Вы находитесь на странице: 1из 13

Resource Letter EM-1: Electromagnetic Momentum

David J. Griffiths

Citation: Am. J. Phys. 80, 7 (2012); doi: 10.1119/1.3641979


View online: http://dx.doi.org/10.1119/1.3641979
View Table of Contents: http://ajp.aapt.org/resource/1/AJPIAS/v80/i1
Published by the American Association of Physics Teachers

Additional information on Am. J. Phys.


Journal Homepage: http://ajp.aapt.org/
Journal Information: http://ajp.aapt.org/about/about_the_journal
Top downloads: http://ajp.aapt.org/most_downloaded
Information for Authors: http://ajp.dickinson.edu/Contributors/contGenInfo.html

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
RESOURCE LETTER
Resource Letters are guides for college and university physicists, astronomers, and other scientists to literature, websites, and other teaching aids.
Each Resource Letter focuses on a particular topic and is intended to help teachers improve course content in a specific field of physics or to
introduce nonspecialists to this field. The Resource Letters Editorial Board meets at the AAPT Winter Meeting to choose topics for which Resource
Letters will be commissioned during the ensuing year. Items in the Resource Letter below are labeled with the letter E to indicate elementary level or
material of general interest to persons seeking to become informed in the field, the letter I to indicate intermediate level or somewhat specialized
material, or the letter A to indicate advanced or specialized material. No Resource Letter is meant to be exhaustive and complete; in time there may
be more than one Resource Letter on a given subject. A complete list by field of all Resource Letters published to date is at the website http://
ajp.dickinson.edu/Readers/resLetters.html. Suggestions for future Resource Letters, including those of high pedagogical value, are welcome and
should be sent to Professor Roger H. Stuewer, Editor, AAPT Resource Letters, School of Physics and Astronomy, University of Minnesota, 116
Church Street SE, Minneapolis, MN 55455; e-mail: rstuewer@physics.umn.edu

Resource Letter EM-1: Electromagnetic Momentum


David J. Griffiths
Department of Physics, Reed College, Portland, Oregon 97202
(Received 29 July 2011; accepted 1 September 2011)
This Resource Letter surveys the literature on momentum in electromagnetic fields, including the
general theory, the relation between electromagnetic momentum and vector potential, hidden
momentum, the 4/3 problem for electromagnetic mass, and the AbrahamMinkowski controversy
regarding the field momentum in polarizable and magnetizable media. VC 2012 American Association of
Physics Teachers.
[DOI: 10.1119/1.3641979]

I. INTRODUCTION But locating this hidden momentum can be subtle


and difficult.
According to classical electrodynamics, electric and mag- (3) A moving charge drags around the momentum in its
netic fields (E and B) store linear momentum, which must fields, which means (in effect) that it has extra mass.
be included if the total momentum of a system is to be con- But this electromagnetic mass is inconsistent with what
served. Specifically, the electromagnetic momentum per unit you get from the energy in the fields (using Einsteins for-
volume is mula E mc2)by a notorious factor of 4/3, in the case
of a spherical shell. Which mass (if either) is correct?
g 0 E  B; (1)
(4) Inside matter, which is subject to polarization and mag-
as first proposed by Poynting (Refs. 3032). Field momen- netization, the effective field momentum is modified.
tum is most dramatically demonstrated in the laboratory by Minkowski proposed
the pressure of light on an absorbing or reflecting surface. [In
1619 Kepler suggested that the pressure of light explains gM D  B; (2)
why comet tails point away from the sun (Ref. 29). The
theory was developed by Maxwell (Ref. 10) and confirmed Abraham advocated
experimentally by Lebedew (Ref. 25) and Nichols and Hull 1
(Ref. 28). Some introductory textbooks offer a quick qualita- gA E  H: (3)
c2
tive explanation for the pressure exerted on a perfect conduc-
tor: E drives charge in (say) the x direction and B (in the y For over a century a debate has raged: which expression is
direction) then exerts a force in the z direction. This naive right? Or are they perhaps both right, and simply describe
argument is faulty (Refs. 27, 33, and 16).] different things? How can the question be settled, theoreti-
But the notion that fields carry momentum leads to several cally and experimentally? Although many distinguished
intriguing problems, some of which are not entirely resolved authors claim to have resolved the issue, the dispute contin-
after more than a century of debate. ues to this day.
(1) For a point charge q in an external field represented by In Section II, I summarize the theory. I then survey each
the vector potential A, the electromagnetic momentum is of the four controversies described qualitatively above. In
qA. This suggests that A can be thought of as potential the final section, I briefly consider electromagnetic momen-
momentum per unit charge, just as the scalar potential tum in quantum mechanics, where the photon makes the
V is potential energy per unit charge. But this interpre- story in some respects more concrete and intuitive.
tation raises questions of its own, and it has never been
universally accepted. II. THEORY
(2) According to Eq. (1), even purely static fields can
A. Nonrelativistic
store momentum. How can a system at rest carry mo-
mentum? It cannot there must be some compensat- Electrodynamics (Refs. 2 and 6) is based on Maxwells
ing non-electromagnetic momentum in such systems. equations, which tell us how the sources (charge density q

7 Am. J. Phys. 80 (1), January 2012 http://aapt.org/ajp C 2012 American Association of Physics Teachers
V 7

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
and current density J) generate electric and magnetic fields The corresponding statement for electromagnetic energy is
(E and B):
@u
1 @B r  S E  J: (14)
r  E q; rE ; @t
0 @t
This is Poyntings theorem; E  J is the power per unit vol-
@E
r  B 0; r  B l0 J l0 0 ; (4) ume delivered by the fields to the electric charges. Except in
@t regions where E  J 0 (empty space, for example) the elec-
and the Lorentz force law, which tells us the force exerted by tromagnetic energy by itself is not conserved, because the
the fields on a point charge q moving with velocity v: fields do work on the charges. Similarly,

F qE v  B: (5) @g $
 r  > qE J  B: (15)
@t
The homogeneous Maxwell equations (the two that do not
involve q or J) allow us to express the fields in terms of sca- Here (qE J  B) is the force per unit volume exerted by
lar and vector potentials the fields on the electric charges. Except (for example) in
empty space, electromagnetic momentum by itself is not
@A conserved. [Nor, therefore, is mechanical momentum sepa-
E rV  ; B r  A: (6) rately conserved. This means that Newtons Third Law
@t
(although it holds in electrostatics and magnetostatics) is not
Electromagnetic fields store energy and momentum (and obeyed in electrodynamics (Refs. 22 and 2).]
for that matter also angular momentum). The energy per unit As we shall see, it is no accident that the same quantity
volume in the fields is (E  B) appears in the Poynting vector and in the momentum
  density (Ref. 1),
1 2 1 2
u 0 E B ; (7) S c2 g (16)
2 l0 $
(or that the same quantity > plays a dual role as force-per-
and the momentum density is unit-area and momentum flux).
g 0 E  B: (8)
B. Relativistic
The fields also transport energy and momentum from one
place to another. The energy flux (energy per unit time, per 1. Notation
unit area) is given by the Poynting vector, The theory is more elegant in covariant (relativistic) nota-
1 tion. The (Cartesian) space-time coordinates are xl (ct, x,
S E  B (9) y, z), Greek indices run from 0the temporal coordi-
l0 nateto 3, while Roman indices go from 1 to 3the
spatial coordinates. We use the metric
(S  da is the energy per unit time transported through a 0 1
window of area da). The momentum flux is related to the 1 0 0 0
Maxwell stress tensor: B 0 1 0 0 C
    gl B@ 0 0 1 0 A
C (17)
1 2 1 1 2
Tij 0 Ei Ej  dij E Bi Bj  dij B (10) 0 0 0 1
2 l0 2

