The exact relation between the displacement current and the conduction

current: Comment on “Time-dependent generalizations of the

Biot–Savart and Coulomb laws,” by D. J. Griffiths and M. A. Heald
†Am. J. Phys. 59 „2…, 111–117 „1991…‡
José A. Herasa兲
Departamento de Ciencias Básicas, Universidad Autónoma Metropolitana, Unidad Azcapotzalco,
Av. San Pablo No. 180, Col. Reynosa, 02200, México D. F. México and Departamento de Física
y Matemáticas, Universidad Iberoamericana, Prolongación Paseo de la Reforma 880, México
D. F. 01210, México
共Received 6 January 2008; accepted 28 February 2008兲
I introduce the exact relation between the displacement current ⑀0⳵E / ⳵t and the ordinary current J.
I show that ⑀0⳵E / ⳵t contains a local term determined by the present values of J plus a nonlocal term
determined by the retarded values of J. The nonlocal term implements quantitatively the suggestion
made by Griffiths and Heald that the displacement current at a point is a surrogate for ordinary
currents at other locations. © 2008 American Association of Physics Teachers.
关DOI: 10.1119/1.2899735兴

Griffiths and Heald have posed and answered the follow-

ing question 关their “recurring question 共3兲”兴:1 “What rela-
tionship exists between the two sources in the Ampere–
⑀0
⳵E
⳵t
J
=− +
1
3 4␲
冕 ⬘冉d 3x
3R̂共R̂ · 关J兴兲 − 关J兴
R3
Maxwell law: 3R̂共R̂ · 关⳵J/⳵t兴兲 − 关⳵J/⳵t兴
+
R 2c

⵱ ⫻ B = ␮ 0J + ␮ 0⑀ 0
⳵E
⳵t
, 共1兲 +
R̂ ⫻ 共R̂ ⫻ 关⳵2J/⳵t2兴兲
Rc2
, 冊 共3兲

for instance, can one be derived from the other?” Their an-
swer was the following:1 “None. The two currents are inde-
pendent and both are needed, in general. J represents the
ordinary current at a point where the curl 关of B兴 is evaluated,
while the displacement current ⑀0⳵E / ⳵t is a surrogate for
ordinary currents at other locations.” The assumed lack of a
specific relation between ⑀0⳵E / ⳵t and J led to the claim that
the Biot–Savart law extended to include the displacement
current:
where the square brackets 关 兴 indicate that the enclosed quan-
tity is to be evaluated at the retarded time t⬘ = t − R / c. Equa-
tion 共3兲 provides the exact relation between the currents
⑀0⳵E / ⳵t and J and implies that the two currents are not in-
dependent. The first term represents a local part which exists
in the region where J is defined and depends on the present
values of J. The second term is a nonlocal part which ex-
tends over all space and depends on the retarded values of J.
In short, the local contribution to ⑀0⳵E / ⳵t is instantaneously
produced by J, and the nonlocal contribution to ⑀0⳵E / ⳵t is

冕
causally produced by J. It follows that the displacement cur-
␮0 共J + ⑀0⳵E/⳵t兲 ⫻ R̂
B= d 3x ⬘ , 共2兲 rent is produced by the present and retarded values of the
4␲ R2 ordinary current J. The nonlocal part of ⑀0⳵E / ⳵t shows ex-
plicitly how the displacement current at a point represents
the contribution from ordinary currents elsewhere. Equation
“…is not useful because it is self-referential: To calculate B
共3兲 implements quantitatively the suggestion made by Grif-
at a point, one must know E everywhere—but you can’t de-
termine E unless you already know B 共everywhere兲, because fiths and Heald that:1 “the displacement current at a point is
of Faraday induction 关ⵜ ⫻ E = −⳵B / ⳵t兴.” In Eq. 共2兲 the inte- a surrogate for ordinary currents at other locations.” These
authors also claimed that Eq. 共2兲 is self-referential. However,
gration is over all space and R̂ = R / R = 共x − x⬘兲 / 兩x − x⬘兩.
In this comment I derive the exact relation between the the substitution of Eq. 共3兲 into Eq. 共2兲 removes the self-
displacement current ⑀0⳵E / ⳵t and the current J. The exis- referential aspect of Eq. 共2兲. To calculate B by means of Eq.
tence of this relation allows us to answer in the affirmative 共2兲 we really need to know ⑀0⳵E / ⳵t and not necessarily E,
the question posed and answered negatively by Griffiths and and we can calculate ⑀0⳵E / ⳵t from J using Eq. 共3兲 with no
Heald.1 explicit reference to B.2
Let me first introduce this relation which we will derive in To obtain Eq. 共3兲 we first consider the generalized Biot–
the following: Savart law:3

