Lecture 25

Radiation by a Hertzian Dipole

Radiation of electromagnetic field is of ultimate importance for wireless communication sys-

tems. The first demonstration of the wave nature of electromagnetic field was by Heinrich
Hertz in 1888 [18], some 23 years after Maxwell’s equations were fully established. Guglielmo
Marconi, after much perserverence with a series of experiments, successfully transmitted
wireless radio signal from Cornwall, England to Newfoundland, Canada in 1901 [136]. The
experiment was serendipitous since he did not know that the ionosphere was on his side:
The ionosphere helped to bounce the radio wave back to earth from outer space. Marconi’s
success ushered in the age of wireless communication, which is omni-present in our daily
lives. Hence, radiation by arbitrary sources is an important topic for antennas and wireless
communications. We will start with studying the Hertzian dipole which is the simplest of
radiation sources we can think of.

25.1 History

The original historic Hertzian dipole experiment is shown in Figure 25.1. It was done in 1887
by Heinrich Hertz [18]. The schematic for the original experiment is also shown in Figure
25.2.
A metallic sphere has a capacitance in closed form with respect to infinity or a ground
plane.1 Hertz could use those knowledge to estimate the capacitance of the sphere, and also,
he could estimate the inductance of the leads that are attached to the dipole, and hence, the
resonance frequency of his antenna. The large sphere is needed to have a large capacitance,
so that current can be driven through the wires. As we shall see, the radiation strength of
the dipole is proportional to p = ql the dipole moment.

1 We shall learn later that this problem can be solved in closed form using image theorem.

271
272 Electromagnetic Field Theory

Figure 25.1: Hertz’s original experiment on a small dipole (courtesy of Wikipedia [18]).

Figure 25.2: More on Hertz’s original experiment on a small dipole (courtesy of Wikipedia
[18]). The antenna was powered by a transformer. The radiated electromagnetic field was
picked up by a loop receiver antenna that generates a spark at its gap M .
Radiation by a Hertzian Dipole 273

25.2 Approximation by a Point Source

Figure 25.3: Schematic of a small Hertzian dipole which is a close approximation of that first
proposed by Hertz.

Figure 25.3 is the schematic of a small Hertzian dipole resembling the original dipole that
Hertz made. Assuming that the spheres at the ends store charges of value q, and l is the
effective length of the dipole, then the dipole moment p = ql. The charge q is varying in time
harmonically because it is driven by the generator. Since

dq
= I,
dt

we have the current moment

dq
Il = l = jωql = jωp (25.2.1)
dt

for this Hertzian dipole.

A Hertzian dipole is a dipole which is much smaller than the wavelength under consid-
eration so that we can approximate it by a point current distribution, or a current density.
Mathematically, it is given by [32, 44]

J(r) = ẑIlδ(x)δ(y)δ(z) = ẑIlδ(r) (25.2.2)

The dipole is as shown in Figure 25.3 schematically. As long as we are not too close to the
dipole so that it does not look like a point source anymore, the above is a good mathematical
model and approximation for describing a Hertzian dipole.
274 Electromagnetic Field Theory

Figure 25.4: The field of a point dipole field versus that of a dipole field. When one is far
away from the dipole sources, their fields are similar to each other (courtesy of Wikepedia).

We have learnt previously that the vector potential is related to the current as follows:
 0
e−jβ|r−r |
A(r) = µ dr0 J(r0 ) (25.2.3)
4π|r − r0 |
Since the current is a 3D delta function in space, using the sifting property of a delta function,
the corresponding vector potential is given by
µIl −jβr
A(r) = ẑ e (25.2.4)
4πr
Since the vector potential A(r) is cylindrically symmetric, the corresponding magnetic field
is obtained, using cylindrical coordinates, as
 
1 1 1 ∂ ∂
H= ∇×A= ρ̂ Az − φ̂ Az (25.2.5)
µ µ ρ ∂φ ∂ρ
∂
p
where ∂φ = 0, r = ρ2 + z 2 . In the above, we have used the chain rule that

∂ ∂r ∂ ρ ∂ ρ ∂
= =p = .
∂ρ ∂ρ ∂r ρ2 + z 2 ∂r r ∂r

As a result,
 
ρ Il 1 1
H = −φ̂ − − jβ e−jβr (25.2.6)
r 4π r2 r
ρ
In spherical coordinates, r = sin θ, and (25.2.6) becomes [32]
Il
H = φ̂ (1 + jβr)e−jβr sin θ (25.2.7)
4πr2
Radiation by a Hertzian Dipole 275

The electric field can be derived using Maxwell’s equations.

 
1 1 1 ∂ 1 ∂
E= ∇×H= r̂ sin θHφ − θ̂ rHφ
jω jω r sin θ ∂θ r ∂r
Ile−jβr h 2 2
i
= r̂2 cos θ(1 + jβr) + θ̂ sin θ(1 + jβr − β r ) (25.2.8)
jω4πr3

Figure 25.5: Spherical coordinates are used to calculate the fields of a Hertzian dipole.