(Specifically, the momentum per unit time transported and the Einstein convention (sum repeated indices). The
$ energy density u, the energy flux S, the momentum density
through a window da is >  da ). For example, the energy $
and momentum per unit time radiated (to infinity) by a non- g, and the stress tensor > go together to make the stress-
relativistic point charge q are energy tensor:
 
dE l0 q2 2 u S=c
a ; (11) Hl $ (18)
dt 6pc cg >
dp l0 q2 2 $
a v; (12) This is entirely generalin other contexts u, S, g, and > will
dt 6pc3 not have their electromagnetic form [Eqs. (7)(10)]. If the
where v is the velocity of the charge and a is its acceleration. stress-energy tensor is divergenceless:
[The uniqueness of these expressions [Eqs. (7)(10)] is open
@l Hl 0 (19)
to some question (Ref. 23), but I shall take them as
definitions.] then
Several conservation laws follow from Maxwells equa-
tions. Local conservation of charge is expressed by the conti-
nuity equation, pl  H0l d 3 r (20)

@q transforms as a four-vector [this is sometimes called von


r  J 0: (13)
@t Laues theorem (Refs. 24, 26, and 7)], and the total energy

8 Am. J. Phys., Vol. 80, No. 1, January 2012 David J. Griffiths 8

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
and momentum (E cp0 and p) are conserved. If the stress a conserved four-vector. And (as always) the complete
tensor is symmetric (Hl Hl ), then angular momentum is stress-energy tensor,
also conserved (Ref. 11). In a well-formulated theory the
complete stress-energy tensor is always divergenceless and Hl T l Hl
o ; (31)
symmetric, but this may not be true for individual portions
of it, such as the electromagnetic contribution alone. (where Hl
o is the non-electromagnetic contribution) is diver-
genceless (and symmetric).
The energy/momentum radiated by a point charge q is
2. Electrodynamics
The charge and current densities combine to form a four- dpl l0 q2 
a a gl ; (32)
vector: dt 6pc3
J l cq; Jx ; Jy ; Jz ; (21) where gl : dxl/ds is the four-velocity and al : dgl/ds is
the four-acceleration (ds is the proper time).
the fields constitute an antisymmetric tensor:
0 1 III. MOMENTUM AND VECTOR POTENTIAL
0 Ex Ey Ez
B Ex 0 cBz cBy C
Fl B@ Ey cBz
C; (22) For a localized configuration the total momentum in the
0 cBx A fields is
Ez cBy cBx 0
p g d3 r 0 E  B d 3 r: (33)
and the potentials make a four-vector

Al V=c; Ax ; Ay ; Az : (23) [I assume all fields go to zero sufficiently rapidly at infinity


that the integrals converge, and surface terms can be
The inhomogeneous Maxwell equations read neglected. It is notoriously dangerous to speak of the mo-
mentum (or energy) of a configuration that is not localized in
1 
@l Fl J (24) space, and when I refer to a uniform field this should
0 c always be interpreted to mean locally uniform, but going to
zero at infinity.] In the static case two equivalent expressions
(where @ l is short for @/@xl). The homogeneous Maxwell can be obtained, by writing either E or B in terms of the
equations are enforced by the potential representation potentials (E rV, or B r  A with r  A 0) (Refs.
Fl @ l A  @  Al : (25) 51, 47, 38, and 61):

The electromagnetic stress-energy tensor is: p qA d3 r; (34)
0 1
u Sx =c Sy =c Sz =c
B cgx Txx Txy Txz C 1
p VJ d 3 r: (35)
T l B@ cgy Tyx Tyy Tyz A:
C (26) c2
cgz Tzx Tzy Tzz In particular, the electromagnetic momentum of a stationary
l point charge q, in a magnetic field represented by the vector
In view of Eq. (16), T is symmetric; in terms of the fields: potential A, is
 
1 p qA: (36)
T l 0 glj Fjk Fk gl Fjk Fjk : (27)
4
This suggests that A can be interpreted as potential
The continuity equation becomes the statement that Jl is momentum per unit charge, just as V is potential energy per
divergenceless: unit charge.
The association between momentum and vector potential
@l J l 0: (28) goes back to Maxwell, who called A electromagnetic
momentum (Ref. 41; p. 481) and later electrokinetic
The electromagnetic stress-energy tensor is not by itself momentum (Ref. 10; Art. 590), and Thomson (Ref. 21). But
divergenceless; from Maxwells equations it follows that the idea did not catch on; any physical interpretation of A
1 was disparaged by Heaviside and Hertz (Refs. 34 and 36),
@l T l Fj Jj : (29) who regarded A as a purely mathematical device. So genera-
c
tions of teachers were left with no good answer to their stu-
Ordinarily, therefore, electromagnetic energy and dents persistent question: What does the vector potential
momentum represent, physically? Few were satisfied by the safe but
unilluminating response, It is that function whose curl is B
(Ref. 39). From time to time the connection to momentum
p0em T 00 d 3 r and piem T 0i d3 r; (30) was rediscovered [by Calkin (Ref. 35), for example], but it
was not widely recognized until Konopinskis pivotal paper
do not constitute a four-vector, and they are not conserved. (Ref. 40). Konopinski was apparently unaware of the histori-
However, if Jj 0 (for instance, in empty space) then plem is cal background, which was supplied by Gingras (Ref. 37).