592 Am. J. Phys. 76 共6兲, June 2008 http://aapt.org/ajp © 2008 American Association of Physics Teachers 592

Downloaded 26 Nov 2012 to 132.248.9.8. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
 B=
␮0
4␲
冕 ⬘冉d 3x
关J兴 ⫻ R̂ 关⳵J/⳵t兴 ⫻ R̂
R2
+
Rc
. 冊 共4兲 E=−
P
+
1
3⑀0 4␲⑀0
冕 ⬘冉d 3x
3R̂共R̂ · 关P兴兲 − 关P兴
R3
We next take the curl of Eq. 共4兲

再 冎
3R̂共R̂ · 关⳵P/⳵t兴兲 − 关⳵P/⳵t兴

冕 ⬘冉
+
␮0 关J兴 ⫻ R̂ R 2c

冊
⵱⫻B= d 3x ⵱⫻
4␲ R2

再 冎冊
R̂ ⫻ 共R̂ ⫻ 关⳵2P/⳵t2兴兲
+ . 共10兲
关⳵J/⳵t兴 ⫻ R̂ Rc2
+⵱⫻ . 共5兲
Rc We can apply Eq. 共10兲 to a point electric dipole possessing
a time varying dipole moment p共t兲 located at the point x0.

再 冎
The following identities are demonstrated in the Appendix:
The polarization density for such a dipole is given by
关J兴 ⫻ R̂ 3R̂共R̂ · 关J兴兲 − 关J兴 P共x , t兲 = p共t兲␦兵x − x0其. After substituting this polarization den-
⵱⫻ = sity into Eq. 共10兲 and integrating over all space, we find the
R2 R3
Hertzian electric field:

+
R̂共R̂ · 关⳵J/⳵t兴兲 − 关⳵J/⳵t兴
R 2c E=−
p共t兲
3⑀0
␦兵x − x0其 +
1
4␲⑀0 R3
冉
3R̂共R̂ · 关p兴兲 − 关p兴

冊
8␲
+ 关J兴␦共x − x⬘兲, 共6兲 3R̂共R̂ · 关ṗ兴兲 − 关ṗ兴 R̂ ⫻ 共R̂ ⫻ 关p̈兴兲
3 + + , 共11兲

再 冎
R 2c Rc2
关⳵J/⳵t兴 ⫻ R̂ 2R̂共R̂ · 关⳵J/⳵t兴兲
⵱⫻ = where R̂ = R / R = 共x − x0兲 / 兩x − x0兩, ṗ = ⳵p / ⳵t, and the square
Rc R 2c brackets 关 兴 now indicate that the enclosed quantity is to be
evaluated at the retarded time t⬘ = t − 兩x − x0兩 / c. The electric
R̂ ⫻ 共R̂ ⫻ 关⳵2J/⳵t2兴兲
+ , 共7兲 field in Eq. 共11兲 is usually presented without the delta func-
Rc2 tion term.
where ␦共x − x⬘兲 is the Dirac-delta function. If we insert Eq. The origin of the local term J / 3 in Eq. 共3兲 might seem
共6兲 关with its last term written as 4␲关J兴␦共x − x⬘兲 − 共4␲ / 3兲 intriguing. However, if we use the Panofsky and Phillips
⫻关J兴␦共x − x⬘兲兴 and Eq. 共7兲 into Eq. 共5兲, and integrate the formula for the retarded electric field,5

再 冕 ⬘冉
Dirac-delta terms, we obtain

冕 冉
1 关␳兴R̂ 2R̂共R̂ · 关J兴兲 − 关J兴
E= d 3x +
J 1 3R̂共R̂ · 关J兴兲 − 关J兴 4␲⑀0 R2 R 2c
ⵜ ⫻ B = ␮ 0J + ␮ 0 d 3x ⬘

冊
− +
3 4␲ R3
R̂ ⫻ 共R̂ ⫻ 关⳵J/⳵t兴兲
+ , 共12兲
3R̂共R̂ · 关⳵J/⳵t兴兲 − 关⳵J/⳵t兴 Rc2
+
R 2c

