Home Search Collections Journals About Contact us My IOPscience

An introduction to the Poynting vector

This content has been downloaded from IOPscience. Please scroll down to see the full text.

1979 Phys. Educ. 14 301

(http://iopscience.iop.org/0031-9120/14/5/004)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 129.186.1.55
This content was downloaded on 29/09/2013 at 16:02

Please note that terms and conditions apply.

Phys Educ V01 14, 1979 Printed In Great Brllain

Figure 6 Comparison of holography with optical

imaging using lenses. In (a) the image is real; in (b)
and (c) the image is virtual ~ *,” --
82
Figure 7 Magnification of virtual image
of magnifying thevirtualimage is to move the
hologram along. the expanding laser beam (figure 7a);
the laser light is diffracted at the same angle, but the
distance AB increases. The dimensions of the image
also depend on the form of the incident wave (figure References
7b, c, d). When the incident wave is convergent, then Born M and Wolf E 1964 Principles of Optics (Oxford:
for a certain angle the virtual image disappears (the Pergamon Press)
real image, of course, still exists but becomes smaller). Butters J 1971 Holography and its Technology (London)
This is so because the expanding wave of the virtual Halliday D and Resnick R 1962 Physics, Part 2 (New York:
image is transformed intoa plane wave (figure 7e). Wiley)
Hands R A 1972 Light for Advanced Courses (London:
Bell)
Conclusion Kock E 1975 ‘Sound visualisation and holography’ Phys.
Teach. 13 14
The experimentsdescribed above lead to a better
Vienot 3 Ch, Smigielski P and Royer H 1971 Holographie
understanding not only of the holographic method of Optique (Pans: Dunod)
producing images, but also of opticalimaging by other Nuffield Advanced Science Electromagnetic Waves-Unit 8
opticalsystems. Holographymay be used as an (London: Longman)
attractive means of teaching the principles of optics, Nuffield Physics 1967 Teacher’s Guide-Book 3 (London:
especially in the newer physics courses. Longman)

Poynting vector at a relatively advanced stage in their

An introduction to education, usually in the context of the flow of energy
in electromagnetic waves, although of coursesome
the Poynting elementary texts (see, for
example, Beiser 1968)
introduce the topic briefly. The purpose of this article
vector is to draw attention to the possibility of introducing
the Poynting vector concept at a much earlier, even
elementary, stage, duringdiscussion of simple d.c.
N MORTON circuits, which usually includes the resistive behaviour
Department of Pure and Applied Physics, andsome description of theelectric and magnetic
University of Salford fields associated with the circuits. This approach has
the advantage that concepts of the manner in which
Most studentsof physics first encounter theconcept of energy is conveyed from a battery, say, to a resistor
electromagnetic
energy flow associated with the connected across the terminals are considerably

0031-9120/79/050301+04$01 00 0 1979ThelnslltuleofPhyslcs 30 1
 b C

I
1
E

~ ~~~

Figure 1 Parallel-plate d.c. transmission line viewed as (a)a parallel-plate capacitor, showing the electric field
lines (broken lines), and (b)a strip section of a coaxial line, showing the magnetic field lines (full lines).
(C) Orientations of the electric and magnetic field lines and thePoynting vector (broad arrow)