9 Am. J. Phys., Vol. 80, No. 1, January 2012 David J. Griffiths 9

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
Many modern authors follow Konopinskis lead, culminating 1
in what remains (to my mind) the definitive discussion by p E  m: (43)
c2
Semon and Taylor (Ref. 42).
An obvious objection is that the vector potential is not [This is for a conventional magnetic dipolea tiny current
gauge invariant, and different choices yield different loop. If the dipole is made of hypothetical magnetic
momenta. Semon and Taylor point out that generalized mo- monopoles, the momentum is zero (Ref. 52).]
mentum itself is ambiguous: canonical momentum, for An ideal (point) electric dipole pe in an external magnetic
instance, does not always coincide with ordinary (kinetic) field B (Ref. 61):
momentum. In any event, Eq. (34) holds only for static
fields. But in truth, the same objections could be raised 1
p B  pe : (44)
against the interpretation of V as potential energy per unit 2
charge. [A point charge at rest can be represented by the
potentials V (r, t) 0, Ar; t 1=4p0 qt=r2 ^r, and while
A sphere (radius R) carrying a surface charge r k cos h,
no physicist in her right mind would choose to do so, the fact where k is a constant and h is the polar angle with respect
remains that the physical meaning of V depends on the to the z axis (its electric dipole moment is
gauge. Moreover, if the fields are time dependent, the work pe 4=3pR3 k ^z). It also carries a surface current
done to move a charge is no longer qDV in any gauge.] K k0 sin h0 /^0 , where k0 is another constant and h0 , /0 are
Another way to get at the association between momentum the polar and aximuthal angles with respect to the z0 axis
and potential is afforded by the Lagrangian formulation of (its magnetic dipole moment is m 4=3pR3 k0 ^z0 ). The mo-
electrodynamics. For a nonrelativistic particle of mass m and mentum in the fields is (Ref. 63)
charge q, moving with velocity v through fields described by l0
the potentials V (r, t) and A(r, t) (Ref. 5; Sec. 4.9) p m  pe (45)
4pR3
1 A charged parallel-plate capacitor (field E, volume V) in a
Lr; v; t mv2 qv  A  qV: (37)
2 uniform magnetic field B (Refs. 61, 80, 57, and 43):
The generalized momentum pi dL=dq_ i (Ref. 20) is 1
p 0 EBV: (46)
p mv qA; (38) 2

the sum of a purely kinetic part (mv) and an electromagnetic [Ordinarily, electromagnetic momentum (like electromag-
part (qA). The Hamiltonian p  v  L is netic energy), being quadratic in the fields, does not obey the
superposition principle (that is, the momentum of a compos-
1
H p  qA2 qV: (39) ite system is not the sum of the momenta of its parts, consid-
2m ered in isolation). However, if static charges are placed in an
external magnetic field, the momentum is linear in the elec-
It differs from the free particle Hamiltonian (H p2/2m) by
tric field they produce, and hence the total momentum is the
the substitution
sum of the individual momenta. Thats how McDonald (Ref.
p ! p  qA; H ! H  qV: (40) 61) discovered the (surprising) factor of 1/2 in Eq. (46),
which is due to momentum in the fringing fields.]
This is the so-called minimal coupling rulean efficient Now, there is a very general theorem in relativistic field
device for constructing the Hamiltonian of a charged particle theory that says
in the presence of electromagnetic fields. It is equivalent to
the Lorentz force law and is especially useful in quantum If the center of energy of a closed system is at
mechanics (Ref. 14, Sec. 6.8). In the relativistic theory (Ref. rest, then the total momentum is zero.
6, Sec. 12.1) the generalized four-momentum is
[Center of energy is the relativistic generalization of cen-
pl mgl qAl ; (41) ter of mass, but it takes account
of all forms of energy, not
just rest energy: ru d3 r= u d3 r.] This certainly seems rea-
and minimal coupling becomes (Ref. 3; p. 360) sonable; a heuristic argument is given by Calkin (Ref. 47),
and a more formal proof by Coleman and Van Vleck (Ref.
pl ! pl  qAl : (42) 48). In the configurations described above the center of energy
is clearly at rest, so if there is momentum in the fields there
Thus, relativity reinforces the notion that if V is (potential) must be compensating non-electromagnetic momentum some-
energy per unit charge, then A is (potential) momentum per where else in the system. But it is far from obvious where this
unit charge, drawing a parallel between the four-vectors hidden momentum resides, or what its nature might be.
pl (E/c, p) and Al (V/c, A). Curiously, the phenomenon of hidden momentum was not
noticed until the work of Shockley and James (Ref. 64) and
IV. HIDDEN MOMENTUM Costa de Beauregard (Ref. 50), in 1967. It was picked up im-
mediately by Haus and Penfield (Ref. 53), Coleman and Van
Even purely static electromagnetic fields can harbor mo- Vleck (Ref. 48), Furry (Ref. 51), and eventually by many
mentum (Eqs. (33)(35)). Configurations that have been others (Refs. 47, 65, 49, and 54). Indeed, the subject remains
studied include an active area of research to this day (Refs. 55, 58, and 59).
An ideal (point) magnetic dipole m in an external electric The simplest model for hidden momentum was suggested
field E (Ref. 51): by Calkin (Ref. 47) (or Ref. 2; Example 12.12); it consists of