冊冎
we can show the origin of the J / 3 term. Equation 共12兲 is an
R̂ ⫻ 共R̂ ⫻ 关⳵ J/⳵t 兴兲2 2 equivalent expression of the generalized Coulomb law.3 The
+ . 共8兲 time derivative of Eq. 共12兲 yields
Rc2
The term within the parentheses 兵 其 is identified as the term
⑀0⳵E / ⳵t in Eq. 共1兲. From this identification we obtain Eq. 共3兲.
⑀0
⳵E 1
=
⳵t 4␲
冕 ⬘ d 3x
关⳵␳/⳵t兴R̂
R2

冕 ⬘冉
Equation 共3兲 can be applied to the polarization current
⳵P / ⳵t, where P is a given polarization density. The replace- 1 2R̂共R̂ · 关⳵J/⳵t兴兲 − 关⳵J/⳵t兴
+ d 3x
ment J → ⳵P / ⳵t in Eq. 共3兲 and the fact that ⳵ / ⳵t can be taken 4␲ R 2c

冊
outside the retardation parentheses and even outside the in-
R̂ ⫻ 共R̂ ⫻ 关⳵2J/⳵t2兴兲

再
tegral leads to
共13兲

冕 ⬘冉
+ .
Rc2
⳵E ⳵ P 1 3R̂共R̂ · 关P兴兲 − 关P兴
⑀0 = − + d 3x We use the continuity equation and write the first integral as
⳵t ⳵t 3 4␲ R3
共Appendix A兲

冕 ⬘冉
3R̂共R̂ · 关⳵P/⳵t兴兲 − 关⳵P/⳵t兴
+
R 2c 冕 关⳵␳/⳵t兴R̂ 4␲ 3R̂共R̂ · 关J兴兲 − 关J兴

冊冎
d 3x ⬘ =− J+ d 3x
R2 3 R3

冊
R̂ ⫻ 共R̂ ⫻ 关⳵2P/⳵t2兴兲
+ , 共9兲 R̂共R̂ · 关⳵J/⳵t兴兲
Rc2 + . 共14兲
R 2c
which implies that the electric field produced by the polar-
ization density is4 The insertion of Eq. 共14兲 into Eq. 共13兲 gives Eq. 共3兲 again.

593 Am. J. Phys., Vol. 76, No. 6, June 2008 Notes and Discussions 593

Downloaded 26 Nov 2012 to 132.248.9.8. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
 The existence of the J / 3 term in Eq. 共3兲 can be traced to
the first term in Eq. 共14兲. Thus, the origin of the J / 3 term is ⵱⫻ 再 关⳵J/⳵t兴 ⫻ R̂
Rc
冎
冉 冊冉 冊
attributable to time variations of the charge density and ulti-
mately to the continuity equation 共charge conservation兲. I R̂ R̂
invite readers to decide which of the two derivations of Eq. = 关⳵J/⳵t兴 ⵱ · − ⵱ · 关⳵J/⳵t兴
Rc Rc

冉 冊
共3兲 presented here they find more useful.
R̂ R̂
+ · ⵱ 关⳵J/⳵t兴 − 共关⳵J/⳵t兴 · ⵱兲 . 共A4兲
Rc Rc
ACKNOWLEDGMENTS
If we substitute the results

冉 冊
The author thanks David J. Griffiths and Mark A. Heald
for useful comments. The support of FICSAC is gratefully R̂ 1
acknowledged. ⵱· = 2 , 共A5a兲
Rc Rc

APPENDIX A: DERIVATION OF EQUATIONS (6), (7), R̂ · 关⳵2J/⳵t2兴

⵱ · 关⳵J/⳵t兴 = − , 共A5b兲
AND (14) c

冉 冊
Consider the expansion

再 冎
R̂ 关⳵2J/⳵t2兴
· ⵱ 关⳵J/⳵t兴 = − , 共A5c兲
关J兴 ⫻ R̂ Rc Rc2
⵱⫻
R2

再 冉 冊冎
= − ⵱ ⫻ 关J兴 ⫻ ⵱
1
R
共A1a兲 共关⳵J/⳵t兴 · ⵱兲
R̂
Rc
=−
2R̂共R̂ · 关⳵J/⳵t兴兲 − 关⳵J/⳵t兴
R 2c
, 共A5d兲

冉冊 冉冊
=− 关J兴ⵜ2
1
R
+⵱
1
R
⵱ · 关J兴
into Eq. 共A4兲, we obtain Eq. 共7兲.
The proof of Eq. 共14兲 goes as follows: If the continuity
equation and Eq. 共A2b兲 are substituted into the identity