refined, as compared with the rather imprecise under- As an example we consider the flow of electro-
standing gained from very elementary texts (see, for magnetic energy, say from a battery, along a loss-free,
example, Nelkon and Parker 1977) that the energy is parallel-plate d.c. line, and the ultimate Joulean dissi-
in some wayconveyed by the currentitself. pation in a cylindricalresistorconnected across the
The Poynting vector S is most compactly defined in ends of the line. The geometry chosen has the merit
terms of electric and magnetic field strengths by the thatthe requiredmagnitudes of the electric and
formula magnetic fields are readily obtained, even without
resort to concepts as advanced as line integrals. Of
S=EAH
(1) course no particular originality is claimed forthe
and it represents the local rate of flow of energy per results of the calculations, which are described,for
unit cross-sectional area. Objections can be raised to example, in texts by Barnes (1965) and Scott(1966).
thisstrictlylocal definition of S (see, forexample,
Bleaney and Bleaney 1976); however the definition is
certainly appropriate for the present purposes. In the Parallel-plate d.c. transmission line
elementary case that E and H are mutually We discussfirst the calculation of the power flow
perpendicular, S is perpendicular toboth E and along the line, which comprises two nominally
H, andthe magnitudeof S is the simple product perfectly conducting, thin metal plates, ofwidth W and
of the two field strengths. The direction of flow of the spacing S, where W >> S, having much the appearance
power is given in this case by an adaptation of the of the familiarparallel-plate capacitor in elongated
familiar Fleming’s left-hand rule for the forces form, as shown in figure la. The line is assumed to be
experienced by conductors in magneticinduction, sufficiently long that any distortion of the field dis-
replacing theforce by S, thecurrent by E and the tributions due to the connections at each end may be
induction by H. Theformula for S can clearly be ignored in the subsequentdiscussion. Duetothe
introduced withoutemploying vector notation, if battery the surfaces of the plates become oppositely
desired, as in the elementary treatment of forces on charged, and the corresponding electric field (shown
conductors. by broken lines in figure la) is largely confined to the
It is perhaps noteworthy that the magnetic vector region between the plates, to which it is perpen-
appearingin the most compact formula for S is H, dicular, and takes themagnitude
rather then B, although the latter, as the force vector,
JEl= V/s
is generally regarded asthemore basic, and is the
vector commonly used nowadays in elementary texts. where V is the supply voltage. Unless recourse ismade
If it is desired to omit any reference to the vector H it to a line integralformulation of Ampire’s law the
may be replaced in all formulae quoted in this article strength of the magnetic field in the region of interest,
by B/p, to agoodapproximation, as all the media between the plates (where E and therefore S are finite),
involved are assumed to be essentially nonmagnetic. may be calculated by regarding the plates as a longi-
Note that the symbol N is frequently used instead of tudinal strip section of a large diameter coaxial line,
thepreferredsymbol S for the Poyntingvector, to comprising concentric cylindrical conductors with
avoid confusion with the symbol for area. +
radii r and r S, respectively, where r>> S, as shown in

302
figure lb. If a current i flows along the longitudinal
strip sections the total in the complete circumference
of each cylindrical conductor, say Z, will be given by
II W
,’
r’
Z = 2nr i/w. (3)
From the familiar elementary form of Amp&re’s law
the corresponding magneticfieldissolenoidal and
confined to the annular space between the conductors,
is approximately uniform and of magnitude
I H I =Z/2nr = i/w (4)
using equation (3). H is perpendicular to E, and hence
S is directed along the line as shown in figure IC, as
intuitively expected. It is worth noting that only the
inner cylindrical conductor effectively contributes to Figure 2 Parallel-plate line terminated by a cylindrica
the field. Alternatively the strips can be bent to form resistor (stippled)connected directly betweenthe
completecylindrical conductors, with radii w/2 II, conductors
resulting in an identical value for I H I on application
of Ampere’s law. with the remainder of the circuit, and in particular
Concepts of the storage of energy by electric and with the nearby connectingelements. The simplest
magneticfields are readily introduced in elementary case in which a fairlyaccurate description of the fields
discussionof, for example, the stored energy in a may be given without very detailedcalculations is that
capacitor or solenoid, respectively; therefore the flow depicted in figure 2, where a resistor of length S and
of energy associated with the Poynting vector can small diameter d (i.e. d<<s, W) is connected directly
reasonably be postulated. In the present case the between the plates constituting the transmission line.
product of 1 E 1 and 1 H I , as mentioned, is numerically In this case no charges appear on the cylindrical face
significantonlyover the rectangular cross-sectional of the resistor (see Booker 1959) and the description
area ws enclosed by the plates; thus the total power P of the electric fieldis correspondingly simplified.A
flowing alongthe line is givenby longitudinal electric fieldE’,,will exist inside a resistor
corresponding to the voltage drop across it, and the
P = ( E ( (HI w s = i V (5)
electriclines of forcemay be regarded as running
where the usual product formula for electrical power is between charges on the end faces of the resistor. The
recoveredusing equations ( 2 ) and (4).Notice that electricfield E external to the resistor will havea
energyis carried forward in the vicinity of both component Ellparallel to the surface, equal in strength
conductors, despite the current reversal.Aloss-free to E’I,by conventional boundary condition arguments.
lineis therefore advantageously, and in fact more In the simple case shown in figure 2 it is clear that E
rigorously, regarded as aguideforelectromagnetic itself is infact parallel to the resistor, and
energy.
JE;,1 = VIS.
It is not difficult in
principle to extendthe (6)
calculation shown above to other conductor From Amp&re’s law a solenoidal magnetic field exists
geometries, for example a parallel-wire line, although both inside and outside a linear conductor, and, at the
in such cases only order of magnitude estimates of the surface of the resistor, of diameter d, this has
power flow can be obtained easily. In fact itis magnitude 1 Hd 1, where
generallyonlyin cases where the geometry of the
conductors iscomplicated and that of the field
I Hd I = i/nd.
(7)
carrying regionsrelativelysimple(i.e. the reverse of Strictly equation (7) represents the value appropriate
the usualcase), as in waveguides or the example for an infinitely long conductor, and the currents in the
above, that it becomes economical to discuss a circuit transmission line also makesome contribution. But
in terms of the electric and magnetic fields,rather than since d<< W and the line fields will be of order i/w over
in terms of voltages and currents in the usual way. most of the resistor (as in equation(4)), much lessthan
values givenby equation (7). we mayignorethese
corrections. Since H,, isperpendicular to EPIIthe
Joulean dissipation in a resistor vector formulation of equation (l), or using the le%
We considernow the dissipation of the transmitted handrule,shows that electromagnetic energy enters
energyinacylindrical resistor. It is not strictly the resistor across the cylindrical surface with area
possible to discuss the magnetic and electric field zds, and the total power is given by
distributions associated with the current and electric P = ~ E ’ II(Hd(lldS=
I iv
charges on a resistor in isolation from those associated (8)