10 Am. J. Phys., Vol. 80, No. 1, January 2012 David J. Griffiths 10

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
a steady current loop in an external electric field. The current If the sphere is now set in motion, at a constant nonrelativis-
is treated as a stream of free charges, speeding up and slow- tic velocity v, the momentum in its electromagnetic fields is
ing down in response to the field. [Because the current is the (Refs. 18, 74, 69, 75, and 70)
same all around the loop, in segments where the charges
are moving more rapidly they are farther apart.] Each 2Q2
p v (49)
charge carries a (relativistic) mechanical momentum
q 3Rc2
mv= 1  v=c2 , andeven though the loop is not mov- from which we infer that there is an electromagnetic contri-
ing and the current is steadythese momenta add up to a bution to its mass in the amount
total that exactly cancels the electromagnetic momentum.
Others have noted that this is an artificial model for the 2Q2 4
mem2 2
mem1 : (50)
current, and Vaidman (Ref. 65) considered two more realis- 3Rc 3
tic models, an incompressible fluid, and a metal wire. The
The momentum-derived mass is greater than the energy-
former carries mechanical momentum because of the
derived mass, by an infamous factor of 4/3 (Refs. 73, 71, and
remarkable (relativistic) fact that a moving fluid under pres-
72). This ratio holds for all spherically-symmetrical charge
sure has extra momentum (Ref. 56), whereas the latter,
configurations; other geometries yield different factors (Ref.
because of induced charges on the surface of the wire, has no
70).
momentum in the fields (and no hidden momentum to cancel
The underlying source of the discrepancy has been known
it). [This applies as well to the examples above; to be safe,
for over a century. Poincare (Ref. 76) pointed out that a
we assume that all charges are glued to nonconductors, and
charged sphere is unstable (it would explode, from the mu-
the magnetic fields are produced by charged nonconductors
tual repulsion of its parts), unless some other force is pro-
in motion (Refs. 51, 65, and 57).]
vided, to hold it together. This stabilizing force (whatever its
Hidden momentum has nothing to do, really, with electrody-
nature) has come to be known as Poincare stress, and it too
namics, except that it was first discovered in this context. The
contributes to the energy and momentum of the object.
name itself is perhaps unfortunate, since it sounds mysterious,
When the two contributions are combined, the inconsistency
and a definitive characterization of hidden momentum remains
disappears. [A closely related problem is Rindlers para-
elusive. This much seems clear: it is mechanical, relativistic,
dox (Ref. 77).]
and occurs in systems that are at rest, but have internally moving
More formally, the problem is that the electromagnetic
parts. [Actually, there is no reason the system has to be at rest,
stress-energy tensor is not divergenceless in the presence of
but the phenomenon is much more striking in that case, and
charge and current [Eq. (29)], and as a result the integral
there has not been much discussion of hidden momentum in
moving configurations. Similarly, there exists in principle hid-
l
den angular momentum, but since there is no rotational analog pem T 0l d 3 r (51)
to the center-of-energy theorem, it is less intriguingindeed,
many examples are known in which nothing is rotating, and yet does not constitute a four-vector. It is only the complete
the fields carry angular momentum with no compensating hid- stress-energy tensor (which in this instance would include a
den angular momentum. The extreme example is the Thomson contribution from the Poincare stress) that is divergenceless,
dipole, consisting of a magnetic monopole and an electric and its integral does yield a genuine (and conserved) four-
charge (Ref. 46).] It is hidden only in the sense that it is sur- vector (Eq. (20)).
prising and unexpected, but it is perfectly genuine momentum. In the early years of the 20th century, when Abraham (Ref.
[To this day some authors remain skeptical (Ref. 44); Mansuir- 66), Lorentz (Ref. 9), Schott (Ref. 19), and others dreamed of
pur (Refs. 126 and 60) calls hidden momentum absurdity.] producing a purely electromagnetic model of the electron
Not every case of momentum in static fields involves hid- (Ref. 79), there was a sense that the electromagnetic fields
den momentum. A long coaxial cable connected to a battery ought to yield an energy-momentum four-vector all by
at one end and a resistor at the other carries electromagnetic themselves, and this notion has persistedeven in the face of
momentum (Ref. 2; Example 8.3), but no hidden momentum. Poincares (to my mind decisive) argument that they should
In this case the center of energy is not at rest; energy is flow- not. Rohrlich (Refs. 78 and 18) proposed to retain the equation
ing from the battery to the resistor, and the associated momen- for electromagnetic energy in the particles rest frame, but
tum is precisely the momentum in the fields (Refs. 45 and 62). define the electromagnetic momentum by Lorentz transforma-
tion. This makes plem a four-vector by construction, but it
V. MOMENTUM AND MASS means Eq. (1) is no longer applicable in the presence of
The energy in the electric field of a uniformly charged sta- charges and currents. [In empty space, where Jl 0, the elec-
tionary spherical shell, of radius R and charge Q, is (Ref. 2; tromagnetic stress-energy tensor is divergenceless (Eq. (29)),
Example 2.8): and the problem does not arise. Thus, the application of Eq.
(1) to electromagnetic waves in vacuum has not, to my knowl-
Q2 edge, been challenged.] Rohrlichs suggestion was criticized
E : (47) by Tangherlini (Ref. 81), Boyer (Ref. 67), and others (Ref.
8p0 R 68), but it has left a residue of confusion as to the correct
According to Einsteins formula (E mc2) this means there expression for electromagnetic momentum (Ref. 80).
is an electromagnetic contribution to its mass, in the amount
VI. MOMENTUM IN MATTER
Q2 In the presence of materials subject to polarization (P) and
mem1 (48)
8p0 Rc2 magnetization (M) it is convenient to express the laws of

11 Am. J. Phys., Vol. 80, No. 1, January 2012 David J. Griffiths 11

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
   
electrodynamics in terms of free charges and free currents, 1 2 1 1 2
since these are the ones we directly control (Ref. 2; Sec. Tm ij  Ei Ej  dij E Bi Bj  dij B
2 l 2
7.3.5):    
1 1
Ei Dj  dij E  D Bi Hj  dij B  H : (60)
q qf qb ; J Jf Jb Jp : (52) 2 2

The bound charge, bound current, and polarization Comparison with Eq. (15) invites us to interpret
current are
gM  D  B (61)
@P
qb r  P; Jb r  M; Jp : (53) as the electromagnetic momentum in a linear medium. This
@t
was Minkowskis proposal (Ref. 129).
Introducing the auxiliary fields However, the resulting electromagnetic stress-energy ten-
sor has the form
1
D  0 E P; H B  M; (54)  
l0 um Sm =c
TM l $ ; (62)
cgM  >M 
Maxwells equations become
and Abraham pointed out that because Sm/c = cgM Minkow-
@B
r  D qf ; r  E  ; skis tensor is not symmetric, and hence does not conserve
@t angular momentum. He suggested instead that the electro-
@D magnetic momentum in matter is (Refs. 82, 83, and 84).
r  B 0; r  H Jf ; (55)
@t
1
If the medium is linear (as we shall assume from now on), D gA  E  H; (63)
c2
and H are related to E and B by the constitutive relations $
while um, Sm, and >m are unchanged (Ref. 11; pp. 204205).
1 This entails replacing Eq. (59) by
D E; H B; (56)
l
$ @
where  is the permittivity and l is the permeability of the r  >m  g qf E Jf  B f A ; (64)
@t A
material. [In non-linear mediasuch as ferromagnetsthe
work done depends not only on the initial and final states of with the extra Abraham force density (Refs. 94 and 128)
the system but also on how it was carried from one to the  
other, so the whole notion of stored energy loses its mean- 1 @
fA 1  2 D  B; (65)
ing (Ref. 17; Secs. 4.10 and 10.10).] n @t
Although the original formulas for electromagnetic energy p
(Eq. (7)) and energy flux (Eq. (9)) are still perfectly correct, where n  l=0 l0 is the index of refraction of the me-
they are not very useful in this context: (0/2)E2, for instance, dium. [Most studies concentrate on dielectric materials, with
is the work it would take to bring in all the charges (free and l l0; some treat purely magnetic materials ( 0). I will
bound) from infinity and nail them down in their final loca- keep the discussion general, whenever possible, though I do
tionsbut it does not include the work required to stretch all assume that dispersion can be neglected (i.e.,  and l do not
the atomic springs to which the bound charges are attached. depend significantly on frequency).]
A more useful quantity is the work done on the free charges Thus, began a debate that has raged for more than a cen-
alone, as we bring them in from infinity, with the bound tury, between Minkowskis momentum and Abrahams.
charges (and the springs) responding however they do. The Scores of theoretical papers have claimed to settle the issue in
resulting electromagnetic energy density in matter is (Ref. 2; favor of one or the other; experiments have been performed
Problem 8.15), (Ref. 6; Sec. 6.7), (Refs. 8, 98, 113, and 114) with unambiguous (but contradictory) results. In recent years
the dispute has been particularly intense, because of the criti-
  cal importance of optical forces in nanotechnology (Refs. 85,
1 1 1
um E2 B2 E  D B  H; (57) 97, 106, and 29).
2 l 2
How would one go about determining the momentum of
the electromagnetic fields in a medium? The obvious test
and the corresponding flux (the Poynting vector) is would be the pressure of light on the interface between two
1 transparent media, or on a mirror embedded in the material.
Sm E  B E  H: (58) Conservation of momentum should dictate the correct for-
l mula. Imagine a wave packet in vacuum, with total energy U
and momentum p (Fig. 1(a)),
In the same spirit we might calculate the force per unit
volume on the free charges in the material:    
1 2 E2 U
U 0 E AL; p 0 AL (66)
$ @ 2 2c c
fm qf E Jf  B r  >m  D  B; (59)
@t
(here E is the amplitude of the electric field, E/c is the ampli-
where (Ref. 10; Art. 641), (Ref. 13; Chapter X) tude of the magnetic field, and we average over a full cycle).

12 Am. J. Phys., Vol. 80, No. 1, January 2012 David J. Griffiths 12

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
Fig. 2. Pulse 1.
Fig. 1. A pulse of light in vacuum (a) and in a medium (b). M U  v
pbar 1 : (71)
The same packet (with the same energy) travels more slowly s c c
(v c/n) in a transparent medium (Fig. 1(b)): The total momentum before the pulse enters the bar (and
8  2
> E v U also after it exits) is p U/c; conservation requires that this
  >
> p  AL np;
1 2 v < M
2v c v equal the momentum of the pulse while inside the bar (pin)
U E AL ;   plus the momentum of the bar itself,
2 c > > 1 E 2
v vU p
>
: pA 2 AL 2 : p pin pbar ; (72)
c l 2v c c n
(67) and hence
The momentum of the packet is increased (by a factor of n), v p
pin p : (73)
according to Minkowski, whereas according to Abraham it c n
decreases by the same factor (Refs. 137, 117, 135, and 89). It Score one for Abraham (compare Eq. (67)).
should be very easy to discriminate. Example 2. Imagine a monochromatic plane wave, propa-
Example 1. In 1953 Balazs (Ref. 87) proposed the follow- gating in the z direction through a homogeneous linear mate-
ing (now classic) thought experiment. Start with a bar of rial, that encounters a perfectly reflecting surface at z 0.
transparent material, its two ends coated to prevent reflec- The incident and reflected fields are (Ref. 2; Sec. 9.3)
tions. Imagine two identical pulses of light, each of energy U
(the pulses are very short, compared to the length of the bar). EI z; t E0 coskz  xt ^x;
Pulse (1) passes through the bar (Fig. 2); pulse (2) passes
just outside the bar (Fig. 3). The bar itself is free to move, E0
BI z; t coskz  xt ^y
but relatively heavy Mc2  U , so it can absorb momentum v
from pulse (1) without acquiring significant kinetic energy. ER z; t E0 coskz  xt ^x;
Because there are no external forces, the center of energy E0
(X) of the system (pulse plus bar) moves at a constant rate, BR z; t coskz  xt ^y;
v
and at any given time is at the same location in both cases.
During the time s that pulse (1) spends inside the bar, the (for z 0, and zero for z > 0). The discontinuity in B deter-
center of energy of system (2) moves a distance mines the surface current:
1
DX2 Ucs: (68) 2E0
U Mc2 K cosxt ^x: (74)
lv
Pulse (1) does not move as far (because it travels at a
reduced speed in the medium), so the bar itself must move a The force on an area A of surface is
(small) distance  to make up for it.
2E20 A
1 F rE K  BA cos2 xt ^z
DX1 Mc2  Uw L : (69) lv2
U Mc2
2E20 A cos2 xt ^z (75)
Equating the two (DX2 D X1), we find that
(since B is discontinuous at z 0, we use the average). This
U U
 cs  w  L  c  vs (70) is the momentum per unit time imparted to the mirror (no
2
Mc U Mc2 energy is delivered, if we hold the mirror stationary). Mean-
(note that vs  w L). Evidently the momentum of the bar, while, the momentum per unit time dumped by the electro-
during the time pulse (1) is inside, is magnetic wave, as it reverses direction, is

8  2 
>
> E cos2 xt
< 2  0 Av 2E20 A cos2 xt ^z Minkowski
dp v
  (76)
dt > > 1 E2 cos2 xt 2
:2 2 0 Av 2 E20 A cos2 xt ^z Abraham:
c l v n

13 Am. J. Phys., Vol. 80, No. 1, January 2012 David J. Griffiths 13

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
One would like to write down, once and for all, the com-
plete and correct total stress-energy tensorelectromagnetic
plus mechanical. Unfortunately, this depends on the detailed
nature of the material (Refs. 127 and 132), and in realistic
theories can be forbiddingly complicated.
What exactly do the two momenta represent, physically,
and why is it that some contexts seem to favor the one or the
other (Ref. 133)? Recently, several authors (Refs. 90, 104,
117, 107, 91, and 101) and especially Barnett (Refs. 88 and
89) have noted that there is a parallel ambiguity between the
Fig. 3. Pulse 2. kinetic momentum (mv) of a particle and its canonical
momentum. Abrahams field momentum is associated with
the former, and Minkowskis with the latter; the conserved
Point Minkowski. (For related examples see (Refs. 120, 121, total is
122, 123, 108, 125, and 102).)
This experiment was performed by Jones and collabora-
tors (Refs. 109, 110, and 111), and the Minkowski prediction ptotal pkinetic gA d3 r pcanonical gM d 3 r: (77)
was confirmed. [Actually, Jones measured the force on the
mirror as a function of n, for various liquids; Minkowski
says the force is proportional to n, whereas Abraham says it You can use either the Abraham momentum or the Minkow-
should go like 1/n. Mansuripur (Ref. 122) has argued that the ski momentum for the fields, as long as you combine it with
experiment would have supported Abraham had they used a the appropriate momentum for the particles. In Example 1
mirror that did not reverse the phase of the reflected wave (a we used the kinetic momentum of the bar (Eq. (71)), so it
perfect magnetic conductor, instead of a perfect electric was appropriate in this case to use the Abraham momentum
conductor), but Kemp and Grzegorczyk (Ref. 112) show that for the fields. Baxter and Loudon (Ref. (91)) associate
even in that case the Minkowski prediction is ultimately Abraham momentum with the motion of a dielectric
sustained.] specimen as a whole and Minkowski momentum with the
Other models have been explored (Refs. 86, 96, and 135), motion of objects embedded in the dielectric.
including ones involving diffraction (see Example 3 below), There are several excellent reviews of the entire contro-
and even purely static configurations (for which, however, versy [and at least one dissertation (Ref. 92)]: Brevik (Ref.
hidden momentum may have to be considered) (Refs. 116, 93) offers a detailed survey of the relevant experiments, Pfei-
57, and 95). Some seem to favor Minkowski, others Abra- fer et al. (Ref. 132) is the most comprehensive, Milonni and
ham. Beginning in the late 1960s (Refs. 15, 4, 105, 131, and Boyd (Ref. 128) and Baxter and Loudon (Ref. 91) are the
93) something approaching a consensus emerged: Both the most up-to-date (and in my opinion the clearest and most ac-
Minkowski momentum and the Abraham momentum are cessible). But the final word on this vexed subject has cer-
correct, but they speak to different issues, and it is largely tainly not been writtenindeed, the frequency of papers
a matter of taste which of the two (or perhaps even one of continues to grow.
the other candidates that have from time to time been pro-
posed (Refs. 103, 131, 100, 124, and 136) one identifies as
the true electromagnetic momentum. The essential point VII. MOMENTUM OF PHOTONS
goes back to Poincare (Ref. 76): in the presence of matter the
electromagnetic stress-energy tensor by itself is not con- In 1900 Max Planck (Ref. 143) proposed that electromag-
served (divergenceless). Only the total stress-energy tensor netic waves come in little squirts (quanta), with energy
carries unambiguous physical significance, and how one E h; (78)
apportions it between an electromagnetic part and a
matter part depends on context and convenience. Minkow- where  is the frequency and h is the Plancks constant
ski did it one way, Abraham another; they simply regard dif- (empirically, h 6.626  10 34 J s). Planck did not pretend
ferent portions of the total as electromagnetic (Refs. 130, to know why the energy is quantizedhe assumed it had
142, 128, and 134). Except in vacuum, electromagnetic something to do with the emission process. In 1905 Albert
momentum by itself is an intrinsically ambiguous notion. Einstein (Ref. 141) reinterpreted Plancks quanta as
For example, when light passes through matter it exerts particles of light (Gilbert Lewis suggested the name
forces on the charges, setting them in motion, and delivering photon in 1926). Einsteins idea was widely ridiculed
momentum to the medium. Since this is associated with the (Ref. 12; Secs. 18a and 19f) until 1923, when Compton
wave, it is not unreasonable to include some or all of it in the (Ref. 140) accounted for the change in wavelength of light
electromagnetic momentum, even though it is purely mechani- scattered from a charged particle by treating the light as a
cal in nature. But figuring out exactly how and where this mo- massless particle with energy given by Plancks formula
mentum is located can be very tricky. For instance, in Example and momentum dictated by the relativistic invariant
2 momentum is imparted not just to the mirror, but also to the E2  p2c2 m2c4:
dielectric in front of it, so the fact that Minkowskis momentum
balances that of the mirror is to some extent fortuitous. As E h
Baxter and Loudon put it (Ref. 91), the total momentum trans- p : (79)
c c
fer has the Abraham value the [dielectric] liquid takes up the
difference between the Abraham and Minkowski momenta, For an authoritative history see Paiss biography of Einstein,
which is eventually transferred to its container. Subtle is the Lord (Ref. 12).