− ⵱ 冉冉冊 冊 1
R
· ⵱ 关J兴 + 共关J兴 · ⵱兲 ⵱
1
R
. 冉冊 共A1b兲
关⵱⬘ · J兴 = ⵱⬘ · 关J兴 + ⵱ · 关J兴 and the resulting equation multiplied
by R̂ / R2, we obtain

It can be shown that6 关⳵␳/⳵t兴R̂ 共⵱⬘ · 关J兴兲R̂ R̂共R̂ · 关⳵J/⳵t兴兲

共A6兲
冉冊
2 =− +
1 R R2 R 2c
ⵜ2 = − 4␲␦共x − x⬘兲, 共A2a兲
R and hence

⵱ · 关J兴 = −
R̂ · 关⳵J/⳵t兴
c
, 共A2b兲 冕 d 3x ⬘
关⳵␳/⳵t兴R̂
R2
=− 冕 d 3x ⬘
共⵱⬘ · 关J兴兲R̂
R2

冉冉冊 冊
⵱
1
· ⵱ 关J兴 =
关⳵J/⳵t兴
, 共A2c兲
+ 冕 d 3x ⬘
R̂共R̂ · 关⳵J/⳵t兴兲
R 2c
. 共A7兲
R R 2c
The integral involving ⵱⬘ · 关J兴 can be transformed further by

冉冊
examining the components of the integrand
1 3R̂共R̂ · 关J兴兲 − 关J兴 4␲

再 冉 冊冎
共关J兴 · ⵱兲 ⵱ = − 关J兴␦共x − x⬘兲.
R R3 3 共⵱⬘ · 关J兴兲R̂i ⳵ ⳵ 1
共A2d兲 − =− 关J兴k
R 2
⳵xk⬘ ⳵xi⬘ R
To obtain Eq. 共A2d兲 we used the identity:7 3R̂i共R̂ · 关J兴兲 − 关J兴i 4␲ i
+ − 关J兴 ␦共x − x⬘兲,

冉冊
R3 3
⳵ 1 3R̂iR̂ j − ␦ij 4␲
= − ␦ij␦共x − x⬘兲, 共A3兲 共A8兲
⳵ x i⳵ x j R R3 3
where Eq. 共A3兲 has been used and summation over the index
where R̂i = Ri / R = 共xi − xi⬘兲 / 兩xi − xi⬘兩 and ␦ij is the Kronecker k is implied. The volume integral of the first term of Eq. 共A8兲
delta symbol. The substitution of Eq. 共A2兲 into Eq. 共A1兲 can be transformed into a surface integral with the aid of
leads directly to Eq. 共6兲. Gauss’ theorem, and hence vanishes assuming that the cur-
Consider the expansion rents are localized:

594 Am. J. Phys., Vol. 76, No. 6, June 2008 Notes and Discussions 594

Downloaded 26 Nov 2012 to 132.248.9.8. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
 冕 d 3x ⬘
⳵
⳵xk⬘
再
关J兴k
⳵ 1
⳵xi⬘ R
冉 冊冎 冖 = 再
dS⬘ · 关J兴
⳵ 1
冉 冊冎
⳵xi⬘ R
= 0. 3
over all space, the integrand of which is itself a series of integrals over all
space兲 and therefore it is not generally useful in practice.
O. D. Jefimenko, Electricity and Magnetism 共Electrect Scientific Co., Star
City, WV, 1989兲, 2nd ed., p. 516; J. D. Jackson, Classical Electrodynam-
共A9兲
ics 共Wiley, New York, 1999兲, 3rd ed., Sec. 6.5.
4
The volume integral of the last term in Eq. 共A8兲 gives J. A. Heras, “Unified model based on U共1兲 duality symmetry of polariza-
−共4␲ / 3兲J. Therefore, tion and magnetization,” Phys. Rev. E 58, R6951–R6954 共1998兲. Equa-
tion 共9兲 implies Eq. 共10兲 with an additive integration term E0共x兲. Because

− 冕 d 3x ⬘
共⵱⬘ · 关J兴兲R̂
R2
Eq. 共10兲 satisfies a wave equation, the term E0共x兲 is ultimately a solution
of the Laplace equation. This additive term is shown to vanish under the
usual assumption that the electric field vanishes at infinity.