25.2.1 Case I. Near Field, βr  1

Since βr  1, retardation effect within this short distance from the point dipole can be
ignored. Also, we let βr → 0, and keeping the largest terms (or leading order terms in math
parlance), then from (25.2.8), with Il = jωp
p
E∼
= (r̂2 cos θ + θ̂ sin θ), βr  1 (25.2.9)
4πr3
For the H field, from (25.2.7), with βr  1, then
jωp
H = φ̂ sin θ (25.2.10)
4πr2
or
jβrp
η0 H = φ̂ sin θ (25.2.11)
4πεr3
Thus, it is seen that

η0 H  E, when βr  1 (25.2.12)
276 Electromagnetic Field Theory

where p = ql is the dipole moment.2 The above implies that in the near field, the electric
field dominates over the magnetic field.
In the above, βr could be made very small by making λr small or by making ω → 0. The
above is like the static field of a dipole.
Another viewpoint is that in the near field, the field varies rapidly, and space derivatives
are much larger than the time derivative.3
For instance,
∂ ∂

∂x c∂t
Alternatively, we can say that the above is equivalent to
∂ ω

∂x c
or that
1 ∂2
∇2 − ≈ ∇2
c2 ∂t2
In other words, static theory prevails over dynamic theory when βr  1. The above approx-
imations are consistent with that the retardation effect is negligible over this lengthscale.

25.2.2 Case II. Far Field (Radiation Field), βr  1

This is also known as the far zone. In this case, retardation effect is important. In other
words, phase delay cannot be ignored.
Il −jβr
E∼
= θ̂jωµ e sin θ (25.2.13)
4πr
and
Il −jβr
H∼
= φ̂jβ e sin θ (25.2.14)
4πr
ωµ
Eθ
pµ
Note that H φ
= β =  = η0 . Here, E and H are orthogonal to each other and they are
both orthogonal to the direction of propagation, as in the case of a plane wave. Or in a word,
a spherical wave resembles a plane wave in the far field approximation.

25.3 Radiation, Power, and Directive Gain Patterns

The time average power flow in the far field is given by
 2
1 1 2 η0 βIl
hSi = <e[E × H∗ ] = r̂ η0 |Hφ | = r̂ sin2 θ (25.3.1)
2 2 2 4πr
p
2 Here, η0 = µ/ε. We multiply H by η0 so that the quantities we are comparing have the same unit.
3 This is in agreement with our observation that electromagnetic fields are great contortionists: They will
deform themselves to match the boundary first before satisfying Maxwell’s equations. Since the source point
is very small, the fields will deform themselves so as to satisfy the boundary conditions near to the source
region. If this region is small compared to wavelength, the fields will vary rapidly over a small lengthscale
compared to wavelength.
Radiation by a Hertzian Dipole 277

The radiation field pattern of a Hertzian dipole is the plot of |E| as a function of θ at a
constant r. Hence, it is proportional to sin θ, and it can be proved that it is a circle. The
radiation power pattern is the plot of hSr i at a constant r.

Figure 25.6: Radiation field pattern of a Hertzian dipole. It can be shown that the pattern
is a circle.

Figure 25.7: Radiation power pattern of a Hertzian dipole which is also the same as the
directive gain pattern.
278 Electromagnetic Field Theory

The total power radiated by a Hertzian dipole is thus given by

 2π  π  π  2
η0 βIl
P = dφ dθr2 sin θhSr i = 2π dθ sin3 θ (25.3.2)
0 0 0 2 4π

Since

 π  −1  1
3 2 4
dθ sin θ = − (d cos θ)[1 − cos θ] = dx(1 − x2 ) = (25.3.3)
0 1 −1 3

then

2
η0 (βIl)2

4 βIl
P = πη0 = (25.3.4)
3 4π 12π

The directive gain of an antenna, G(θ, φ), is defined as [32]

hSr i hSr i
G(θ, φ) = = P (25.3.5)
hSav i 4πr 2

where

P
hSav i = (25.3.6)
4πr2

is the power density if the power P were uniformly distributed over a sphere of radius r.
Notice that hSav i is independent of angle. Hence, the angular dependence of the directive
gain G(θ, φ) is coming from hSr i.
Substituting (25.3.1) and (25.3.4) into the above, we have

 2
η0 βIl
2 4πr sin2 θ 3 2
G(θ, φ) =  2 = sin θ (25.3.7)
1 4 βIl 2
4πr 2 3 η0 π 4π

The peak of G(θ, φ) is known as the directivity of an antenna. It is 1.5 in the case of
a Hertzian dipole. If an antenna is radiating isotropically, its directivity is 1, which is the
lowest possible value, whereas it can be over 100 for some antennas like reflector antennas
(see Figure 25.8). A directive gain pattern is a plot of the above function G(θ, φ) and it
resembles the radiation power pattern.
Radiation by a Hertzian Dipole 279

Figure 25.8: The gain of a reflector antenna can be increased by deflecting the power radiated
in the desired direction by the use of a reflector (courtesy of racom.eu).