14 Am. J. Phys., Vol. 80, No. 1, January 2012 David J. Griffiths 14

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
The photon picture offers a more tangible way to think 8. Electrodynamics of Continuous Media (Second edition), L. D. Lan-
about electromagnetic momentum: Instead of the rather dau and E. M. Lifshitz (Addison-Wesley, Reading, MA, 1984). (A)
9. The Theory of Electrons, H. A. Lorentz (1909; Dover reprint, New
abstract notion of momentum stored in fields, it is simply the
York, 1952). (I)
total momentum of the photons present. In particular, the 10. A Treatise on Electricity and Magnetism, J. C. Maxwell (Oxford Uni-
photon picture illuminates the AbrahamMinkowski contro- versity, Oxford, 1873; Dover reprint, New York, 1954; Clarendon
versy about electromagnetic momentum in a dielectric me- reprint, Oxford, 1998). (I)
dium. Is the momentum of a photon in matter p/n, as 11. The Theory of Relativity (Second edition), C. Mller (Clarendon,
Abraham would have it, or pn, as Minkowski requires (Refs. Oxford, 1972). (A)
104, 138, 88, and 89)? 12. Subtle is the Lord, A. Pais (Clarendon, Oxford, 1982). (E)
13. Classical Electricity and Magnetism (Second edition), W. K. H. Pan-
ofsky and M. Phillips (Addison-Wesley, Reading, MA, 1962). (A)
8   14. Introduction to the Quantum Theory (Third edition), D. Park
> 1 h
> p
< A n c Abraham (McGraw-Hill, New York, 1992). (I)
15. Electrodynamics of Moving Media, P. Penfield Jr. and H. A. Haus
  (80)
>
> h (MIT, Cambridge, MA, 1967). (A)
: pM n Minkowski 16. The Theory of Heat Radiation, M. Planck (1913; Dover reprint, New
c York, 1959). (I)
17. Electromagnetic Fields: Sources and Media, A. M. Portis (Wiley,
Some of the classic examples can be presented more cleanly New York, 1978). (I)
18. Classical Charged Particles, F. Rohrlich (Addison Wesley, Reading,
in this language (Refs. 119 and 88), and it is essential in ex- MA, 1965), Sections 2-2 and 6-3. (A)
plicitly quantum mechanical arguments such as the follow- 19. Electromagnetic Radiation, G. A. Schott (Cambridge U.P., Cambridge,
ing (Refs. 139, 142, and 91). 1912). (A)
Example 3. The size of the central maximum in single-slit 20. Classical Mechanics, J. R. Taylor (University Science Books, Sausalito,
diffraction can be estimated using the uncertainty principle, CA, 2005), Sections 7.1 and 7.9. (I)
DpDx
 h=2. Here Dx is the width of the slit, w, so 21. Electricity and Matter, J. J. Thomson (Charles Scribners, New York,
1908), pp. 3033. (E)
Dp

h=2w: (81)
JOURNAL ARTICLES
The resulting angular spread of the beam is
A. Background
Dp c 22. Linear momentum density in quasistatic electromagnetic systems, J.
Dh 2 : (82) M. Aguirregabiria, A. Hernandez, and M. Rivas, Eur. J. Phys. 25,
p 2pw
555567 (2004). (I)
23. On the uniqueness of the vector for energy flow density in electromag-
Now suppose the whole system is emmersed in a transparent netic fields, U. Backhaus and K. Schafer, Am. J. Phys. 54, 279280
medium with index of refraction n. Equation (81) is still the (1986). (I)
uncertainty in p, but p itself is now given by Eq. (80). Evi- 24. Zur Dynamik der Relativitatstheorie, M. von Laue, Ann. Phys. 35,
dently the new angular divergence is 524542 (1911). (A)
25. Untersuchungen uber die Druckkrafte des Lichtes, P. Lebedew, Ann.
( Phys. 6, 433458 (1901). (I)
n Dh Abraham 26. Grundlagen einer Theorie der Materie, G. Mie, Ann. Phys. 40, 165
0
Dh 1 (83) (1913). [Partial translation in The Genesis of General Relativity, Vol
Dh Minkowski
n 4: Gravitation in the Twilight of Classical Physics, edited by J. Renn
(Springer, Dordrecht, Netherlands, 2007), pp. 633698]. (A)
Does the pattern spread out, as Abraham would have it, or 27. Repairing an elementary explanation of radiation pressure, C. E. Mun-
does it contract, as Minkowski predicts? In point of fact it gan, Am. J. Phys. 77, 965 (2009). (E)
contracts (that is why the oil-immersion microscope has bet- 28. The Pressure Due to Radiation, E. F. Nichols and G. F. Hull, Phys.
ter resolution). Rev. 17, 2650 and 91104 (1903). (I)
29. Moved by light: Lasers push everyday objects into the quantum
It is no surprise that Minkowski wins this one. The mo- world,D. Powell, Science News (Washington, D. C.) 179, 2427,
mentum that goes in the uncertainty principle is canonical (2011). (E)
momentum, and we have seen (Eq. (77)) that canonical mo- 30. On the Transfer of Energy in the Electromagnetic Field, J. H. Poynt-
mentum is associated with the Minkowski form. ing, Phil. Trans. Roy. Soc. London 175, 343361 (1884). [Reprint in
Collected Scientific Papers, J. H. Poynting (Cambridge U.P., Cam-
bridge, 1920), item 10]. (I)
31. Radiation-Pressure, J. H. Poynting, Philos. Mag. 9, 393406 (1905).
1. The Feynman Lectures on Physics, R. P. Feynman, R. B. Leighton, M. [Reprint in Collected Scientific Papers, J. H. Poynting (Cambridge
Sands (Addison-Wesley, Reading, MA, 1964), Vol. II, Section 27-6. (E) U.P., Cambridge, 1920), item 22.] (I)
32. The pressure of light against the source: The recoil from light, J. H.
2. Introduction to Electrodynamics (Third edition), D. J. Griffiths
(Addison Wesley, Upper Saddle River, NJ, 1999). (I) Poynting and G. Barlow, Roy. Soc. Proc. A 83, 534546 (1910).
3. Introduction to Elementary Particles (Second edition), D. J. Griffiths [Reprint in Collected Scientific Papers, J. H. Poynting (Cambridge
(Wiley-VCH, Weinheim, 2008). (I) U.P., Cambridge, 1920) item 28.] (I)
4. Foundations of Electrodynamics, S. R. de Groot and L. G. Suttorp 33. The Lorentz force and the radiation pressure of light, T. Rothman and
(North-Holland, Amsterdam, 1972). (A) S. Boughn, Am. J. Phys. 77, 122127 (2009); Reply to Repairing an el-
5. Classical Electromagnetic Radiation (Third edition), M. A. Heald ementary explanation of radiation pressure, ibid. 77, 966 (2009). (E)
and J. B. Marion (Saunders, Fort Worth, TX, 1995). (I)
6. Classical Electrodynamics (Third edition), J. D. Jackson (Wiley, New B. Momentum and Vector Potential
York, 1999). (A)
7. Concepts of Mass in Classical and Modern Physics, M. Jammer (Har- 34. Maxwell and the Vector Potential, A. M. Bork, Isis 58, 210222
vard U.P., Cambridge, 1961; Dover reprint, New York, 1997). (E) (1967). (E)