冕 ⬘冉 冊
5
W. K. H. Panofsky and M. Phillips, Classical Electricity and Magnetism
4␲ 3R̂共R̂ · 关J兴兲 − 关J兴 共Addison–Wesley, Reading, MA, 1962兲, 2nd ed. See also K. T. Mc-
=− J+ d 3x . 共A10兲 Donald, “The relation between expressions for time-dependent electro-
3 R3 magnetic fields given by Jefimenko and by Panofsky and Phillips,” Am.
From Eqs. 共A7兲 and 共A10兲 we obtain Eq. 共14兲. J. Sci. 65, 1074–1076 共1997兲 and J. A. Heras, “New approach to the
classical radiation fields of moving dipoles,” Phys. Lett. A 237, 343–348
共1998兲.
a兲 6
Electronic mail: heras@phys.lsu.edu The reader can find an approach for deriving identities involving retarded
1
D. J. Griffiths and M. A. Heald, “Time-dependent generalizations of the quantities in J. A. Heras, “Can Maxwell’s equations be obtained from the
Biot–Savart and Coulomb Laws,” Am. J. Sci. 59, 111–117 共1991兲. continuity equation?,” Am. J. Sci. 75, 652–657 共2007兲.
2 7
The nonself-referential form of Eq. 共2兲 关obtained by substituting Eq. 共3兲 C. P. Frahm, “Some novel delta-function identities,” Am. J. Sci. 51,
into Eq. 共2兲兴 implies complicated integrations 共it is defined as an integral 826–829 共1983兲.

Erratum: “Resource Letter PSNAC-1: Physics and society: Nuclear arms

control” †Am. J. Phys. 76 „1…, 5–14 „2008…‡
Alexander Glaser and Zia Mian
Program on Science and Global Security Woodrow Wilson School of Public and International Affairs
Princeton University, Princeton, New Jersey 08544
共Received 21 January 2008; accepted 15 March 2008兲
关DOI: 10.1119/1.2906230兴

The title of Ref. 10 should read “Nuclear Choices: A Citizen’s Guide to Nuclear Technology,” and the date of publication of
the revised edition should be 1993.
We are grateful to Richard Wolfson for pointing out our mistake.

Erratum: “Two examples of circular motion for introductory courses

in relativity” †Am. J. Phys. 75 „12…, 1123–1133 „2007…‡
Stephanie Wortel and Shimon Malin
Department of Physics, Colgate University, Hamilton, New York 13346
Mark D. Semon
Department of Physics, Bates College, 44 Campus Avenue, Lewiston, Maine 04240
共Received 31 January 2008; accepted 23 March 2008兲
关DOI: 10.1119/1.2908190兴
David Cornell1 has pointed out that our discussion of the
resolution of the triplet paradox in the second paragraph of
Sec. II includes the unfortunately worded sentence: “Just as
in the standard twin paradox, when each of the traveling
triplets compares their clock with the clock on Earth, each
will agree that their clock is running slower than the Earth
clock by the same factor of ␥.”2 This sentence might seem to
imply incorrectly that the oppositely traveling triplets will
find, during the trip, that their clocks are running at a com-
mon rate that is slower than that of the Earth clocks.
The sentence should have read “Just as in the standard
twin paradox, when the traveling triplets return to Earth, they
will both find that less time has elapsed on each of their
clocks than on the Earth clock by the same factor of ␥.”
Rather than “resolve” the triplet paradox by explaining how
the travelers themselves understand the result, we have sim-
ply examined the situation from the Earth observer’s point of
view. We thank David Cornell for bringing this confusion to
our attention.
Appendix B of our paper resolves the triplet and circular
paradoxes between the Earth and traveling observers, and the
paper of Cranor, Heider, and Price3 resolves the paradox be-
tween the two travelers in the circular twin case. The method
of Ref. 2 also can be used to resolve the paradox between the
two travelers in the triplet case.
1
David Cornell 共private communication兲, 6 January 2008.
2
S. Wortel, S. Malin, and M. D. Semon, “Two examples of circular motion
for introductory courses in relativity,” Am. J. Phys. 75, 1123–1133
共2007兲.
3
M. B. Cranor, E. M. Heider, and R. H. Price, “A circular twin paradox,”
Am. J. Phys. 68, 1016–1020 共2000兲.

595 Am. J. Phys. 76 共6兲, June 2008 http://aapt.org/ajp © 2008 American Association of Physics Teachers 595

Downloaded 26 Nov 2012 to 132.248.9.8. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission