Chapter 7 Page 29 of 70
Essential Graduate Physics EM: Classical Electrodynamics

7.5. Transmission lines: TEM waves

So far, we have analyzed plane electromagnetic waves, implying that their cross-section is
infinite – evidently, an unrealistic assumption. The cross-section may be limited, still sustaining wave
propagation, using wave transmission lines:45 long, uniform structures made of either good conductors
or dielectrics. Let us first discuss the first option, using the following simplifying assumptions:
(i) the structure is a cylinder (not necessarily with a round cross-section, see Fig. 16) filled with a
usual (right-handed), uniform dielectric material with negligible energy losses (” = ” = 0), and
(ii) the wave attenuation due to the skin effect is also negligibly low. (As Eq. (78) indicates, for
that the characteristic size a of the line’s cross-section has to be much larger than the skin-depth s of its
wall material. The energy dissipation effects will be analyzed in Sec. 9 below.)

Ezn z E

Fig. 7.16. Electric field’s

decomposition in a transmission
0 y
line (in particular, a waveguide).
x Et
With such exclusion of energy losses, we may look for a particular solution of the macroscopic
Maxwell equations in the form of a monochromatic wave traveling along the line:


E(r , t )  Re E  ( x, y ) e
i ( k z z t )
, 
H (r , t )  Re H  ( x, y ) e
i ( k z z t )
, (7.98)

with real kz, where the z-axis is directed along the transmission line – see Fig. 16. Note that this form
allows a substantial coordinate dependence of the electric and magnetic field within the plane [x, y] of
the transmission line’s cross-section, as well as nonvanishing longitudinal components Ez and/or Hz of
the fields, so the solution (98) is substantially more general than the plane waves discussed above. We
will see in a minute that as a result, the parameter kz may be very much different from its plane-wave
value (13), k  ()1/2, in the same material, at the same frequency.
In order to describe these effects quantitatively, let us decompose the complex amplitudes of the
wave’s fields into their longitudinal and transverse components (Fig. 16):46
E  Ez n z  Et , H  H z n z  H t . (7.99)

45 Another popular term is the waveguide, but it is typically reserved for the transmission lines with singly-
connected cross-sections, to be analyzed in the next section. The first structure for guiding waves was proposed
by J. J. Thomson in 1893, and experimentally tested by O. Lodge in 1894.
46 For the notation simplicity, I am dropping index  in the complex amplitudes of the field components, and also

have dropped the argument  in kz and Z, even though these parameters may depend on the wave’s frequency
rather substantially – see below.

Chapter 7 Page 30 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

Plugging Eqs. (98)-(99) into the source-free Maxwell equations (2), and requiring the longitudinal and
transverse components to be balanced separately, we get
ik z n z  E t  iH t   t  E z n z , ik z n z  H t  iE t   t  H z n z ,
 t  E t  iH z n z ,  t  H t  iE z n z , (7.100)
 t  E t  ik z E z ,  t  H t  ik z H z .
where t is the 2D del operator acting in the transverse plane [x, y] only, i.e. the usual , but with /z
= 0. The system (100) looks even bulkier than the original equations (2), but it is much simpler for
analysis. Indeed, by eliminating the transverse components from these equations (or, even simpler, just
by plugging Eq. (99) into Eqs. (3) and keeping only their z-components), we get a pair of self-consistent
equations for the longitudinal components of the fields, 47
2D Helmholtz
equations for
Ez and Hz
 2
t 
 k t2 E z  0,  2
t 
 k t2 H z  0 , (7.101)

where k is still defined by Eq. (13), k  ()1/2, while

Wave vector
component
balance
k t2  k 2  k z2   2   k z2 (7.102)

After the distributions Ez(x,y) and Hz(x,y) have been found from these equations, they provide right-hand
sides for the rather simple, closed system of equations (100) for the transverse components of field
vectors. Moreover, as we will see below, each of the following three types of solutions:
(i) with Ez = 0 and Hz = 0 (called the transverse electromagnetic, or TEM waves),
(ii) with Ez = 0, but Hz  0 (called either the TE waves or, more frequently, H-modes), and
(iii) with Ez 0, but Hz = 0 (the so-called TM waves or E-modes),
has its own dispersion law and hence its own wave propagation velocity; as a result, these modes (i.e.
the field distribution patterns) may be considered separately.
In the balance of this section, we will focus on the simplest, TEM waves (i), with no longitudinal
components of either field. For them, the top two equations of the system (100) immediately give Eqs.
(6) and (13), and kz = k. In plain English, this means that E = Et and H = Ht are proportional to each
other and are mutually perpendicular (just as in the plane wave) at each point of the cross-section and
that the TEM wave’s impedance Z  E/H and dispersion law (k), and hence the propagation speed, are
the same as in a plane wave in the same material. In particular, if  and  are frequency-independent
within a certain frequency range, the dispersion law within this range is linear,  = k/()1/2, and the
wave’s speed does not depend on its frequency. For practical applications to telecommunications, this is
a very important advantage of the TEM waves over their TM and TE counterparts – to be discussed in
the next sections.
Unfortunately for practice, such waves cannot propagate in every transmission line. To show
this, let us have a look at the two last lines of Eqs. (100). For the TEM waves (Ez = 0, Hz = 0, kz = k),
they are reduced to merely

47The wave equation represented in the form (101), even with the 3D Laplace operator, is called the Helmholtz
equation, named after Hermann von Helmholtz (1821-1894) – the mentor of H. Hertz and M. Planck, among
many others.

Chapter 7 Page 31 of 70
Essential Graduate Physics EM: Classical Electrodynamics

 t  E t  0,  t  H t  0,
(7.103)
 t  E t  0,  t  H t  0.
Within the coarse-grain description of the conducting walls of the line (i. e., neglecting not only the
screening depth but also the skin depth in comparison with the cross-section dimensions), we have to
require that inside them, E = H = 0. Close to a wall but outside it, the normal component En of the
electric field may be different from zero, because surface charges may sustain its jump – see Sec. 2.1, in
particular Eq. (2.3). Similarly, the tangential component H of the magnetic field may have a finite jump
at the surface due to skin currents – see Sec. 6.3, in particular Eq. (6.38). However, the tangential
component of the electric field and the normal component of the magnetic field cannot experience such
jumps, and to have them equal to zero inside the walls they have to equal zero just outside the walls as
well:
E  0, Hn  0 . (7.104)
But the left columns of Eqs. (103)-(104) coincide with the formulation of the 2D boundary
problem of electrostatics for the electric field induced by electric charges of the conducting walls, with
the only difference that in our current case, the value of  actually means (). Similarly, the right
columns of those relations coincide with the formulation of the 2D boundary problem of magnetostatics
for the magnetic field induced by currents in the walls, with   (), with the only difference is that in
our current coarse-grain approximation, the magnetic fields cannot penetrate into the conductors.
Now we immediately see that in waveguides with a singly-connected wall, for example, a hollow
conducting tube (see, e.g., Fig. 16), the TEM waves are impossible, because there is no way to create a
non-zero electrostatic field inside a conductor with such cross-section. However, such fields (and hence
the TEM waves) are possible in structures with cross-sections consisting of two or more disconnected
(galvanically-insulated) parts – see, e.g., Fig. 17.

Ht C

 
   
  Fig. 7.17. An example of the cross-
  section of a transmission line that may
support the TEM wave propagation.
Et

In order to derive “global” relations for such a transmission line, let us consider the contour C
drawn very close to the surface of one of its conductors – see, e.g., the red dashed line in Fig. 17. We
can consider it, on one hand, as the cross-section of a cylindrically-shaped Gaussian volume of a certain
elementary length dz <<  2/k. Using the generalized Gauss law (3.34), we get

 E  dr  , (7.105)

t n
C

where  (not to be confused with the wavelength !) is the complex amplitude of the linear density of
the electric charge of the conductor. On the other hand, the same contour C may be used in the
generalized Ampère law (5.116) to write

Chapter 7 Page 32 of 70
Essential Graduate Physics EM: Classical Electrodynamics

 H  dr  I  ,
C
t (7.106)

where I is the total current flowing along the conductor (or rather its complex amplitude). But, as was
mentioned above, in the TEM wave the ratio Et/Ht of the field components participating in these two
integrals is constant and equal to Z = (/)1/2, so Eqs. (105)-(106) give the following simple relation
between the “global” variables of the conductor:
 /   
I     . (7.107)
Z  1/ 2
k
This important relation may be also obtained in a different way; let me describe it as well,
because (as we will see below) it has an independent heuristic value. Let us consider a small segment dz
<<  = 2/k of the line’s conductor, and apply the electric charge conservation law (4.1) to the instant
values of the linear charge density and current. The cancellation of dz in both parts yields
( z , t ) I ( z , t )
 . (7.108)
t z
If we accept the sinusoidal waveform, exp{i(kz – t)}, for both these variables, we immediately recover
Eq. (107) for their complex amplitudes, showing that this relation expresses just the charge continuity
law.
The global equation (108) may be made more specific in the case when the frequency
dependence of  and  is negligible, and the transmission line consists of just two isolated conductors –
see, e.g., Fig. 17. In this case, to have the wave localized in the space near the two conductors, we need
a sufficiently fast decrease of its electric field at large distances. For that, their linear charge densities for
each value of z should be equal and opposite, and we can simply relate them to the potential difference V
between the conductors:
( z, t )
 C0 , (7.109)
V ( z, t )
where C0 is the mutual capacitance of the conductors (per unit length) – which was repeatedly discussed
in Chapter 2. Then Eq. (108) takes the following form:
V ( z , t ) I ( z , t )
C0  . (7.110)
t z
Next, let us consider the contour shown with the red dashed line in Fig. 18 (which shows a
different cross-section of the transmission line – by a plane containing the wave propagation axis z), and
apply to it the Faraday induction law (6.3).

I
I ( z, t ) I ( z, t )  dz
z

V
V ( z, t ) d  L0 I ( z , t ) dz V ( z, t )  dz
z Fig. 7.18. Electric current, magnetic flux, and
voltage in a two-conductor transmission line.

dz

Chapter 7 Page 33 of 70
Essential Graduate Physics EM: Classical Electrodynamics

Since, in the coarse-grain approximation, the electric field inside the conductors (in Fig. 18, on
the horizontal segments of the contour) vanishes, the total e.m.f. equals the difference of the voltages V
at the ends of the segment dz, while the only sources of the magnetic flux through the area limited by the
contour are the (equal and opposite) currents I in the conductors, we can use Eq. (5.70) to express the
flux. As a result, by canceling dz in both parts of the equation, we get
I ( z , t ) V ( z , t )
L0  , (7.111)
t z
where L0 is the mutual inductance of the conductors per unit length. The only difference between this L0
and the dc mutual inductances discussed in Chapter 5 is that at the high frequencies we are analyzing
now, L0 should be calculated neglecting the magnetic field penetration into the conductors. (In the dc
case, we had the same situation for superconductor electrodes within their coarse-grain, ideal-diamagnet
description.)
The system of Eqs. (110) and (111) is frequently called the telegrapher’s equations. Combined,
they give for any “global” variable f (either V, or I, or ) the usual 1D wave equation,
2 f 2 f
 L C
0 0  0, (7.112)
z 2 t 2
which describes dispersion-free TEM wave’s propagation. Again, this equation is only valid within the
frequency range where the frequency dependence of both  and  is negligible. If this is not so, the
global approach may still be used for sinusoidal waves f = Re[fexp{i(kz – t)}]. Repeating the above
arguments, instead of Eqs. (110)-(111) we get a more general system of two algebraic equations

C 0V  kI  , L0 I   kV , (7.113)

in which L0   and C0   may now depend on frequency. These equations are consistent only if
k2 1 L0C0
L0 C 0     . (7.114) product
2 v2 invariance

Besides the fact we have already known (that the TEM wave’s speed is the same as that of the plane
wave), Eq. (114) gives us the result that I confess was not emphasized enough in Chapter 5: the product
L0C0 does not depend on the shape or size of line’s cross-section, provided that the magnetic field’s
penetration into the conductors is negligible). Hence, if we have calculated the mutual capacitance C0 of
a system of two cylindrical conductors, the result immediately gives us their mutual inductance: L0 =
/C0. This relationship stems from the fact that both the electric and magnetic fields may be expressed
via the solution of the same 2D Laplace equation for the system’s cross-section.
With Eq. (114) satisfied, any of Eqs. (113) gives the same result for the following ratio:
1/ 2
V L  Transmission
ZW     0  , (7.115) line’s TEM
I   C0  Impedance

which is called the transmission line’s impedance. This parameter has the same dimensionality (in SI
units – ohms, denoted ) as the wave impedance (7),

Chapter 7 Page 34 of 70
Essential Graduate Physics EM: Classical Electrodynamics

1/ 2
E   
Z   , (7.116)
H   
but these parameters should not be confused, because ZW depends on the cross-section’s geometry, while
Z does not. In particular, ZW is the only important parameter of a transmission line for its matching with
a lumped load circuit (Fig. 19) in the important case when both the cable cross-section’s size and the
load’s linear dimensions are much smaller than the wavelength.48

ZW I Z L ( ) V Fig. 7.19. Passive, lumped

termination of a TEM
transmission line.
z0 z

Indeed, in this case, we may consider the load in the quasistatic limit and write
V ( z0 )  Z L ( ) I  ( z0 ) , (7.117)
where ZL() is the (generally complex) impedance of the load. Taking V(z,t) and I(z,t) in the form
similar to Eqs. (61) and (62), and writing the two Kirchhoff’s circuit laws for the point z = z0, we get for
the reflection coefficient a result similar to Eq. (68):

Z L ( )  ZW
R . (7.118)
Z L ( )  ZW
This formula shows that for the perfect matching (i.e. the total wave absorption in the load), the load’s
impedance ZL() should be real and equal to ZW – but not necessarily to Z.
As an example, let us consider one of the simplest (and most practically important) transmission
lines: the coaxial cable (Fig. 20).49

  ,    ba

0 a
Fig. 7. 20. The cross-section of a coaxial cable
with (possibly, dispersive) dielectric filling.

For this geometry, we already know the expressions for both C0 and L0,50 though they have to be
modified for the account of arbitrary dielectric and magnetic constants, and the magnetic field’s non-
penetration into the conductors. As a result of this (elementary) modification, we get the formulas,

48 The ability of TEM lines to have such a small cross-section is another important practical advantage.
49 Itwas invented by the same O. Heaviside in 1880.
50 See, respectively, Eqs. (2.49) and (5.79).

Chapter 7 Page 35 of 70
Essential Graduate Physics EM: Classical Electrodynamics

2  Coaxial
C0  , L0  ln(b / a) , (7.119) cable’s
ln(b / a) 2 C0 and L0

illustrating that the universal relationship (114) is indeed valid. For the cable’s impedance (115), Eqs.
(119) yield a geometry-dependent value
1/ 2
 ln(b / a) ln(b / a)
ZW   Z Z. (7.120)
  2 2
For the standard TV antenna cables (such as RG-6/U, with b/a ~ 3, /0  2.2), ZW = 75 , while for
most computer component connections, coaxial cables with ZW = 50  (such as RG-58/U) are
prescribed by electronic engineering standards. Such cables are broadly used for the transmission of
electromagnetic waves with frequencies up to 1 GHz over distances of a few km, and up to ~20 GHz on
the tabletop scale (a few meters), limited by wave attenuation – see Sec. 9 below.
Moreover, the following two facts enable a wide application, in electrical engineering and
physical experiment, of coaxial-cable-like systems. First, as Eq. (5.78) shows, in a cable with a << b,
most energy of the wave is localized near the internal conductor. Second, the theory to be discussed in
the next section shows that excitation of other (H- and E-) waves in the cable is impossible until the
wavelength  becomes smaller than ~(a + b). As a result, the TEM mode propagation in a cable with a
<< b < / is not much affected even if the internal conductor is not straight, but bent – for example,
into a helix – see, e.g., Fig. 21.

Fig. 7.21. A typical traveling-wave tube:

(1) electron gun, (2) ac input, (3) beam-
focusing magnets, (4) wave attenuator,
(5) helix coil, (6) ac output, (7) vacuum
tube, (8) electron collector. Adapted from
https://en.wikipedia.org/wiki/Traveling-
wave_tube under the Creative Commons
BY-SA 3.0 license.

In such a system, called the traveling-wave tube (TWT), a quasi-TEM wave propagates with
velocity v  c along the helix’s length, so the velocity’s component along the cable’s axis may be made
close to the velocity u << c of the electron beam moving ballistically along the tube’s axis, enabling
their effective interaction, and as a result, a length-accumulating amplification of the wave.51
Another important example of a TEM transmission line is a set of two parallel wires. In the form
of twisted pairs,52 they allow communications, in particular long-range telephone and DSL Internet

Despite the current prevalence of semiconductor devices in electronics, TWTs are still used in satellite
51

TV and radio systems, because they may work at very high microwave power – e.g., up to 200W at 20
GHz and pulsed 50W at 200 GHz. Very unfortunately, in this course, I will not have time/space to discuss
even the (rather elegant) basic theory of such devices. The reader interested in this field may be referred, for
example, to the detailed monograph by J. Whitaker, Power Vacuum Tubes Handbook, 3rd ed., CRC Press, 2017.
52 Such twisting, around the line’s direction axis, reduces the crosstalk between adjacent lines, and the parasitic
radiation at their bends.

Chapter 7 Page 36 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

connections, at frequencies up to a few hundred kHz, as well as relatively short, multi-line Ethernet and
TV cables at frequencies up to ~ 1 GHz, limited mostly by the mutual interference (“crosstalk”) between
the individual lines of the same cable, and the unintentional radiation of the wave into the environment.

7.6. Waveguides: H and E waves

Let us now return to Eqs. (100) and explore the H- and E-waves – with, respectively, either Hz or
Ez different from zero. At the first sight, they may seem more complex. However, Eqs. (101), which
determine the distribution of these longitudinal components over the cross-section, are just the 2D
Helmholtz equations for scalar functions. For simple cross-section geometries, they may be readily
solved using the methods discussed for the Laplace equation in Chapter 2, in particular the variable
separation. After the solution of such an equation has been found, the transverse components of the
fields may be calculated by differentiation, using the simple formulas,
i i  k 
Et  k z  t E z  kZ ( n z   t H z ), Ht  k z  t H z  Z (n z   t E z ), (7.121)
k t2 k t2
which follow from the first line of Eqs. (100).53
In comparison with the boundary problems of electro- and magnetostatics, the only conceptually
new feature of Eqs. (101) is that they form the so-called eigenproblems, with typically many solutions
(eigenfunctions), each describing a specific wave mode, and corresponding to a specific eigenvalue of
the parameter kt. The good news here is that these values of kt are determined by this 2D boundary
problem and hence do not depend on kz. As a result, the dispersion law (kz) of any mode, which
follows from the last form of Eq. (102),
1/ 2
 k 2  k t2 
 
Universal
1/ 2
dispersion    z 
  v 2 k z2   c2 , (7.122)
 
relation

is functionally similar for all modes. It is also similar to that of plane waves in plasma (see Eq. (38), Fig.
6, and their discussion in Sec. 2), with the only difference that the speed in light c is generally replaced
with v = 1/()1/2 – the speed of the plane or TEM waves in the medium filling the waveguide, and that
p is replaced with the so-called cutoff frequency
 c  vk t , (7.123)
specific for each mode. (As Eq. (101) implies, and as we will see from several examples below, kt has
the order of 1/a, where a is the characteristic dimension of the waveguide’s cross-section, so the critical
value of the free-space wavelength   2c/ is of the order of a.) Below the cutoff frequency of each
particular mode, this wave cannot propagate in the waveguide.54 As a result, the modes with the lowest

53 For the derivation of Eqs. (121), one of these two linear equations should be first vector-multiplied by nz. Note
also that this approach could not be used to analyze the TEM waves, because for them kt = 0, Ez = 0, Hz = 0, and
Eqs. (121) yield uncertainty.
54 An interesting twist in the ideas of electromagnetic metamaterials (mentioned in Sec. 5 above) is the so-called

-near-zero materials, designed to have the effective product  much lower than 00 within certain frequency
ranges. Since at these frequencies, the speed v (4) becomes much lower than c, the cutoff frequency (123)

Chapter 7 Page 37 of 70
Essential Graduate Physics EM: Classical Electrodynamics

values of c present special practical interest, because the choice of the signal frequency  between the
two lowest values of the cutoff frequency (123) guarantees that the waves propagate in the form of only
one mode, with the lowest kt. Such a choice enables engineers to simplify the excitation of the desired
mode by wave generators and to avoid the unintentional transfer of electromagnetic wave energy to
undesirable modes by (virtually unavoidable) small inhomogeneities of the system.
The boundary conditions for the Helmholtz equations (101) depend on the propagating wave
type. For the E-modes, with Hz = 0 but Ez  0, the condition E = 0 immediately gives
Ez C  0, (7.124)
where C is the inner contour limiting the conducting wall’s cross-section. For the H-modes, with Ez = 0
but Hz  0, the boundary condition is slightly less obvious and may be obtained using, for example, the
second equation of the system (100), vector-multiplied by nz. Indeed, for the component normal to the
conductor surface, the result of such multiplication is

ik z (H t ) n  i
k
n z  E t n  H z . (7.125)
Z n
But the first term on the left-hand side of this relation must be zero on the wall surface, because of the
second of Eqs. (104), while according to the first of Eqs. (104), the vector Et in the second term cannot
have a component tangential to the wall. As a result, the vector product in that term cannot have a
normal component, so the term should equal zero as well, and Eq. (125) is reduced to
H z
C  0. (7.126)
n
Let us see how all this machinery works for a simple but practically important case of a metallic-
wall waveguide with a rectangular cross-section – see Fig. 22
y

b
Ht
Fig. 7.22. A rectangular waveguide, and the
E transverse field distribution in its
0 a fundamental mode H10 (schematically).
x

In the natural Cartesian coordinates shown in this figure, both Eqs. (101) take the simple form
 2 2   E , for E - modes,
 2  2  k t2  f  0, where f   z (7.127)
 x y   H z , for H - modes.
From Chapter 2, we know that the most effective way of solving such equations in a rectangular region
is the variable separation, in which the general solution is represented as a sum of partial solutions of the
type

virtually vanishes. As a result, the waves may “tunnel” through very narrow sections of metallic waveguides filled
with such materials – see, e.g., M. Silveirinha and N. Engheta, Phys. Rev. Lett. 97, 157403 (2006).

Chapter 7 Page 38 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

f  X ( x)Y ( y ) . (7.128)

Plugging this expression into Eq. (127), and dividing each term by XY, we get the equation,
1 d 2 X 1 d 2Y
  k t2  0 , (7.129)
X dx 2 Y dy 2
which should be satisfied for all values of x and y within the waveguide’s interior. This is only possible
if each term of the sum equals a constant. Taking the X-term and Y-term constants in the form (–kx2) and
(–ky2), respectfully, and solving the corresponding ordinary differential equations,55 for the
eigenfunction (128) we get

f  c x cos k x x  s x sin k x x  c y cos k y y  s y sin k y y , with k x2  k y2  k t2 , (7.130)

where the constants c and s should be found from the boundary conditions. Here the difference between
the H-modes and E-modes kicks in.
For the H-modes, Eq. (130) is valid for Hz, and we should use the boundary condition (126) on
all metallic walls of the waveguide, i.e. at x = 0 and a; and y = 0 and b – see Fig. 22. As a result, we get
very simple expressions for eigenfunctions and eigenvalues:
 nx  my
H z nm  H l cos cos , (7.131)
a b
1/ 2
n m  n  2  m  2 
kx 
a
, ky 
b
, so that (k t ) nm  k  k  2
x y
2 1/ 2
        , (7.132)
 a   b  
where Hl is the longitudinal field’s amplitude, and n and m are two integer numbers – each of them
arbitrary besides that they cannot be equal to zero simultaneously.56 Assuming, just for certainty, that a
 b (as shown in Fig. 22), we see that the lowest eigenvalue of kt and hence the lowest cutoff frequency
(123) are achieved for the so-called H10 mode with n = 1 and m = 0, and hence with
Fundamental 
mode’s (k t )10  , (7.133)
cutoff a
thus confirming our prior estimate of kt.
Depending on the a/b ratio, the second-lowest kt (and hence c) belongs to either the H11 mode
with n = 1 and m = 1:
1/ 2
 1 1 
1/ 2
  a 2 
(k t )11    2  2   1     (k t )10 , (7.134)
a b    b  
or to the H20 mode with n = 2 and m = 0:

55 Let me hope that the solution of equations of the type d2X/dx2 + kx2X = 0 does not present any problem for the
reader, at least due to their prior experience with problems such as standing waves on a guitar string,
wavefunctions in a flat 1D quantum well, or (with the replacement x  t) a classical harmonic oscillator.
56 Otherwise, the function H (x,y) would be constant, so, according to Eq. (121), the transverse components of the
z
electric and magnetic field would equal zero. As a result, as the last two lines of Eqs. (100) show, the whole field
would be zero for any kz  0.

Chapter 7 Page 39 of 70
Essential Graduate Physics EM: Classical Electrodynamics

2
(k t ) 20   2(k t )10 . (7.135)
a
These values become equal at a/b = 3  1.7; in practical waveguides, the a/b ratio is not too far from
this value. For example, in the standard X-band (~10-GHz) waveguide WR90, a  2.3 cm (fc  c/2 
6.5 GHz), and b  1.0 cm.
Now let us have a look at the alternative E-modes. For them, we still should use the general
solution (130) with f = Ez, but now with the boundary condition (124). This gives us the eigenfunctions
 nx  my
E z nm  El sin sin , (7.136)
a b
and the same eigenvalue spectrum (132) as for the H modes. However, now neither n nor m can be equal
to zero; otherwise, Eq. (136) would give the trivial solution Ez(x,y) = 0. Hence the lowest cutoff
frequency of TM waves is achieved at the so-called E11 mode with n =1, m = 1, and with the eigenvalue
given by Eq. (134), always higher than (kt)10.
Thus the fundamental H10 mode is certainly the most important wave in rectangular waveguides;
let us have a better look at this field distribution. Plugging the corresponding solution (131) with n = 1
and m = 0 into the general relation (121), we easily get
kza x
( H x )10  i H l sin , ( H y )10  0, (7.137)
 a
ka x
( E x )10  0, ( E y )10  i . ZH l sin (7.138)
a 
This field distribution is (schematically) shown in Fig. 22. Neither of the fields depends on the
coordinate y – the feature very convenient, in particular, for microwave experiments with small samples.
The electric field has only one (in Fig. 22, vertical) component that vanishes at the side walls and
reaches its maximum at the waveguide’s center; its field lines are straight, starting and ending on wall
surface charges (whose distribution propagates along the waveguide together with the wave). In
contrast, the magnetic field has two non-zero components (Hx and Hz), and its field lines are shaped as
horizontal loops wrapped around the electric field maxima.
An important question is whether the H10 wave may be usefully characterized by a unique
impedance introduced similarly to ZW of the TEM modes – see Eq. (115). The answer is not, because the
main value of ZW is a convenient description of the impedance matching of a transmission line with a
lumped load – see Fig. 19 and Eq. (118). As was discussed above, such a simple description is possible
(i.e., does not depend on the exact geometry of the connection) only if both dimensions of the line’s
cross-section are much less than . But for the H10 wave (and more generally, any non-TEM mode) this
is impossible – see, e.g., Eq. (129): its lowest frequency corresponds to the TEM wavelength max =
2/(kt)min = 2/(kt)10 = 2a. (The reader is challenged to find a simple interpretation of this equality.)
Now let us consider metallic-wall waveguides with a round cross-section (Fig. 23a). In this
single-connected geometry, the TEM waves are impossible again, while for the analysis of H-modes and
E-modes, the polar coordinates {, } are most natural. In these coordinates, the 2D Helmholtz equation
(101) takes the following form:

Chapter 7 Page 40 of 70
Essential Graduate Physics EM: Classical Electrodynamics

 1     1 2   E , for E - modes,
     2  k t2  f  0, where f   z (7.139)
       
2
  H z , for H - modes.
Separating the variables as f = R()F(), we get

1 d  dR  1 d 2F
   2  k t2  0 . (7.140)
R d  d   F d
2

But this is exactly the Eq. (2.127) that was studied in Sec. 2.7 in the context of electrostatics, just with a
replacement of notation:   kt. So we already know that to have 2-periodic functions F() and finite
values R(0) (which are evidently necessary for our current case – see Fig. 23a), the general solution
must have the form given by Eq. (2.136), i.e. the eigenfunctions are expressed via integer-order Bessel
functions of the first kind:
f nm  J n k nm  c n cos n  s n sin n   const  J n k nm   cos n   0  , (7.141)
with the eigenvalues knm of the transverse wave number kt to be determined from appropriate boundary
conditions, and an arbitrary constant 0.

(a)   ,  (b)

R   , 
R
0 0
Fig. 7.23. (a) Metallic and (b) dielectric
, waveguides with circular cross-sections.

As for the rectangular waveguide, let us start with the H-modes (f = Hz). Then the boundary
condition on the wall surface ( = R) is given by Eq. (126), which, for the solution (141), takes the form
d
J n ( )  0, where   kR . (7.142)
d
This means that the eigenvalues of Eq. (139) are
' nm
k t  k nm 
, (7.143)
R
where ’nm is the mth zero of the function dJn()/d. Approximate values of these zeros for several
lowest n and m may be read out from Fig. 2.18; their more accurate values are given in Table 1 below.

Table 7.1. Zeros ’nm of the function dJn()/d for a few lowest
values of the Bessel function’s index n and the root’s number m.
m=1 2 3
n=0 3.83171 7.015587 10.1735
1 1.84118 5.33144 8.53632
2 3.05424 6.70613 9.96947
3 4.20119 8.01524 11.34592

Chapter 7 Page 41 of 70
Essential Graduate Physics EM: Classical Electrodynamics

The table shows, in particular, that the lowest of the zeros is ’11  1.84.57 Thus, perhaps a bit
counter-intuitively, the fundamental mode, providing the lowest cutoff frequency c = vknm, is H11,
corresponding to n = 1 rather than n = 0:
 
H z  H l J 1  11'  cos   0  . (7.144)
 R
It has the transverse wave number is kt = k11 = ’11/R  1.84/R, and hence the cutoff frequency
corresponding to the TEM wavelength max = 2/k11  3.41 R. Thus the ratio of max to the waveguide’s
diameter 2R is about 1.7, i.e. is close to the ratio max/a = 2 for the rectangular waveguide. The origin of
this proximity is clear from Fig. 24, which shows the transverse field distribution in the H11 mode. (It
may be readily calculated from Eqs. (121) with Ez = 0 and Hz given by Eq. (144).)

H Fig. 7.24. Transverse field components in the

fundamental H11 mode of a metallic, circular
waveguide (schematically).

One can see that the field structure is actually very similar to that of the fundamental mode in the
rectangular waveguide, shown in Fig. 22, despite the different nomenclature (which is due to the
different coordinate system used for the solution). However, note the arbitrary constant angle 0,
indicating that in circular waveguides, the transverse field’s polarization is arbitrary. For some practical
applications, such degeneracy of these “quasi-linearly-polarized” waves creates problems; some of them
may be avoided by using waves with circular polarization.
As Table 1 shows, the next lowest H-mode is H21, for which kt = k21 = ’21/R  3.05/R, almost
twice larger than that of the fundamental mode, and only then comes the first mode with no angular
dependence of any field, H01, with kt = k01 = ’01/R  3.83/R,58 followed by several angle-dependent
modes: H31, H12, etc.
For the E modes, we may still use Eq. (141) (with f = Ez), but with the boundary condition (124)
at  = R. This gives the following equation for the problem eigenvalues:
 nm
J n k nm R   0, i.e. k nm  , (7.145)
R
where nm is the mth zero of function Jn() – see Table 2.1. That table shows that the lowest kt is equal to
01/R  2.405/R. Hence the corresponding mode (E01), with no angular dependence of its fields, e.g.

57 Mathematically, the lowest root of Eq. (142) with n = 0 equals 0. However, it would yield k = 0 and hence a
constant field Hz, which, according to the first of Eqs. (121), would give a vanishing electric field.
58 The electric field lines in the H
01 mode (as well as all higher H0m modes) are directed straight from the
symmetry axis to the walls, reminding those of the TEM waves in the coaxial cable. Due to this property, these
modes provide, at  >> c, much lower energy losses (see Sec. 9 below) than the fundamental H11 mode, and are
sometimes used in practice, despite the inconvenience of working in the multimode frequency range.

Chapter 7 Page 42 of 70
Essential Graduate Physics EM: Classical Electrodynamics

 
E z  El J 0   01  , (7.146)
 R
has the second-lowest cutoff frequency, ~30% higher than that of the fundamental mode H11.
Finally, let us discuss one more topic of general importance – the number N of electromagnetic
modes that may propagate in a waveguide within a certain range of relatively large frequencies  >> c.
It is easy to calculate for a rectangular waveguide, with its simple expressions (132) for the eigenvalues
of {kx, ky}. Indeed, these expressions describe a rectangular mesh on the [kx, ky] plane, so each point
corresponds to the plane area Ak = (/a)(/b), and the number of modes in a large k-plane area Ak >>
Ak is N = Ak/Ak = abAk/2 = AAk/2, where A is the waveguide’s cross-section area.59 However, it is
frequently more convenient to discuss transverse wave vectors kt of arbitrary direction, i.e. with an
arbitrary sign of their components kx and ky. Taking into account that the opposite values of each
component actually give the same wave, the actual number of different modes of each type (E- or H-) is
a factor of 22  4 lower than was calculated above. This means that the number of modes of both types is
Ak A
N 2 . (7.147)
(2 ) 2
Let me leave it for the reader to find hand-waving (but convincing :-) arguments that this mode
counting rule is valid for waveguides with cross-sections of any shape, and any boundary conditions on
the walls, provided that N >> 1.

7.7. Dielectric waveguides, optical fibers, and paraxial beams

Now let us discuss electromagnetic wave propagation in dielectric waveguides. The simplest,
step-index waveguide (see Figs. 23b and 25) consists of an inner core and an outer shell (in the optical
fiber technology lingo, called cladding) with a higher wave propagation speed, i.e. a lower index of
refraction:
v   v  , i.e. n   n  , k   k  ,          . (7.148)

at the same frequency. (In most cases the difference is achieved due to that in the electric permittivity, +
< -, while magnetically both materials are virtually passive: -  +  0, so their refraction indices n,
defined by Eq. (84), are very close to (/0)1/2; I will limit my discussion to this approximation.)
The basic idea of the waveguide’s operation may be readily understood in the limit when the
wavelength  is much smaller than the characteristic size R of the core’s cross-section. In this
“geometric-optics” limit, at distances of the order of  from the core-cladding interface, which provides
the wave reflection, we can neglect the interface’s curvature and approximate its geometry with a plane.
As we know from Sec. 4, if the angle  of the wave’s incidence on such a plane interface is larger than
the critical value c specified by Eq. (85), the wave is totally reflected. As a result, the waves launched
into the fiber core at such “grazing” angles, propagate inside the core, being repeatedly reflected from
the cladding – see Fig. 25.

59This formula ignores the fact that, according to the above analysis, some modes (with n = 0 and m = 0 for the H
modes, and n = 0 or m = 0 for the E modes) are forbidden. However, for N >> 1, the associated corrections of Eq.
(147) are negligible.

Chapter 7 Page 43 of 70
Essential Graduate Physics EM: Classical Electrodynamics

“cladding”   ,  

“core”   ,   
Fig. 7.25. Wave propagation in
 a thick optical fiber at  > c.

The most important type of dielectric waveguides is optical fibers.60 Due to a heroic
technological effort over three decades starting from the mid-1960s, the attenuation of such fibers has
been decreased from values of the order of 20 db/km (typical for a window glass) to the fantastically
low values of ~0.2 db/km (meaning virtually perfect transparency of 10-km-long fiber segments!),
combined with the extremely low plane-wave (“chromatic”) dispersion below 10 ps/kmnm.61 In
conjunction with the development of inexpensive erbium-based quantum amplifiers, this breakthrough
has enabled inter-city and inter-continental (undersea), broadband62 optical cables, which are the
backbone of all modern telecommunication infrastructure.
The only bad news is that these breakthroughs were achieved for just one kind of materials
(silica-based glasses)63 within a very narrow range of their chemical composition. As a result, the
dielectric constants   /0 of the cladding and core of practical optical fibers are both close to 2.2
(giving n  1.5) and hence very close to each other, so the relative difference of the refraction indices,
n  n  1 / 2   1 / 2      (7.149)
   ,
n  1 / 2 2 
is typically below 0.5%. This factor limits the fiber bandwidth. Indeed, let us use the geometric-optics
picture to calculate the number of quasi-plane-wave modes that may propagate in the fiber. For the
complementary angle (Fig. 25)

  , so that sin   cos  , (7.150)
2
Eq. (85) gives the following propagation condition:

60 For a comprehensive discussion of this vital technology see, e.g., A. Yariv and P. Yeh, Photonics, 6th ed.,
Oxford U. Press, 2007.
61 Both these parameters have their best values not in the visible light range (with wavelengths from 380 to 740
nm), but in the near-infrared, with the attenuation lowest between approximately 1,500 and 1,630 nm. As a result,
most modern communication systems use two spectral windows – the so-called C-band (1,530-1,565 nm) and L-
band (1,570-1,610 nm) within that range.
62 Each of the spectral bands mentioned above, at a typical signal-to-noise ratio S/N > 105, corresponds to the

Shannon bandwidth f log2(S/N) exceeding 1014 bits per second, some five orders of magnitude (!) higher than
that of a modern Ethernet cable. The practically usable bandwidth of each fiber is somewhat lower, but a typical
optical cable, with several fibers in parallel, has a proportionately higher aggregate bandwidth. A relatively recent
(circa 2017) example is the C-band transatlantic (6,600-km-long) cable Marea, with eight fiber pairs and an
aggregate useable bandwidth of 160 terabits per second. By 2024, by using additional spectrum portions, the
bandwidth of similar cables has been pushed beyond 400 terabits per second – so far, in laboratory conditions.
63 The silica-based fibers were developed in 1966 by an industrial research group led by Charles Kao (who shared
the 2009 Nobel Prize in physics), but the very idea of using optical fibers for long-range communications may be
traced back at least to the 1963 work by Jun-ichi Nishizawa – who also invented semiconductor lasers.

Chapter 7 Page 44 of 70
Essential Graduate Physics EM: Classical Electrodynamics

n
cos    1  . (7.151)
n
In the limit  << 1, when the incidence angles  > c of all propagating waves are very close to /2, and
hence the complementary angles are  small, we may keep only two first terms in the Taylor expansion
of the left-hand side of Eq. (151) and get
max
2
 2 . (7.152)

(Even for the higher-end value  = 0.005, this critical angle is only ~0.1 radian, i.e. is close to 5.) Due
to this smallness, we may approximate the maximum transverse component of the wave vector as

(k t ) max  k (sin  ) max  kmax  2k , (7.153)

and use Eq. (147) to calculate the number N of propagating modes:

N 2
R k   kR   .
2 2 2
max 2 (7.154)
2 2
For typical values k = 0.73107 m-1 (corresponding to the free-space wavelength 0 = n = 2n/k  1.3
m), R = 25 m, and  = 0.005, this formula gives N  150.
Now we can calculate the geometric dispersion of such a fiber, i.e. the difference in the mode
propagation speed, which is commonly characterized in terms of the difference between the wave delay
times (traditionally measured in picoseconds per kilometer) of the fastest and slowest modes. Within the
geometric optics approximation, the difference in time delays of the fastest mode (with kz = k) and the
slowest mode (with kz = k sinc) at distance l is

 l  k l l l l  n  l
t      z   k z  1  sin  c   1     . (7.155)
 vz     v v  n   v
For the example considered above, the TEM wave’s speed in the glass, v = c/n  2108 m/s, and the
geometric dispersion t/l is close to 25 ps/m, i.e. 25,000 ps/km. (This means, for example, that a 1-ns
pulse, being distributed between the modes, would spread to a ~25-ns pulse after passing a just 1-km
fiber segment.) This result should be compared with the chromatic dispersion mentioned above, below
10 ps/kmnm, which gives dt/l is of the order of only 1,000 ps/km in the whole communication band d
~ 100 nm. Due to this high geometric dispersion, such relatively thick (2R ~ 50 nm) multi-mode fibers
are used for the transfer of signals over only short distances below ~ 100 m. (As compensation, they
may carry relatively large power, beyond 10 mW, without being damaged by the field.)
Long-range telecommunications are based on single-mode fibers, with thin cores (typically with
diameters 2R ~ 5 m, i. e. of the order of /1/2). For such structures, Eq. (154) yields N ~ 1, but in this
case, the geometric optics approximation is not quantitatively valid, and for the fiber analysis, we should
get back to the Maxwell equations. In particular, this analysis should take into explicit account the
evanescent wave in the cladding, because its penetration depth may be comparable with R.64

64 The following quantitative analysis of the single-mode fibers is very valuable – both for practice and as a very
good example of Maxwell equations’ solution. However, I have to confess that its results will not be used in the
following parts of the course. So, if the reader is not interested in this topic, they may safely jump to the text

Chapter 7 Page 45 of 70
Essential Graduate Physics EM: Classical Electrodynamics

Since the cross-section of an optical fiber lacks metallic walls, the Maxwell equations describing
them cannot be exactly satisfied with either TEM-wave, or H-mode, or E-mode solutions. Instead, the
fibers can carry the so-called HE and EH modes, with both vectors H and E having longitudinal
components simultaneously. In such modes, both Ez and Hz inside the core (  R) have a form similar
to Eq. (141):
f   f l J n k t   cos n   0 , where k t2  k 2  k z2  0, and k 2   2     , (7.156)

where the constant angles 0 may be different for each field. On the other hand, for the evanescent wave
in the cladding, we may rewrite Eqs. (101) as

 2

  t2 f   0, where  t2  k z2  k 2  0, and k 2   2     . (7.157)

Figure 26 illustrates these relations between kt, t, kz, and k; note that the following sum, Universal
relation
k     (     )  0  2k  ,
t
2
t
2 2 2
(7.158) between
kt and t

is fixed (at a given frequency) and, for typical fibers, is very small (<< k2). In particular, Fig. 26 shows
that neither kt nor t can be larger than  [(- – +)0]1/2 = (2)1/2 k. This means that the depth  = 1/t of
the wave penetration into the cladding is at least 1/k(2)1/2 = /2(2)1/2 >> /2. This is why the
cladding layers in practical optical fibers are made as thick as ~50 m, so only a negligibly small tail of
this evanescent wave field reaches their outer surfaces.

k 2  k 2   2 (     )  0

 t2 k t2
Fig. 7.26. The relation between the transverse
exponents kt and t for waves in optical fibers.
k 2 k z2 k 2 k2

In the polar coordinates, Eq. (157) becomes

 1     1 2 
     2   t2  f   0 , (7.159)
       
2

- the equation to be compared with Eq. (139) for the circular metallic-wall waveguide. From Sec. 2.7,
we know that the eigenfunctions of Eq. (159) are the products of the sine and cosine functions of n by
a linear combination of the modified Bessel functions In and Kn, shown in Fig. 2.22, now of the
argument t. The fields have to vanish at   , so only the latter functions (of the second kind) can
participate in the solution:
f   K n ( t  ) cos n   0  (7.160)

Now we have to reconcile Eqs. (156) and (160), using the boundary conditions at  = R for both
longitudinal and transverse components of both fields, with the latter components first calculated using

following Eq. (181). (I believe that the discussion of the angular momentum of electromagnetic radiation, starting
at that point, is compulsory for every professional physicist.)

Chapter 7 Page 46 of 70
Essential Graduate Physics EM: Classical Electrodynamics

Eqs. (121). Such a conceptually simple, but a bit bulky calculation (which I am leaving for the reader’s
exercise), yields a system of two homogeneous linear equations for the complex amplitudes El and Hl,
which are compatible if

 k 2 J n' k 2 K n'  1 J n' 1 K n'  n 2  k 2 k 2  1 1 

      2  2  2  2  2  , (7.161)
 k t J n  t K n  k t J n  t K n  R  k t  t  k t  t 
where the prime signs denote the derivatives of each Bessel function over its full argument: kt for Jn,
and t for Kn, and the functions and their derivatives are taken at  = R.
For any given frequency , the system of equations (158) and (161) determines the values of kt
and t, and hence kz. Actually, for any n > 0, this system provides two different solutions: one
corresponding to the so-called HE wave, with a larger ratio of Ez/Hz, and the EH wave, with a smaller
value of that ratio. For angular-symmetric modes with n = 0 (for whom we might naively expect the
lowest cutoff frequency), the equations may be satisfied by the fields having just one non-zero
longitudinal component (either Ez or Hz), so the HE wave are the usual E-modes, while the EH modes
are the H-waves. For the H-modes, the characteristic equation is reduced to the requirement that the
expression in the second parentheses on the left-hand side of Eq. (161) is equal to zero. Using the Bessel
function identities J0’ = –J1 and K0’ = –K1, this equation may be rewritten in a simpler form:
1 J 1 (k t R) 1 K 1 ( t R)
 . (7.162)
k t J 0 (k t R)  t K 0 ( t R)
Using the universal relation between kt and t given by Eq. (158), we may plot both sides of Eq.
(162) as functions of the same argument, say,   ktR – see Fig. 27.

LHS
 01  11  02  12  03
0
Fig. 7.27. Two sides of the characteristic
equation (162), plotted as functions of ktR,
RHS for two values of its dimensionless
parameter: V = 8 (blue line) and V = 3 (red
line). Note that according to Eq. (158), the
argument of the functions K0 and K1 is
3 tR = [V 2 – (ktR)2]1/2  (V 2 – 2)1/2.
0 5 10
  kt R

The right-hand side of Eq. (162) depends not only on , but also on the dimensionless parameter
V defined as the normalized right-hand side of Eq. (158):
V 2   2 (     )  0 R 2  2 k 2 R 2 . (7.163)
(According to Eq. (154), if V >> 1, it gives twice the number N of the fiber modes – the conclusion
confirmed by Fig. 27, taking into account that it describes only the H-modes.) Since the ratio K1/K0 is
positive for all values of the functions’ argument (see, e.g., the right panel of Fig. 2.22), the right-hand

Chapter 7 Page 47 of 70
Essential Graduate Physics EM: Classical Electrodynamics

side of Eq. (162) is always negative, so the equation may have solutions only in the intervals where the
ratio J1/J0 is negative, i.e. at
 01  k t R  11 ,  02  k t R  12 ,... , (7.164)

where nm is the m-th zero of the function Jn() – see Table 2.1. The right-hand side of the characteristic
equation (162) diverges at tR  0, i.e. at ktR  V, so no solutions are possible if V is below the critical
value Vc = 01  2.405. At this cutoff point, Eq. (163) yields k  01/R(2)1/2. Hence, the cutoff
frequency of the lowest H mode corresponds to the TEM wavelength
2R
max  2 1 / 2  3.7 R1 / 2 . (7.165)
 01
For typical parameters  = 0.005 and R = 2.5 m, this result yields max ~ 0.65 m, corresponding to the
free-space wavelength 0 ~ 1 m. A similar analysis of the first parentheses on the left-hand side of Eq.
(161) shows that at   0, the cutoff frequency for the E modes is similar.
This situation may look exactly like that in metallic-wall waveguides, with no waves possible at
frequencies below c, but this is not so. The basic reason for the difference is that in the metallic
waveguides, the approach to c results in the divergence of the longitudinal wavelength z  2/kz. On
the other hand, in dielectric waveguides, the approach leaves z finite (kz  k+). Due to this difference, a
certain linear superposition of HE and EH waves with n = 1 can propagate at frequencies well below the
cutoff frequency for n = 0, which we have just calculated.65 This mode, in the limit +  - (i.e.  << 1)
allows a very interesting and simple description using the Cartesian (rather than polar) components of
the fields, but still expressed as functions of the polar coordinates  and . The reason is that this mode
is very close to a linearly polarized TEM wave. (Due to this reason, this mode is referred to as LP01.)
Let us select the x-axis parallel to the transverse component of the magnetic field vector at  = 0,
so Ex=0 = 0, but Ey=0  0, and Hx=0 0, but Hy=0 = 0. The only suitable solutions of the 2D
Helmholtz equation (that should be obeyed not only by the z-components of the fields but also their x-
and y-components) are proportional to J0(kt), with zero coefficients for Ex and Hy:

E x  0, E y  E 0 J 0 (k t  ), H x  H 0 J 0 (k t  ), H y  0, for   R . (7.166) LP01 mode

Now we can use the last two equations of Eqs. (100) to calculate the longitudinal components of the
fields:
1 E y k 1 H x k
Ez   i t E 0 J 1 (k t  ) sin  , Hz   i t H 0 J 1 (k t  ) cos  , (7.167)
 ik z y kz  ik z x kz

where I have used the following mathematical identities: J0 = –J1’, /x = x/ = cos, and /y = y/
= sin. As a sanity check, we see that the longitudinal component or each field is a (legitimate!)
eigenfunction of the type (141), with n = 1. Note also that if kt << kz (this relation is always true if  <<
1 – see either Eq. (158) or Fig. 26), the longitudinal components of the fields are much smaller than their

65This fact becomes less surprising if we recall that in the circular metallic waveguide, discussed in Sec. 6, the
fundamental mode (H11, see Fig. 23) also corresponded to n = 1 rather than n = 0.

Chapter 7 Page 48 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

transverse counterparts, so the wave is indeed very close to the TEM one. Because of that, the ratio of
the electric and magnetic field amplitudes is also close to that in the TEM wave: E0/H0  Z–  Z+.
Now to satisfy the boundary conditions at the core-to-cladding interface ( = R), we need to have
a similar angular dependence of these components at   R. The longitudinal components of the fields
are tangential to the interface and thus should be continuous. Using the solutions similar to Eq. (160)
with n = 1, we get
k t J 1 (k t R) k t J 1 (k t R)
E z  i E 0 K 1 ( t  ) sin  , H z  i H 0 K 1 ( t  ) cos  , for   R. (7.168)
k z K 1 ( t R) k z K 1 ( t R)
For the transverse components, we should require the continuity of the normal magnetic field Hn, for
our simple field structure equal to just Hxcos, of the tangential electric field E = Eysin, and of the
normal component of Dn = En = Eycos. Assuming that - = + = 0, and +  -–66 we can satisfy these
conditions with the following solutions:
J 0 (k t R) J (k R)
E x  0, E y  E 0 K 0 ( t  ), H x  0 t H 0 K 0 ( t  ), H y  0, for   R. (7.169)
K 0 ( t R) K 0 (k t R)
From here, we can calculate components from Ez and Hz, using the same approach as for   R:
1 E y  J (k R)
Ez   i t 0 t E 0 K 1 ( t  ) sin  ,
 ik z y k z K 0 ( t R)
(7.170)
1 H x  J (k R)
Hz   i t 0 t H 0 K 1 ( t  ) cos  , for   R.
 ik z x k z K 0 ( t R)
These relations provide the same functional dependence of the fields as Eqs. (167), i.e. the internal and
external fields are compatible, but their amplitudes at the interface coincide only if
LP01 mode: J 1 (k t R) K ( R)
characteristic kt  t 1 t . (7.171)
equation J 0 (k t R) K 0 ( t R )
This characteristic equation (which may be also derived from Eq. (161) with n = 1 in the limit 
 0) looks close to Eq. (162), but functionally is much different from it – see Fig. 28. Indeed, its right-
hand side is always positive, and the left-hand side tends to zero at ktR  0. As a result, Eq. (171) may
have a solution for arbitrary small values of the parameter V defined by Eq. (163), i.e. for arbitrary low
frequencies (large wavelengths). This is why this mode is used in practical single-mode fibers: there are
no other modes with wavelengths larger than the max given by Eq. (165), so they cannot be
unintentionally excited on small inhomogeneities of the fiber.
It is easy to use the Bessel function approximations by the first terms of the Taylor expansions
(2.132) and (2.157) to show that in the limit V  0, tR tends to zero much faster than ktR  V: tR 

66 This is the core assumption of this approximate theory, which accounts only for the most important effect of the
small difference of the dielectric constants + and –: the difference between (k+2 – kz2) = kt2 > 0 and (k-2 – kz2) = –
t2 < 0. For more discussion of the accuracy of this approximation and some exact results, the interested reader
may be referred either to the monograph by A. Snyder and D. Love, Optical Waveguide Theory, Chapman and
Hill, 1983, or to Chapter 3 and Appendix B in the monograph by Yariv and Yeh, that was cited above.

Chapter 7 Page 49 of 70
Essential Graduate Physics EM: Classical Electrodynamics

2exp{-1/ V } << V. This means that the scale c  1/t of the radial distribution of the LP01 wave’s fields
in the cladding becomes very large. In this limit, this mode may be interpreted as a virtually TEM wave
propagating in the cladding, just slightly deformed (and guided) by the fiber’s core. The drawback of
this feature is that it requires very thick cladding, to avoid energy losses in its outer (“buffer” and
“jacket”) layers that defend the silica layers from the elements and mechanical damages, but lack their
low optical absorption. Due to this reason, the core radius is usually selected so that the parameter V is
just slightly less than the critical value Vc = 01  2.4 for higher modes, thus ensuring the single-mode
operation.
10

RHS LHS
Fig. 7.28. Two sides of the
0
characteristic equation (171) for the
LP01 mode, plotted as a function of
ktR, for two values of the
dimensionless parameter: V = 8
(blue line) and V = 1 (red line).
 10
0 5 10
  kt R

In order to reduce the field spread into the cladding, the step-index fibers discussed above may
be replaced with graded-index fibers whose dielectric constant  is gradually and slowly decreased from
the center to the periphery.67 Keeping only the main two terms in the Taylor expansion of the function
() at  = 0, we may approximate such reduction as68

 (  )   (0)1   2 , (7.172)

where   –[(d2/d2)/2]=0 is a positive constant characterizing the fiber composition gradient.

Moreover, if this constant is sufficiently small ( << k2), the field distribution across the fiber’s cross-
section may be described by the same 2D Helmholtz equation (101), but with a space-dependent
transverse wave vector:69
 
 t2  k t2   f  0,
where (7.173)
k t (  )  k (  )  k z  k t (0)  k 0  , and k 0     0  0 .
2 2 2 2 2 2 2 2

Surprisingly for such an axially-symmetric problem, because of its special dependence on the radius,
this equation may be most readily solved in the Cartesian coordinates. Indeed, rewriting it as

67 Due to the difficulty of fabrication of graded-index fibers with wave attenuation below a few dm/km, they are
not used as broadly as the step-index ones.
68 For an axially-symmetric smooth function (), the first derivative d/d always vanishes at  = 0, so Eq. (172)

does not have a term linear in .

69 This approach is invalid at arbitrary (large)  because in the macroscopic Maxwell equations, (r) is under the

differentiation sign, and the exact Helmholtz-type equations for fields have additional terms containing .

Chapter 7 Page 50 of 70
Essential Graduate Physics EM: Classical Electrodynamics

 2 2 
 2  
 k t2 (0)  k 2 0  x 2  y 2  f  0 ,  (7.174)
 x y 2

and separating the variables as f = X(x)Y(y), we get

 k t2 (0)  k 2 0 x 2  y 2   0 ,
1 d 2 X 1 d 2Y
2
 2
(7.175)
X dx Y dy
so the functions X and Y obey similar differential equations, for example
d2X
dx 2

 k x2  k 2 0x 2 X  0,  (7.176)

with the separation constants satisfying the following condition:

k x2  k y2  k t2 (0)  k 2 0   k z2 . (7.177)

The ordinary differential equation (176) is well known in quantum mechanics, because the
stationary Schrödinger equation for one of the most important basic quantum systems, a 1D harmonic
oscillator, may be rewritten in this form. Its eigenvalues are very simple:

(k x2 ) n  k 0  1 / 2 2n  1, (k y2 ) m  k 0  1 / 2 2m  1, with n, m  0,1, 2,... , (7.178)

but the corresponding eigenfunctions Xn(x) and Ym(y) are expressed via not quite elementary functions –
the Hermite polynomials.70 For most practical purposes, however, the lowest eigenfunctions X0(x) and
Y0(y) are sufficient, because they correspond to the lowest kx,y, and hence the lowest

k t
2
0min  (k x2 ) 0  (k y2 ) 0  2k 0  1 / 2 , (7.179)

and the lowest cutoff frequency. As may be readily verified by the substitution to Eq. (176), the
eigenfunctions corresponding to this fundamental mode are also simple:
 k 0  1 / 2 x 2 
X 0 ( x)  const  exp  , (7.180)
 2 
and similarly for Y0(y), so the field distribution follows the Gaussian function
 k 0  1 / 2  2   2 
f 0 (  )  f 0 (0) exp   2 , with a  1 / k 0  ,
1/ 2 1/ 4
 f 0 ( 0 ) exp (7.181)
 2   2a 
where a >> 1/k(0) has the sense of the effective width of the field’s extension in the radial direction,
normal to the wave’s propagation axis z. This is the so-called Gaussian beam, very convenient for some
applications.
The Gaussian beam (181) is just one example of the so-called paraxial beams, which may be
represented as a result of modulation of a plane wave with a wave number k, by an axially-symmetric

70 See, e.g., QM Sec. 2.9.

Chapter 7 Page 51 of 70
Essential Graduate Physics EM: Classical Electrodynamics

envelope function f(), where   {x, y}, with a relatively large effective radius a >> 1/k.71 Such beams
give me a convenient opportunity to deliver on the promise made in Sec. 1: calculate the angular
momentum L of a circularly polarized wave propagating in free space, and prove its fundamental
relation to the wave’s energy U. Let us start from the calculation of U for a paraxial beam (with an
arbitrary, but spatially localized envelope f) of a circularly polarized wave, with the transverse electric
field components given by Eq. (19):

E x  E 0 f   cos , E y   E 0 f  sin , (7.182a)

where E0 is the real amplitude of the wave’s electric field at the propagation axis,   kz – t +  is its
total phase, and the two signs correspond to two possible directions of the circular polarization.72
According to Eq. (6), the corresponding transverse components of the magnetic field are
E E
H x   0 f  sin , H y  0 f   cos . (7.182b)
Z0 Z0
These expressions are sufficient to calculate the energy density (6.113) of the wave,73
 0 E x2  E y2   0 H x2  H y2   0 E 02 f 2  0 E 02 f 2
u    2
  0 E 02 f 2 , (7.183)
2 2 2 2Z 0

and hence the full energy (per unit length in the direction z of the wave’s propagation) of the beam:
 
U   ud r  2  ud  2 0 E
2 2
0 f
2
d . (7.184)
0 0

However, the transverse fields (182) are insufficient to calculate a non-zero average of L.
Indeed, following the angular moment’s definition in mechanics,74 L  rp, where p is the particle’s
(linear) momentum, we may use Eq. (6.115) for the electromagnetic field momentum’s density g in free
space, to define the field’s angular momentum’s density as
1 1 EM field’s
l  rg  2
r  S  2 r  E  H  . (7.185) angular
c c momentum

Let us use the familiar bac minus cab rule of the vector algebra75 to transform this expression to
1
l 2
Er  H   Hr  E  12 n z E z r  H   H z r  E  E t r  H   H t r  E. (7.186)
c c

71 Note that propagating in a uniform medium, i.e. outside of grade-index fibers or other focusing systems, such
beams gradually increase their width a due to diffraction – the effect to be analyzed in the next chapter.
72 For our task of calculating two quadratic forms of the fields (L and U), their real representation (182) is more
convenient than the complex-exponent one. However, for linear manipulations, the latter representation of the
circularly polarized waves, Et = E0f()Re[(nx  iny)exp{i}], Ht = (E0/Z0)f()Re[(inx + ny)exp{i}], is usually
more convenient, and is broadly used.
73 Note that, in contrast to a linearly-polarized wave (16), the energy density of a circularly-polarized wave does

not depend on the full phase  – in particular, on t at fixed z, or vice versa. This is natural because its field vectors
rotate (keeping their magnitude) rather than oscillate – see Fig. 3b.
74 See, e.g., CM Eq. (1.31).
75 See, e.g., MA Eq. (7.5).

Chapter 7 Page 52 of 70
Essential Graduate Physics EM: Classical Electrodynamics

If the field is purely transverse (Ez = Hz = 0), as it is in a strictly plane wave, the first square
brackets in the last expression vanish, while the second bracket gives an azimuthal component of l,
which oscillates in time and vanishes at its time averaging. (This is exactly the reason why I have not
tried to calculate L during our first discussion of the circularly polarized waves in Sec. 1.)
Fortunately, our discussion of optical fibers, in particular, the derivation of Eqs. (167), (168), and
(170) gives us a clear clue on how to resolve this paradox. If the envelope function f() differs from a
constant, the transverse wave components (182) alone do not satisfy the Maxwell equations (2b), which
necessitates longitudinal components Ez and Hz of the fields, with76
E z E E y H z H x H y
 x  ,   . (7.187)
z x y z x y
However, as these expressions show, if the envelope function f changes very slowly in the sense df/d ~
f/a << kf, the longitudinal components are very small and do not have a back effect on the transverse
components. Hence, the above calculation of U is still valid (asymptotically, at ka  0), and we may
still use Eqs. (182) on the right-hand side of Eqs. (187),
E z  f f  H z E 0  f f 
 E 0   cos  sin ,    sin  cos  , (7.188)
z  x x  z Z0  x x 
and integrate them over z as
 f f  E  f f 
E z  E 0    cos  sin dz  0    cos d   sin d 
 x x  k  x x 
(7.189a)
E  f f 
 0   sin  cos  .
k  x x 
Here the integration constant is taken for zero because no wave field component may have a time-
independent part. Integrating, absolutely similarly, the second of Eqs. (188), we get
E0  f f 
Hz    cos  sin  . (7.189b)
kZ 0  x y 
With the same approximation, we may calculate the longitudinal (z–) component of l, given by
the first term of Eq. (186), keeping only the dominating, transverse fields (182) in the scalar products:
l z  E z r  H t   H z r  E t   E z xH x  yH y   H z xE x  yE y  . (7.190)

Plugging in Eqs. (182) and (189), and taking into account that in free space, k = /c, and hence 1/Z0c2k
= 0/, we get:

lz  
 0 E 02  f
 xf
f   E2
 y    0 0 x
 
  f2
 y
 f2   
 
 0 E 02
ρ f      2E
2 0
2
0

 
d f2
. (7.191)
 
  x y  2  x y  2 d
Hence the total angular momentum of the beam (per unit length), is

76 The complex-exponential versions of these equalities are given by the bottom line of Eq. (100).

Chapter 7 Page 53 of 70
Essential Graduate Physics EM: Classical Electrodynamics


 E2  df 2   E2  
L z   l z d r  2  l z d   0 0   2 d   0 0   d  f .
2 2 2
(7.192)
0
 0 d  0

Taking this integral by parts, with the assumption that f  0 at   0 and    (as it is true for the
Gaussian beam (181) and all realistic paraxial beams), we finally get
 
 0 E 02  0 E 02
Lz  
   
f 2 d  2  2
 f
2
d . (7.193)
0 0

Now comparing this expression with Eq, (184), we see that remarkably, the ratio Lz/U does not
depend on the shape and the width of the beam (and of course on the wave’s amplitude E0), so these
parameters are very simply and universally related:
Angular
U momentum
Lz   . (7.194) at circular
 polarization

Since this relation is valid in the plane-wave limit a  , it may be attributed to plane waves as well,
with the understanding that in reality, they always have some width (“aperture”) restriction.
As the reader certainly knows, in quantum mechanics the energy excitations of any harmonic
oscillator of frequency  are quantized in the units of , while the internal angular momentum of a
particle is quantized in the units of s, where s is its spin. In this context, the classical relation (194) is
used in quantum electrodynamics as the basis for treating the electromagnetic field excitation quanta
(photons) as some sort of quantum particles with spin s = 1. (Such integer spin also fits the Bose-
Einstein statistics of the electromagnetic radiation.)
Unfortunately, I do not have time/space for a further discussion of the (very interesting) physics
of paraxial beams but cannot help noticing, at least in passing, the very curious effect of helical waves –
the beams carrying not only the “spin” momentum (194), but also an additional “orbital” angular
momentum. The distribution of their energy in space is not monotonic, as it is in the Gaussian beam
(181), but reminds several threads twisted around the propagation axis – hence the term “helical”.77
Mathematically, their field structure is described by the associate Laguerre polynomials – the same
special functions that are used for the quantum-mechanical description of hydrogen-like atoms.78
Presently, there are efforts to use such beams for the so-called orbital angular momentum (OAM)
multiplexing for high-rate information transmission.79

7.8. Resonant cavities

Generally, resonators are structures that may sustain oscillations (in electrodynamics, of the
electromagnetic field) even without an external source, until the oscillation amplitude slowly decreases
in time due to unavoidable energy losses. If the resonator quality (described by the so-called Q-factor,
which will be defined and discussed in the next section) is high, Q >> 1, this decay takes many

77 Noticing such solutions of the Maxwell equations may be traced back to at least a 1943 theoretical work by J.
Humblet; however, this issue had not been discussed in literature too much until experiments carried out in 1992 –
see, e.g. L. Allen et al., Optical Angular Momentum; IOP, 2003.
78 See, e.g., QM Sec. 3.7.
79 See, e.g., J. Wang et al., Nature Photonics 6, 488 (2012).

Chapter 7 Page 54 of 70
Essential Graduate Physics EM: Classical Electrodynamics

oscillation periods. Alternatively, high-Q resonators may sustain high oscillating fields permanently, if
driven by relatively weak incident waves. In contrast to lumped-element resonators, say, the well-known
LC tank circuit, the subject of this section is resonant cavities (or “distributed resonators”) limited by
either conducting or dielectric walls that contain distributed standing waves inside them.
Conceptually the simplest resonant cavity is the Fabry-Pérot interferometer80 that may be
obtained by placing two well-conducting planes parallel to each other.81 Indeed, in Sec. 3 we have seen
that if a plane wave is normally incident on such a “perfect mirror”, located at z = 0, its reflection, at
negligible skin depth, results in a standing wave described by Eq. (61b):

 
E ( z, t )  Re 2 E e it i / 2 sin kz . (7.195)

This wave would not change if we place the second mirror (isolating the segment of length l from the
external wave source) at any position z = l with sin kl = 0, i.e. with
kl  p , where p  1, 2,.... (7.196)
This condition, which determines the spectrum of own (or resonance, or eigen-) frequencies of the
resonator of fixed length l,
v 1
 p  vk p  p, with v  , (7.197)
a  1 / 2
has a simple physical sense: the resonator’s length l equals exactly p half-waves of the frequency p.
Though this is all very simple, please note a considerable change of philosophy from what we have been
doing in the previous sections: the main task of the resonator’s analysis is finding its own frequencies
p, which are now determined by the system’s geometry rather than by an external wave source.
Before we move to cavities of more complex shapes, let us use Eq. (62) to represent the
magnetic field in the Fabry-Pérot interferometer:
 E 
H ( z, t )  Re 2  e it  cos kz . (7.198)
 Z 
Expressions (195) and (198) show that in contrast to traveling waves, each field of the standing wave
changes simultaneously (proportionately) at all points of the Fabry-Pérot resonator, turning to zero
everywhere twice a period. At these instants, the energy of the corresponding field vanishes, but the
total energy of the two fields stays constant because the counterpart field oscillates with the phase shift
/2. Such behavior is typical for all electromagnetic resonators.
A more technical remark is that we can readily get the same results (195)-(198) by solving the
Maxwell equations from scratch. For example, we already know that in the absence of dispersion,
losses, and sources, they are reduced to wave equations (3) for any field components. For the Fabry-
Pérot resonator’s analysis, we can use the 1D form of these equations, say, for the transverse component
of the electric field:

80 This device, named after its inventors, Charles Fabry and Alfred Pérot; is also called the Fabry-Pérot etalon
(meaning “gauge”), because of its initial usage for light wavelength measurements.
81 The resonators formed by well-conducting (usually, metallic) walls are frequently called resonant cavities.

Chapter 7 Page 55 of 70
Essential Graduate Physics EM: Classical Electrodynamics

 2 1 2 
 2  2 2  E  0, (7.199)
 z v t 
and solve it as a part of an eigenvalue problem with the corresponding boundary conditions. Indeed, by
separating time and space variables as E(z, t) = Z(z)T(t), we obtain
1 d 2 Z 1 1 d 2T
  0. (7.200)
Z dz 2 v 2 T dt 2
Calling the separation constant k2, we get two similar ordinary differential equations,
d 2Z d 2T
2
 k 2 Z  0, 2
 k 2 v 2T  0, (7.201)
dz dt
both with sinusoidal solutions, so the product Z(z)T(t) is a standing wave with the wave vector k and
frequency  = kv. (In this form, the equations are valid even in the presence of dispersion, but with a
frequency-dependent wave speed: v2 = 1/()().) Now using the boundary conditions E(0, t) = E(l, t)
= 0,82 we get the eigenvalue spectrum for kp and hence for p = vkp, given by Eqs. (196) and (197).
Lessons from this simple case study may be readily generalized to any cavity formed as a
transmission line’s section:83 there are two approaches to finding the resonant frequency spectrum:
(i) We may look at a traveling wave solution and find where reflecting mirrors may be inserted
without affecting the wave’s structure.
(ii) We may solve the general 3D wave equations,
 2 1 2 
   2 2  f (r, t )  0 , (7.202)
 v t 
for field components, as an eigenvalue problem with appropriate boundary conditions. If the system’s
parameters (and hence the coefficient v) do not change in time, the spatial and temporal variables of Eq.
(202) may be always separated by taking
f (r, t )   fk (r )T k (t ) , (7.203)
k

where each function Tk(t) always obeys the same equation as in Eq. (201), having the sinusoidal solution
of frequency k = vk. Plugging this solution back into Eqs. (202), for the spatial distribution of the field,
we get the 3D Helmholtz equation,
 
3D
2
 k 2 fk (r )  0 , (7.204) Helmholtz
equation

whose eigenfunctions fk(r) may be much more involved, especially for non-symmetric geometries.
Let us use these approaches to find the resonant frequency spectrum of a few simple, but
practically important cavities. First of all, the first method is completely sufficient for the analysis of any

82 This is of course the expression of the first of the general boundary conditions (104). The second of these
conditions (for the magnetic field) is satisfied automatically for the transverse waves we are considering.
83 The resonators may have different geometries as well, and in many cases, only the second approach may be
used.

Chapter 7 Page 56 of 70
Essential Graduate Physics EM: Classical Electrodynamics

resonator formed as a fragment of a uniform TEM transmission line (e.g., a coaxial cable), confined
with two conducting lids normal to the line’s direction. Indeed, since in such lines kz = k = /v, and the
electric field is perpendicular to the propagation axis, e.g., parallel to the lid surface, the boundary
conditions are exactly the same as in the Fabry-Pérot resonator, and we again arrive at the
eigenfrequency spectrum (197).
Now let us analyze a slightly more complex system: a rectangular metallic-wall cavity of volume
abl – see Fig. 29. To use the first approach outlined above, let us consider the resonator as a finite-
length (z = l) section of the rectangular waveguide extended along the z-axis, which was analyzed in
detail in Sec. 6. As a reminder, at a < b, in the fundamental H10 traveling wave mode, both vectors E and
H do not depend on y, with E having only a y-component. In contrast, H has two components, Hx and
Hz, with the phase shift /2 between them, and with Hx having the same phase as Ey – see Eqs. (131),
(137), and (138). Hence, if a plane perpendicular to the z-axis, is placed so that the electric field
vanishes on it, Hx also vanishes, so both boundary conditions (104), pertinent to a perfect metallic wall,
are fulfilled simultaneously.
y
b

0 a x Fig. 7.29. Rectangular metallic-wall resonator

as a finite section of a waveguide with the
l cross-section shown in Fig. 22.
z

As a result, the H10 wave would not be perturbed by two metallic walls separated by an integer
number of half-wavelengths z/2 corresponding to the wave number given by the combination of Eqs.
(102) and (133):
2  2 

k z  k 2  k t2
1/ 2

  2  2  . (7.205)
v a 
Using this expression, we see that the smallest of these distances, l = z/2 = /kz, gives the resonance
frequency84
1/ 2
   2    2 
101  v       , (7.206)
 a   l  
where the indices of  show the numbers of half-waves along each dimension of the system, in the order
[a, b, l]. This is the lowest (“fundamental”) frequency of the resonator (if b < a, l).
The field distribution in this mode is close to that in the corresponding waveguide mode H10
(Fig. 22), with the important difference that the magnetic and electric fields are now shifted by phase /2
both in space and time, just as in the Fabry-Pérot resonator – see Eqs. (195) and (198). Such a time shift
allows for a very simple interpretation of the H101 mode, which is especially adequate for very flat

84 In most electrical engineering handbooks, the index corresponding to the shortest side of the resonator is listed
last, so the fundamental mode is nominated as H110 and its eigenfrequency as 110.

Chapter 7 Page 57 of 70
Essential Graduate Physics EM: Classical Electrodynamics

resonators, with b << a, l. At the instant when the electric field reaches its maximum (Fig. 30a), i.e.
when the magnetic field vanishes in the whole volume, the surface electric charge of the broadest (in
Fig. 30, horizontal) walls of the resonator is largest, being localized mostly near the centers of the walls.
At the immediate later times, the walls start to recharge via surface currents, whose density J is largest
in the side walls, and reaches its maximal value in a quarter of the oscillation period T = 2/101 – see
Fig. 30b. The currents generate the vortex magnetic field, with looped field lines in the plane of the
broadest face of the resonator. The surface currents continue to flow in this direction until (in one more
quarter period) the broader walls of the resonator are fully recharged in the polarity opposite to that
shown in Fig. 30a. After that, the surface currents start to flow in the direction opposite to that shown in
Fig. 30b. This process, which repeats again and again, is conceptually similar to the well-known
oscillations in a lumped LC circuit, with the role of (now, distributed) capacitance played mostly by the
broadest walls of the resonator, and that of (now, distributed) inductance, by its narrower walls.

(a) (b)

b - - - J J Fig. 7.30. Fields, charges, and

E  currents in the fundamental (H101)
 H mode of a rectangular metallic
 resonator, at two instants separated
0 + + + a by t = /2101 – schematically.

In order to generalize Eq. (206) to higher oscillation modes, the second of the approaches
discussed above is more prudent. Separating the variables in the Helmholtz equation (204) as R(r) =
X(x)Y(y)Z(z), we see that X, Y, and Z have to be either sinusoidal or cosinusoidal functions of their
arguments, with the wave vector components satisfying the characteristic equation
2
k x2  k y2  k z2  k 2  . (7.207)
v2
In contrast to the wave propagation problem, now we are dealing with standing waves along all three
dimensions, and have to satisfy the macroscopic boundary conditions (104) on all sets of parallel walls.
It is straightforward to check that these conditions (E = 0, Hn = 0) are fulfilled at the following field
component distribution:
E x  E1 cos k x x sin k y y sin k z z , H x  H 1 sin k x x cos k y y cos k z z ,
E y  E 2 sin k x x cos k y y sin k z z , H y  H 2 cos k x x sin k y y cos k z z , (7.208)
E z  E3 sin k x x sin k y y cos k z z , H z  H 3 cos k x x cos k y y sin k z z ,
with each of the wave vector components having an equidistant spectrum similar to Eq. (196):
n m p
kx  , ky  , kz  , (7.209)
a b l
so the full spectrum of resonance frequencies is given by the following formula:
1/ 2
  n  2   m  2   p  2 
 nmp  vk  v        , (7.210)
 a   b   l  

Chapter 7 Page 58 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

which is a natural generalization of Eq. (206). Note, however, that of the three integers m, n, and p, at
least two have to be different from zero to keep the fields (208) from vanishing at all points.
We may use Eq. (210), in particular, to evaluate the number of different modes in a relatively
small range d3k << k3 of the wave vector space volume that is, on the other hand, much larger than the
reciprocal volume, 1/V = 1/abl, of the cavity. Taking into account that each eigenfrequency (210), with
nml  0, corresponds to two field modes with different polarizations,85 the argumentation absolutely
similar to the one used for the 2D case at the end of Sec. 7, yields
Oscillation
d 3k
mode dN  2V . (7.211)
density (2 ) 3
This property, valid for resonators of arbitrary shape, is broadly used in classical and quantum statistical
physics,86 in the following form. If some electromagnetic mode functional f(k) is a smooth function of
the wave vector k, and the volume V is large enough, then Eq. (211) may be used to approximate the
sum of the functional’s values over the modes by an integral:
dN 3 V
 f (k )   f (k )dN   f (k ) d
k
3
k
d k2
2 3  f (k )d
3
k. (7.212)
N k k

Leaving similar analyses of resonant cavities of some other simple shapes for the reader’s
exercises, let me finish this section by noting that low-loss resonators may be also formed by finite-
length sections of not only metallic-wall waveguides of various cross-sections but also of dielectric
waveguides. Moreover, even a simple slab of a dielectric material with a / ratio substantially different
from that of its environment (say, of the free space) may be used as a high-Q Fabry-Pérot interferometer
(Fig. 31), due to an effective wave reflection from its surfaces at the normal and especially an inclined
incidence – see, respectively, Eqs. (68), and Eqs. (91) and (95).

  0

Fig. 7.31. A dielectric Fabry-Pérot interferometer.

d ~

Actually, such dielectric Fabry-Pérot interferometers are frequently more convenient for
practical purposes than metallic-wall resonators, not only due to possibly lower losses (especially in the
optical range) but also due to a natural coupling to the environment, which offers a ready way of wave
insertion and extraction – see Fig. 31 again. The backside of the same medal is that this coupling to the
environment provides an additional mechanism of power losses, limiting the resonance’s quality factor –
see the next section.

85 This fact becomes evident from plugging Eqs. (208) into the Maxwell equation E = 0. The resulting
equation, kxE1 + kyE2 + kzE3 =0, with the discrete, equidistant spectrum (209) for each wave vector component,
may be always satisfied by two linearly independent sets of the constants E1,2,3.
86 See, e.g., QM Sec. 1.1 and SM Sec. 2.6.

Chapter 7 Page 59 of 70
Essential Graduate Physics EM: Classical Electrodynamics

7.9. Energy loss effects

The inevitable energy losses (“dissipation”) in passive media lead, in two basic situations, to two
different effects. In a long transmission line fed by a constant wave source, the losses lead to a gradual
attenuation of the wave, i.e. to a decrease of its amplitude, and hence its power P, with the growing
distance z from the source. In linear materials, the power losses are proportional to the power P carried
by the wave, so the energy balance on a small segment dz takes the form
dP
dPloss  dP   dz  Pdz . (7.213)
dz
The coefficient  participating in the last form of Eq. (213), and hence defined as
dP / dz
  , (7.214)
P
is called the attenuation constant.87 Comparing the solution of Eq. (213),

P ( z )  P (0)e z , (7.215) Wave’s

attenuation

with Eq. (29), where k is replaced with kz, we see that  may be expressed as
  2 Im k z , (7.216)
where kz is the component of the wave vector along the transmission line. In the most important limit
when the losses are low in the sense  <<  kz   Re kz, its effects on the field distribution along the
line’s cross-section are negligible, making the calculation of  rather straightforward. In particular, in
this limit, the contributions to attenuation from two major sources, the energy losses in the filling
dielectric and the skin-effect losses in conducting walls, are independent and additive.
The dielectric losses are especially simple to describe. Indeed, a review of our calculations in
Secs. 5-7 shows that all of them remain valid if either (), or (), or both, and hence k(), have small
imaginary parts:
 
k"   Im  1 / 2   1 / 2    k' . (7.217)

In TEM transmission lines, kz = k, and hence Eq. (216) yields

 filling  2k"  2 Im 1 / 2   1 / 2   . (7.218) Attenuation

due to filling

For dielectric waveguides, in particular optical fibers, these losses are the main attenuation mechanism.
As was discussed in Sec. 7, in practical optical fibers tR >> 1, i.e. most of the field propagates (as an
evanescent wave) in the cladding, with a field distribution very close to the TEM wave. This is why Eq.

87 In engineering, wave attenuation is most frequently measured in decibels per meter, abbreviated as db/m (the
term not to be confused with dBm standing for decibel-milliwatt):
  1/ m 
 db/m  10 log10
P ( z)
P ( z  1 m)
 10 log10 e 
10
ln 10
 
 m 1  4.343 m 1 .  
Alternatively, it is sometimes measured in nepers per meter (Np/m) defined as np/m  /2, so db/m  8.686 Np/m.

Chapter 7 Page 60 of 70
Essential Graduate Physics EM: Classical Electrodynamics

(218) is approximately valid if it is applied to the cladding material alone. In waveguides with non-
TEM waves, we can use the relations between kz and k, derived in the previous sections, to re-calculate
k” into Im kz. (Note that at this recalculation, the values of kt have to be kept real, because they are just
the eigenvalues of the Helmholtz equation (101), which does not include the filling media parameters.).
In transmission lines and waveguides and with metallic walls, higher energy losses may come
from the skin effect. If the wavelength  is much larger than s, as it usually is,88 the losses may be
readily evaluated using Eq. (6.36):
dPloss dP  s
  H wall
2
, (7.219)
dA dA 4
where Hwall is the real amplitude of the tangential component of the magnetic field at the wall’s surface.
The total power loss Ploss/dz per unit length of a waveguide, i.e. the right-hand side of Eq. (213), now
may be calculated by the integration of this expression along the contour(s) limiting the cross-section of
all conducting walls. Since our calculation is only valid for low losses, we may ignore their effect on the
field distribution, so the unperturbed distributions may be used both in Eq. (219), i.e. in the numerator of
Eq. (214), and also for the calculation of the average propagating power, i.e. the denominator of Eq.
(214) – as the integral of the Poynting vector over the cross-section of the waveguide.
Let us see how this approach works for the TEM mode in one of the simplest transmission lines,
the coaxial cable (Fig. 20). As we already know from Sec. 5, in the coarse-grain approximation,
implying negligible power loss, the TEM mode field distributions between the two conductors are the
same as in statics, namely:
a
H z  0, H   0, H ( )  H 0 , (7.220)

where H0 is the field’s amplitude on the surface of the inner conductor, and
1/ 2
a 
E z  0, E  (  )  ZH  (  )  ZH 0 , E  0, where Z    . (7.221)
  
Neglecting the power losses for a minute, we may plug these expressions into Eq. (42) to calculate the
time-averaged Poynting vector:
2
Z H ( )
2 2
Z H0  a 
S    , (7.222)
2 2   
and from it, the total wave power flow through the cross-section:
2
Z H0 a2 b
d b
P   Sd r  2 
2
2
 Z H 0 a 2 ln . (7.223)
A
2 a  2
a
Next, for this particular system (Fig. 20), the contours limiting the wall cross-section are circles
of radii  = a (where the surface field amplitude Hwalls equals, in our notation, H0), and  = b (where,
according to Eq. (214), the field is a factor of b/a lower). As a result, for the power loss per unit length,
Eq. (219) yields

88As follows from Eq. (78), which may be used for crude estimates even in cases of arbitrary wave incidence, this
condition is necessary for low attenuation:  << k only if f << 1.

Chapter 7 Page 61 of 70
Essential Graduate Physics EM: Classical Electrodynamics

dPloss dPloss  a   0 s   a 

2

  dl   2a H 0
2
 2b H 0
2
 a 1    s H 0 . (7.224)
dz  
Ca Cb dA  b  4 2  b

Note that at a << b, the losses in the inner conductor dominate, despite its smaller surface, because of
the higher surface field.
Now we may plug Eqs. (223) and (224) into the definition (214) of , to calculate the skin-effect
contribution to the attenuation constant:
dPloss / dz 1  1 1   s k s  1 1 
 skin         . (7.225)
P 2 ln(b / a )  a b  Z 2 ln(b / a )  a b 
This result shows that the relative (dimensionless) attenuation, /k, scales approximately as the ratio
s/min[a, b], in a semi-quantitative agreement with the plane-wave result (78).
Let us use this result to evaluate  for the standard TV cable RG-6/U, with copper conductors of
diameters 2a = 1 mm, 2b = 4.7 mm, and   2.20 and   0. According to Eq. (6.33), for f = 100 MHz
(i.e.   6.3108 s-1) the skin depth of pure copper at room temperature (with   6.0107 S/m) is close
to 6.510-6 m, while k = ()1/2 = (/0)1/2(/c)  3.1 m–1. As a result, the attenuation is rather low: skin
 0.016 m-1, so the attenuation length scale ld  1/ is about 60 m. Hence the attenuation in a cable
connecting a roof TV antenna or a cable distribution box to a TV set is not a big problem, though using
a worse conductor, e.g., steel, would make the losses rather noticeable. (Hence the current worldwide
shortage of copper.) However, the use of such cable in the X-band (f ~ 10 GHz) is more problematic.
Indeed, though the skin depth s  –1/2 decreases with frequency, the wavelength drops, i.e. k
increases, even faster (k  ), so the attenuation skin  1/2 becomes close to 0.16 m–1, i.e. ld to ~6 m.
This is why at such frequencies, it may be necessary to use rectangular waveguides, with their larger
internal dimensions a, b ~ 1/k, and hence lower attenuation. Let me leave the calculation of this
attenuation, using Eq. (219) and the results derived in Sec. 7, for the reader’s exercise.
The main effect of dissipation on free oscillations in resonators is different: here it leads to a
gradual decay of the oscillating fields’ energy U in time. A useful dimensionless measure of this decay,
called the Q factor, is commonly defined by writing the following temporal analog of Eq. (213):89

 dU  Ploss dt  Udt , (7.226)
Q
where  in the resonance frequency in the loss-free limit, and
 Ploss
 . (7.227) Q-factor
Q U
The solution of Eq. (226),
Q Q / 2 QT
U (t )  U (0)e t / , with     , (7.228)
  / 2 2

89 As losses grow, the oscillation waveform deviates from the sinusoidal one, and the very notion of “oscillation
frequency” becomes vague. As a result, the parameter Q is well-defined only if it is much higher than 1.

Chapter 7 Page 62 of 70
 Essential Graduate Physics EM: Classical Electrodynamics

which is the temporal analog of Eq. (215), shows the physical meaning of the Q-factor: the characteristic
time  of the oscillation energy’s decay is (Q/2) times longer than the oscillation period T = 2/.
(Another useful interpretation of Q comes from the universal relation90

Q , (7.229)

where  is the so-called FWHM91 bandwidth of the resonance, namely the difference between the two
values of the external signal frequency, one above and one below , at which the energy of the
oscillations induced in the resonator by an input signal is twice lower than its resonance value.)
In the important particular case of a resonant cavity formed by the insertion of metallic walls into
a TEM transmission line of a small cross-section (with the linear size scale a much less than the
wavelength ), there is no need to calculate the Q-factor directly, provided that the line attenuation
coefficient  is already known. In fact, as was discussed in Sec. 8 above, the standing waves in such a
resonator, of the length given by Eq. (196): l = p(/2) with p = 1, 2,…, may be understood as an overlap
of two TEM waves propagating in opposite directions, or in other words, a traveling wave plus its
reflection from one of the ends, the whole roundtrip taking time t = 2l/v = p/v = 2p/ = pT.
According to Eq. (215), at this distance, the wave’s power drops by a factor of exp{-2l} = exp{-p}.
On the other hand, the same decay may be viewed as taking place in time, and according to Eq. (228),
results in the drop by a factor of exp{-t/} = exp{-(pT )/(Q/)} = exp{-2p/Q}. Comparing these two
exponents, we get
2 k
Q vs. 
Q  . (7.230)
 
This simple relation neglects the losses at the wave reflection from the walls limiting the
resonator length. This approximation is indeed legitimate at l >> ; if this relation is violated, or if we
are dealing with more complex resonator modes (such as those based on the reflection of E or H waves),
the Q-factor may be different from that given by Eq. (230), and needs to be calculated directly from Eq.
(227). A substantial relief for such a direct calculation is that, just at the calculation of small attenuation
in waveguides, in the low-loss limit (Q >> 1), both the numerator and denominator of the right-hand
side of that formula may be calculated neglecting the effects of the energy loss on the field distribution
in the resonator. I am leaving such a calculation, for a few simple resonant cavities, including the
rectangular and the circular ones, for the reader’s exercise.
To conclude this chapter, the last remark: in some distributed resonators (including certain
dielectric resonators92 and metallic cavities with holes in their walls), additional energy losses due to the
wave radiation into the environment are also possible. In some simple cases (say, the Fabry-Pérot
interferometer shown in Fig. 31), the calculation of these radiative losses is straightforward, but
sometimes it requires more elaborated approaches that will be discussed in the next chapter.

90 See, e.g., CM Sec. 5.1.

91 FWHM is the acronym for Full Width at Half-Maximum.
92 Due to the absence of metallic walls and the associated skin-effect losses, some microwave dielectric resonators
(e.g., those based on pure sapphire crystals cooled to helium temperatures) may feature Q-factors as high as 109 –
see, e.g., D. Creedon et al., Appl. Phys. 98, 222903 (2011).

Chapter 7 Page 63 of 70
Essential Graduate Physics EM: Classical Electrodynamics

7.10. Exercise problems

7.1.* Find the temporal Green’s function of a medium whose complex permittivity () obeys
the Lorentz oscillator model given by Eq. (32), by using:
(i) the Fourier transform of the underlying Eq. (7.30), and
(ii) the direct solution of that equation.
Hint: For the Fourier-transform approach, you may like to use the Cauchy integral.93

7.2. The electric polarization of some material responds to an electric field step94 in the
following way:
 
Pt    1 E0 1  e t / ,
0,
if E t   E0  
for t  0,
for 0  t ,
1,
where  > 0 and 1 are some constants. Calculate the complex permittivity () of this material, and
discuss a possible simple physical model giving such dielectric response.

7.3. Calculate the complex permittivity () of a material whose dielectric-response Green’s
function defined by Eq. (23), is

G    G0 1  e  / ,
with some positive constants G0 and . What is the difference between this dielectric response and the
apparently similar one considered in the previous problem?

7.4. Use the oscillator model of an atom, given by Eq. (30), to calculate its average potential
energy in a uniform, sinusoidal ac electric field, and use the result to calculate the potential profile
created for the atom by a standing electromagnetic wave with the electric field amplitude E(r).

7.5. The solution of the previous problem shows that a standing electromagnetic wave may exert
a time-averaged force on an otherwise free non-relativistic charged particle. Reveal the physics of this
force by writing and solving the equations of motion of such a particle in:
(i) a linearly-polarized monochromatic plane traveling wave, and
(ii) a similar but standing wave.

7.6. Use the first of Eqs. (54) to relate the integral  "  d to the plasma frequency for the
0

Lorentz oscillator model of a system of non-interacting particles.

7.7. Prove that Eq. (6.42) cannot be correct for all frequencies, and suggest its correction making
the result compatible with both the causality principle and the physical model (6.39).

7.8. Calculate, sketch, and discuss the dispersion relation for electromagnetic waves propagating
in a medium described by the Lorentz oscillator model (32), for the case of negligible damping.

93See, e.g., MA Eq. (15.2).

94This function E(t) is of course proportional to the well-known Heaviside step function (t) – see, e.g., MA Eq.
(14.3). I am not using this notion here just to avoid confusion between two different uses of the Greek letter .

Chapter 7 Page 64 of 70
Essential Graduate Physics EM: Classical Electrodynamics

7.9. As was briefly discussed in Sec. 2,95 a wave pulse of a finite but relatively large spatial
extension z >>   2/k may be formed as a wave packet – a sum of sinusoidal waves with wave
vectors k within a relatively narrow interval. Consider an electromagnetic plane wave packet of this
type, with the electric field distribution

E(r, t )  Re  E k e

i kz   k t dk , with k   k   k   k  ,
1/ 2



propagating along the z-axis in an isotropic, linear, and dissipation-free (but not necessarily dispersion-
free) medium. Express the full energy of the packet (per unit area of the wave’s front) via the complex
amplitudes Ek, and discuss its dependence on time.

7.10. Prove the Lorentz reciprocity relation (6.121) for a linear isotropic medium.

7.11.* A plane wave of frequency  is normally incident, from free space, on a plane surface of a
collision-free plasma with the electron density growing linearly and slowly with the distance from the
surface: n = z for z  0, where  > 0 is a small constant. Calculate the functional form of the resulting
standing wave’s “tail” inside the plasma.

7.12.* Analyze the effect of a time-independent uniform magnetic field B0, parallel to the
direction n of an electromagnetic wave propagation, on the wave’s dispersion in plasma, within the
same simple model that was used in Sec. 2 for the derivation of Eq. (38). (Limit your analysis to
relatively weak waves, whose magnetic field is much smaller than B0.)
Hint: You may like to represent the incident wave as a linear superposition of two circularly
polarized waves, with opposite polarization directions.

7.13. A monochromatic plane electromagnetic wave is normally incident, from free space, on a
uniform slab with electric permittivity  and magnetic permeability , with the slab’s thickness d
comparable with the wavelength.
(i) Calculate the power transmission coefficient T, i.e. the fraction of the incident wave’s power,
that is transmitted through the slab.
(ii) Assuming that  and  are frequency-independent and positive, analyze in detail the
frequency dependence of T. In particular, how does the function T () depend on the slab’s thickness d
and the wave impedance Z  (/)1/2 of its material?

7.14. A plane electromagnetic wave with a free-space wave number k0 is normally incident on a
planar conducting film of thickness d ~ s << 1/k0. Calculate the power transmission coefficient of the
system and analyze the result in the limits of small and large values of the ratio d/s.

95 For even more detail, see CM Sec. 5.3 and especially QM Sec. 2.2.

Chapter 7 Page 65 of 70
Essential Graduate Physics EM: Classical Electrodynamics

7.15. One of the results of the previous problem’s solution was the following expression for the
coefficient of power transmission of a plane electromagnetic wave through a thin conducting film of
thickness d << s, , at normal incidence:
1
T  ,
1  Z 0 / 2R 2
where R  1/d is the sheet resistance (“resistance per square”) of the film. Derive this formula in a
simpler way, utilizing the smallness of d from the very beginning. Also, calculate the power reflection
coefficient R, compare it with T, and comment.
 0 , 0 ,  ' , '
7.16. A plane wave of frequency  is normally incident, from
free space, on a plane surface of a material with real electric
permittivity ’ and magnetic permeability ’. To minimize the wave’s
reflection from the surface, it may be covered with a layer, of thickness
d, of another transparent material – see the figure on the right.
Calculate the optimal values of , , and d. d

7.17. A monochromatic plane wave is incident from inside a medium with  > 00 on its planar
surface, at an incidence angle  larger than the critical angle c = sin–1(00/)1/2. Calculate the depth 
of the evanescent wave penetration into the free space, and analyze its dependence on . Does the result
depend on the wave’s polarization?

7.18. Calculate the critical angle c for a wave of frequency , incident from free space upon a
planar surface of a plasma with electron density n, and discuss the implications of the result for
ultraviolet and X-ray optics.

7.19. Analyze the possibility of propagation of surface electromagnetic waves along a planar
boundary between plasma and free space. In particular, calculate and analyze the dispersion relation of
the waves.
Hint: Assume that the magnetic field of the wave is parallel to the boundary and perpendicular to
the wave’s propagation direction. (After solving the problem, justify this mode choice.)

z
7.20. Light from a very distant source arrives to an observer through a
plane layer of nonuniform medium with a certain 1D gradient of its refraction i
index n(z), at angle 0 – see the figure on the right. What is the genuine direction
i to the source, if n(z)  1 at z  ? (This problem is evidently important for n z  0
high-precision astronomical measurements at the Earth’s surface.)

7.21. Calculate the TEM impedance ZW of uniform TEM transmission lines with well-conducting
electrodes and the cross-sections shown in the figure below:

Chapter 7 Page 66 of 70
Essential Graduate Physics EM: Classical Electrodynamics

(i) w  d (ii) (iii)

2R
d1
d  R d d2
2R w  d1, 2

(i) two parallel round wires separated by distance d >> R,

(ii) a microstrip line of width w >> d,
(iii) a stripline with w >> d1 ~ d2,
in all cases using the coarse-grain boundary conditions on conductor surfaces. Assume that the
conductors are embedded into a linear dielectric with constant  and .

7.22. Modify the solution of Task (ii) of the previous problem for a superconductor microstrip
line, taking into account the magnetic field’s penetration into both the strip and the ground plane.

7.23.* What lumped ac circuit would be equivalent to the TEM-line system shown in Fig. 19,
with an incident wave’s power Pi? Assume that the wave reflected from the lumped load circuit does
not return to it.

7.24. Find the lumped ac circuit equivalent to a loss-free

TEM transmission line of length l ~ , with a small cross-section  
area A << 2, as “seen” (measured) from one end, if the line’s
conductors are galvanically connected (“shortened”) at the other l~
end – see the figure on the right. Discuss the result’s dependence on
the signal frequency.

7.25. Represent the fundamental H10 wave in a rectangular waveguide (Fig. 22) with a sum of
two plane waves, and discuss the physics behind such a representation.

7.26.* For the coaxial cable (see, e.g., Fig. 20), find the lowest non-TEM mode and calculate its
cutoff frequency.

7.27. Two coaxial cable sections are connected coaxially –

see the figure on the right, which shows the system’s cut along its b a' b'
symmetry axis. Relations (118) and (120) seem to imply that if the a
ratios b/a of these sections are equal, their impedance matching is
perfect, i.e. a TEM wave incident from one side on the connection
would pass it without any reflection at all: R = 0. Is this statement
correct?

7.28. Calculate the cutoff frequencies c of the fundamental mode

and the next lowest mode in the so-called ridge waveguide with the cross- a a
w
section shown in the figure on the right, in the limit t << a, b, w. Briefly b
t

Chapter 7 Page 67 of 70
Essential Graduate Physics EM: Classical Electrodynamics

discuss possible advantages and drawbacks of such waveguides for signal transfer and physical
experiment.

7.29. Prove that TEM-like waves may propagate, in the radial direction, in the 0
free space between two coaxial, round, well-conducting cones – see the figure on the
right. Can this system be characterized by a certain transmission line impedance ZW, as
defined by Eq. (115)?
7.30. Use the recipe outlined in Sec. 7 to prove the characteristic equation (161)
for the HE and EH waves in step-index optical fibers with a round cross-section.

7.31. Derive an approximate equation describing spatial variations of the complex amplitude of a
general monochromatic paraxial beam propagating in a uniform medium, for the case when these
variations are sufficiently slow. Is the Gaussian beam described by Eq. (181) one of the possible
solutions of this equation? Give your interpretation of the last result.
z
7.32. Calculate the lowest resonance frequencies and the corresponding l
field distributions of standing electromagnetic waves inside a round cylindrical
cavity with well-conducting walls (see the figure on the right), neglecting the 0
skin depth s in comparison with l and R.
R

7.33. Analyze electromagnetic waves that may propagate inside a relatively narrow gap between
two well-conducting concentric spherical shells of radii R and R + d, in the limit d << R.
(i) Within the coarse-grain approximation, derive the 2D equation describing such waves with
relatively large wavelengths  ~ R >> d.
(ii) Calculate the lowest resonance frequencies of the system.

7.34. A molecule with an electric polarizability  is placed inside an otherwise empty

macroscopic cavity with well-conducting walls. Express the resulting shifts of its resonance frequencies
via the unperturbed field distribution in the corresponding mode.

7.35. A plane monochromatic wave propagates through a medium with an Ohmic conductivity 
and negligible electric and magnetic polarization effects. Calculate the wave’s attenuation and relate the
result to a certain calculation carried out in Chapter 6.

7.36. Generalize the telegrapher’s equations (110)-(111) by accounting for small energy losses:
(i) in the transmission line’s conductors, and
(ii) in the medium separating the conductors,
using their simplest (Ohmic) models. Formulate the conditions of validity of the resulting equations.

7.37. Calculate the skin-effect contribution to the attenuation constant  of a TEM wave in the
microstrip line discussed in Problem 21 (ii).

Chapter 7 Page 68 of 70
Essential Graduate Physics EM: Classical Electrodynamics

7.38. Calculate the skin-effect contribution to the attenuation coefficient  defined by Eq. (214),
for the fundamental (H10) mode propagating in a waveguide with well-conducting walls, of a rectangular
cross-section – see Fig.22. Use the results to evaluate the wave decay length ld  1/ of a 10 GHz wave
in the standard X-band waveguide WR-90 (with copper walls, a = 23 mm, b = 10 mm, and no dielectric
filling), at room temperature. Compare the result with that (obtained in Sec. 9) for the standard TV
coaxial cable, at the same frequency.

7.39. Calculate the skin-effect contribution to the attenuation coefficient  of

(i) the fundamental (H11) wave, and
(ii) the H01 wave,
in a conductor-wall waveguide with the circular cross-section (see Fig. 23a), and analyze the low-
frequency ( c) and high-frequency ( >> c) behaviors of  for each of these modes.

7.40. For a rectangular cavity of dimensions abl, with b  a, l, calculate the Q-factor of the
fundamental oscillation mode, due to the skin-effect losses in its conducting walls. Evaluate the factor
for a 232310 mm3 cavity with copper walls, at room temperature.
z
7.41.* Calculate the lowest eigenfrequency and the Q-factor (due to r
the skin-effect losses) of the axially symmetric toroidal cavity with well-
d
conducting walls and the interior’s cross-section shown in the figure on the 0 R
right, for the case d << r, R.96

7.42. Express the contribution to the damping coefficient (the reciprocal Q-factor) of a resonant
cavity, by small energy losses in the dielectric that fills it, via the complex functions () and () of
the material.

7.43. For the dielectric Fabry-Pérot resonator (Fig. 31) with the normal wave incidence, calculate
the Q-factor due to radiation losses, in the limit of a strong impedance mismatch (Z >> Z0), by using two
approaches:
(i) from the energy balance, using Eq. (227), and
(ii) from the frequency dependence of the power transmission coefficient, using Eq. (229).
Compare the results.

96 Such resonators are broadly used in particle accelerators and also in vacuum electron devices for high-power
microwave amplification and generation (e.g., the so-called klystrons), where the electric field has to be
concentrated in the region of charged particle passage – typically, along the symmetry axis (in the figure above,
the z-axis), through a pair of small holes in the cavity’s walls, which do not affect the field distribution
substantially.

Chapter 7 Page 69 of 70
Essential Graduate Physics EM: Classical Electrodynamics

This page is
intentionally left
blank

Chapter 7 Page 70 of 70