303
 charged ones, as shown schematically in figure 3.
t Thus in general theelectric field external to the
resistor has components Ell and E l parallel and per-
pendicular to thesurface, respectively. The component
El allows the passage of power parallel to the resistor,
in the case shown in figure 3, from both ends, due to
the change in direction of El along the resistor. The
component Ell diverts the power into theresistor.
Detailed computations of theelectric and magnetic
field distributions in the vicinity of the resistor and
connecting wires might in principle be made from
appropriate solutions of Laplace’s equation and
integration of the Biot-Savart expression, employing
numerical techniques, and this represents a formidable
problem in general. However, the formulaforthe
power entering the resistor is easily seen to be the
elementary result shown in equation (g), where S is the

“ 4
=c+c 1 length of the resistor, of course.

Discussion
Interms of the elementary Drude model of con-
Figure 3 Schematic diagram of the charge duction the flow of electromagnetic energy associated
distribution, the electric and magnetic lines of force with the Poynting vector, into the resistor, supplies the
and thePoynting vector in the vicinity of a cylindrical kinetic energy picked up by the conduction electrons
resistor connected by loss-free lines to a battery. Note during periods of acceleration in the internal electric
that theelectric field lines meet the resistor at anangle, field E’ll, before theydissipateitasJouleheat in
but are normal to the loss-free connections. The top collisions with the metal lattice. Thus the electrons act
and bottomof the resistor carry positive and negative as mediating agents between the source of electro-
charge, respectively magnetic energy and the thermalenergy content of the
resistor. They can be regarded as converting potential
using equations (6)and (7). Hence the usual formula energy into thermal energy in the resistor, but they do
for the Joulean dissipation in the resistor is recovered. not strictly carry the potential energy from the battery
A schematic representationof the electric and themselves during their relatively slow drift along the
magnetic field lines and the Poynting vector directions transmission line. Instead the energytravelsoutside
in the vicinity of the resistor is shownin figure 2. the actual conductors,with the velocity of light for the
Further elementary calculations utilising the internal surrounding medium. Conduction electrons in a
solenoidal magnetic field H‘ and E’,,, and assuming a circuit return to the battery with essentially the same
uniform current density, allow oneto show that a kinetic energy with which they left it: this behaviour
fraction d Y d 2 of the total power passes through any contrasts with that of a beam of electrons previously
cylindrical surface, concentric with the outer surface accelerated,say, with an electrongun, which on
and with diameter d’, leading to uniformpower striking a target experience a permanent loss of kinetic
dissipation throughout theresistor. energy. The latter is the mode by which the dissipated
A morecomplexconfigurationof the fields and energy is transported in this case.
charges is shown in figure 3, where the resistor forms
the central portionof a linear circuit connected across
the ends of a parallel-wire line. This configuration is References
similar to that realised if a resistor is connected to the Barnes T G 1965 Foundations of Electricity and Magnetism
ends of the parallel-plate line by loss-free wires. As (Boston: Heath) pp15941
before, a longitudinal electric field E’,, exists inside the Beiser A 1968 ModernPhysics: A n IntroductorySurvey
resistor, running between charges on the end faces of (Reading, Mass.: Addison-Wesley) pp8-11
the resistor. In addition to the charges on the ends, Bleaney B I and Bleaney B 1976 Electricity and Magnetism
3rd edn (Oxford: University Press) p234
with thiscircuitgeometry electric charges will also
Booker H G 1959 Approachto ElectricalScience (New
exist on the cylindrical surface of the resistor, varying York: McGraw-Hill) p189
smoothly from positive to negative along the length of Nelkon M and Parker P 1977 Advanced Level Physics 4th
the resistor in asymmetricalcircuit. The associated edn (London: Heinernann) pp7 10-2
external electric field lines form loops which run from Scott W T 1966 The Physics of Electricity and Magnetism
the positively charged regions to the negatively 2nd edn (New York: Wiley) pp554-6

304