If the total power fed into the antenna instead of the total radiated power is used in the
denominator of (25.3.5), the ratio is known as the power gain or just gain and the pattern
is the power gain pattern. The total power fed into the antenna is not equal to the total
radiated power because there could be some loss in the antenna system like metallic loss.

25.3.1 Radiation Resistance

Engineers love to replace complex systems with simpler systems. Simplicity rules again! This
will make interface with electronic driving circuits for the antenna easier to derive. A raw
Hertzian dipole, when driven by a voltage source, essentially looks like a capacitor due to the
preponderance of electric field energy stored in the dipole field. But at the same time, the
dipole radiates giving rise to radiation loss. Thus a simple circuit equivalence of a Hertzian
dipole is a capacitor in series with a resistor. The resistor accounts for radiation loss of the
dipole.
280 Electromagnetic Field Theory

Figure 25.9: (a) Equivalent circuit of a raw Hertzian dipole without matching. (b) Equivalent
circuit of a matched Hertzian dipole (using maximum power transfer theorem). (c) Equivalent
circuit of a matched dipole at the resonance frequency of the LC tank circuit.

Hence, the way to drive the Hertzian dipole effectively is to use matching network for
maximum power transfer. Or an inductor has to be added in series with the intrinsic ca-
pacitance of the Hertzian dipole to cancel it at the resonance frequency of the tank circuit.
Eventually, after matching, the Hertzian dipole can be modeled as just a resistor. Then the
power absorbed by the Hertzian dipole from the driving source is P = 12 I 2 Rr . Thus, the
radiation resistance Rr is the effective resistance that will dissipate the same power as the
radiation power P when a current I flows through the resistor. Hence, it is defined by [32]

2P (βl)2
Rr = = η0 ≈ 20(βl)2 , where η0 = 377 ≈ 120π Ω (25.3.8)
I2 6π

For example, for a Hertzian dipole with l = 0.1λ, Rr ≈ 8Ω.

The above assumes that the current is uniformly distributed over the length of the Hertzian
dipole. This is true if there are two charge reservoirs at its two ends. For a small dipole with
no charge reservoir at the two ends, the currents have to vanish at the tips of the dipole as
shown in Figure 25.10. The effective length of an equivalent Hertzian dipole for the dipole
with triangular distribution is half of its actual length due to the manner the currents are
distributed.4 Such a formula can be used to estimate the radiation resistance of a dipole.
For example, a half-wave dipole does not have a triangular current distribution a sinusoidal
one as shown in Figure 25.11. Nevertheless, we approximate the current distribution of a half-
wave dipole with a triangular distribution, and apply the above formula. We pick a = λ2 , and
let leff = λ4 in (25.3.8), we have

Rr ≈ 50Ω (25.3.9)

4 As shall be shown, when the dipole is short, the details of the current distribution is inessential in

determining the radiation field. It is the area under the current distribution that is important.
Radiation by a Hertzian Dipole 281

Figure 25.10: The current pattern on a short dipole can be approximated by a triangle since
the current has to vanish at the end points of the short dipole. Furthermore, this dipole can
be approximated by an effective Hertzian dipole half its length with uniform current.

The true current distribution on a half-wave dipole resembles that shown in Figure 25.11.
The current is zero at the end points, but the current has a more sinusoidal-like distribution
as in a transmission line. Hence, a half-wave dipole is not much smaller than a wavelength
and does not qualify to be a Hertzian dipole. Furthermore, the current distribution on the
half-wave dipole is not triangular in shape as above. A more precise calculation shows that
Rr = 73Ω for a half-wave dipole [54]. This also implies that a half-wave dipole with sinusoidal
current distribution is a better radiator than a dipole with triangular current distribution.

In fact, one can think of a half-wave dipole as a flared, open transmission line. In the
beginning, this flared open transmission line came in the form of biconical antennas which are
shown inp Figure 25.12 [137]. If we recall that the characteristic impedance of a transmission
line is L/C, then as the spacing of the two metal pieces becomes bigger, the equivalent
characteristic impedance gets bigger. Therefore, the impedance can gradually transform
from a small impedance like 50 Ω to that of free space, which is 377 Ω. This impedance
matching helps mitigate reflection from the ends of the flared transmission line, and enhances
radiation. Because of the matching nature of bicone antennas, they are better radiators with
higher radiaton loss and lower Q. Thus they have a broader bandwidth, and are important
in UWB (ultra-wide band) antennas [138].
282 Electromagnetic Field Theory

Figure 25.11: Approximate current distribution on a half-wave dipole (courtesy of electronics-

notes.co). The currents are zero at the two end tips due to the current continuity equation,
or KCL.

Figure 25.12: A bicone antenna can be thought of as a transmission line with gradually
changing characteristic impedance. This enhances impedance matching and the radiation of
the antenna (courtesy of antennasproduct.com).