15 Am. J. Phys., Vol. 80, No. 1, January 2012 David J. Griffiths 15

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
35. Linear Momentum of Quasistatic Electromagnetic Fields, M. G. Cal- 63. Editorial: Questions and Answers, a new section of the American
kin, Am. J. Phys. 34, 921925 (1966). (I) Journal of Physics, R. H. Romer, Am. J. Phys. 62, 487489 (1994);
36. Origins and consolidation of field theory in nineteenth-century Answer to Question #1956. (E)
Britain: From the mechanical to the electromagnetic view of 64. Try simplest cases discovery of hidden momentum forces on mag-
nature, B. G. Doran, Hist. Study Phys. Sci. 6, 133260 (1975), netic currents, W. Shockley and R. P. James, Phys. Rev. Lett. 18,
226230. (E) 876879 (1967). (I)
37. Comment on What the electromagnetic vector potential describes, 65. Torque and force on a magnetic dipole, L. Vaidman, Am. J. Phys. 58,
Y. Gingras, Am. J. Phys. 48, 84 (1980). (E) 978983 (1990). (I)
38. On the electromagnetic momentum of static charge and steady
current distributions, A. Gsponer, Eur. J. Phys. 28, 10211042
(2007). (I) D. Momentum and Mass
39. An introduction to the vector potential, D. Iencinella and G. Mat-
teucci, Eur. J. Phys. 25, 249256 (2004). (I) 66. Prinzipien der Dynamik des Elektrons, M. Abraham, Ann. Phys. 10,
40. What the electromagnetic vector potential describes, E. J. Konopinski, 105179 (1903). (A)
Am. J. Phys. 46, 499502 (1978). (E) 67. Classical model of the electron and the definition of electromagnetic
41. A dynamical theory of the electromagnetic field, J. C. Maxwell, field momentum, T. H. Boyer, Phys. Rev. D 25, 32463250 (1982);
Philos. Trans. R. Soc. London 155, 459512 (1865). (I) Comment on the preceding paper by T. H. Boyer, F. Rohrlich, ibid.
42. Thoughts on the magnetic vector potential, M. D. Semon and J. R. 25, 32513255 (1982). (A)
Taylor, Am. J. Phys. 64, 13611369 (1996). (I) 68. Electromagnetic interaction momentum and simultaneity, R. Cosson,
and G. Guidi, Am. J. Phys. 69, 462463 (2001). (I)
69. Classical electron theory from a modern standpoint, S. Coleman, in
C. Hidden Momentum Electromagnetism: Paths to Research, edited by D. Teplitz (Plenum,
New York, 1982). (A)
43. Hidden momentum, field momentum, and electromagnetic impulse, 70. Mass renormalization in classical electrodynamics, D. J. Griffiths and
D. Babson, et al., Am. J. Phys. 77, 826833 (2009). (I) R. E. Owen, Am. J. Phys. 51, 11201126 (1983). (I)
44. Interaction of a Point Charge and a Magnet: Comments on Hidden 71. The connection between inertial forces and the vector potential, A. A.
Mechanical Momentum Due to Hidden Nonelectromagnetic Forces, T. Martins and M. J. Pinheiro, in Space Technology and Applications
H. Boyer, (2007); arXiv:0708.3367v1). (I) International Forum edited by M. S. El-Genk (AIP, Melville, NY,
45. Concerning hidden momentum, T. H. Boyer, Am. J. Phys. 76, 2007), pp. 11891200. (A)
190191 (2008). (I) 72. On the electromagnetic origin of inertia and inertial mass, A. A.
46. Angular momentum of a charge-monopole pair, K. R. Brownstein, Martins and M. J. Pinheiro, Int. J. Theor. Phys. 47, 27062715
Am. J. Phys. 57, 420421 (1989). (I) (2008). (A)
47. Linear momentum of the source of a static electromagnetic field, M. 73. An elementary account of the factor of 4/3 in the electromagnetic
G. Calkin, Am. J. Phys. 39, 513516 (1971). (I) mass, P. Moylan, Am J. Phys. 63, 818820 (1995). (I)
48. Origin of hidden momentum forces on magnets, S. Coleman and J. 74. The early history of the theory of the electron: 18971947, A. Pais, in
H. Van Vleck, Phys. Rev. 171, 13701375 (1968). (A) Aspects of Quantum Theory, edited by A. Salam and E. Wigner (Cam-
49. Exposing hidden momentum, E. Comay, Am. J. Phys. 64, bridge U.P., London, 1972). (E)
10281034 (1996). (I) 75. Classical electron models, P. Pearle, in Electromagnetism: Paths to
50. A new law in electrodynamics, O. Costa de Beauregard, Phys. Lett. Research, edited by D. Teplitz (Plenum, New York, 1982). (I)
24A, 177178 (1967); Hidden momentum in magnets and interaction 76. Sur la dynamique de lelectron, H. Poincare, Acad. Sci., Paris, C.
energy, Phys. Lett. 28A, 365 (1968). (I) R. 140, 15041508 (1905); Sur la dynamique de lelectron, Rend.
51. Examples of momentum distributions in the electromagnetic Circ. Mat. Palermo 21, 129175 (1906). [Several English translations
field and in matter, W. H. Furry, Am. J. Phys. 37, 621636 of the latter have been published; see The Genesis of General
(1969). (I) Relativity Vol. 3: Theories of Gravitation in the Twilight of Clas-
52. Dipoles at rest, D. J. Griffiths, Am. J. Phys. 60, 979987 (1992). (I) sical Physics; Part I, edited by J. Renn and M. Schemmel (Boston
53. Force on a current loop, H. A. Haus and P. Penfield Jr., Phys. Lett. Studies in the Philosophy of Science 250, Springer, 2007), pp.
26A, 412413 (1968). (I) 253271.] (A)
54. Hidden momentum and the electromagnetic mass of a charge and cur- 77. A simple relativistic paradox about electrostatic energy, W. Rindler
rent carrying body, V. Hnizdo, Am. J. Phys. 65, 5565 (1997); Hidden and J. Denur, Am. J. Phys. 56, 795 (1988). (I)
momentum of a relativistic fluid carrying current in an external electric 78. Self-energy and stability of the classical electron, F. Rohrlich, Am. J.
field, ibid. 65, 9294 (1997); Hidden mechanical momentum and the Phys. 28, 639643 (1960); Electromagnetic momentum, energy, and
field momentum in stationary electromagnetic and gravitational sys- mass, ibid. 38, 13101316 (1970). (I)
tems, ibid. 65, 515518 (1997). (I) 79. The electron: development of the first elementary particle theory, F.
55. Energy and momentum considerations in an ideal solenoid, S. M. AL- Rohrlich, in The Physicists Conception of Nature, edited by J.
Jaber, J. Electromagn. Anal. Appl. 2, 169173 (2010). (I) Mehra (D. Reidel, Dordrecht, Holland, 1973), Chap. XVI pp.
56. Momentum due to pressure: A simple model, K. Jagannathan, Am. J. 331369). (I)
Phys. 77, 432433 (2009). (I) 80. Question #26. Electromagnetic field momentum, R. H. Romer, Am. J.
57. Electromagnetic momentum density and the Poynting vector in static Phys. 63, 777779 (1995); Answer to Question #26 [Electromagnetic
fields, F. S. Johnson, B. L. Cragin, and R. R. Hodges, Am. J. Phys. 62, field momentum, ibid. 63, 777779 (1995)], K. McDonald, ibid. 63,
3341 (1994). (I) 1516 (1996), F. Rohrlich, ibid. 63, 16 (1996), B. Holstein, ibid. 63, 17
58. On the classical analysis of spin-orbit coupling in hydrogen-like (1996). (I)
atoms, A. L. Kholmetskii, O. V. Missevitch, and T. Yaman, Am. J. 81. Self-stress, covariance and the classical electron, F. R. Tangherlini,
Phys. 78, 428432 (2010). (A) Am. J. Phys. 51, 285288 (1963). (I)
59. On the influence of hidden momentum and hidden energy in the classi-
cal analysis of spin-orbit coupling in hydrogenlike atoms, D. C. Lush
(2010); arXiv:1009.0495v4. (A) E. Momentum in Matter
60. Nature of electric and magnetic dipoles gleaned from the Poynting the-
orem and the Lorentz force law of classical electrodynamics, M. Man- 82. Zur Elektrodynamik Bewegter Korper, M. Abraham, Rend. Circ. Mat.
suripur, Opt. Commun. 284, 594602 (2011). (A) Palermo 28, 128 (1909). (A)
61. Electromagnetic momentum of a capacitor in a uniform magnetic 83. SullElettrodinamica di Minkowski, M. Abraham, Rend. Circ. Mat.
field, K. T. McDonald, <www.hep.princeton.edu/ mcdonald/exam- Palermo 30, 3346 (1910). (A)
ples/cap_momentum.pdf> (2007). (I) 84. Zur Frage der Symmetrie des electromagnetischen Spannungstensors,
62. Hidden Momentum in a Coaxial Cable, K. T. McDonald, <www. M. Abraham, Z. Phys. 15, 537544 (1914). [English translation by P.
hep.princeton.edu/ mcdonald/examples/hidden.pdf> (2007). (I) Enders: enders@dekasges.de.] (A)

16 Am. J. Phys., Vol. 80, No. 1, January 2012 David J. Griffiths 16

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
85. Acceleration and trapping of particles by radiation pressure, A. Ash- 112. The observable pressure of light in dielectric fluids, B. A.
kin, Phys. Rev. Lett. 24, 156159 (1970). (A) Kemp and T. M. Grzegorczyk, Opt. Lett. 36, 493495 (2011).
86. Radiation pressure on a free liquid surface, A. Ashkin and J. M. (A)
Dziedzic, Phys. Rev. Lett. 130, 139142 (1973). (A) 113. Four Poynting theorems, P. Kinsler, A. Favaro, and M. W. McCall,
87. The energy-momentum tensor of the electromagnetic field inside Eur. J. Phys. 30, 983993 (2009). (I)
matter, N. L. Balazs, Phys. Rev. 91, 408411 (1953). (I) 114. On the energy and momentum of electromagnetic fields, I. P. Kras-
88. Resolution of the Abraham-Minkowski dilemma, S. M. Barnett, Phys. nov, Tech. Phys. Lett. 35, 141143 (2009). (A)
Rev. Lett. 104, 070401 (2010). (I) 115. Experimental decision on the electromagnetic momentum expression
89. The enigma of optical momentum in a medium, S. M. Barnett and for magnetic media, D. G. Lahoz and G. M. Graham, J. Phys. A 15,
R. Loudon, Philos. Trans. R. Soc. London, Ser. A 368, 927939 303318 (1982). (A)
(2010). (I) 116. Electromagnetic momentum in static fields and the Abraham-
90. Canonical approach to photon pressure, C. Baxter, M. Babiker, and R. Minkowski controversy, H. M. Lai, Am. J. Phys. 48, 658659 (1980).
Loudon, Phys. Rev. A 47, 12781287 (1993). (A) (I)
91. Radiation pressure and the photon momentum in dielectrics, C. Baxter 117. Energy-momentum balance in quantum dielectrics, U. Leonhardt,
and R. Loudon, J. Mod. Opt. 57, 830842 (2010). (I) Phys. Rev. A 73, 032108 (2006). (A)
92. The momentum of light in media: The Abraham-Minkowski con- 118. Momentum in an uncertain light, U. Leonhardt, Nature (London)
troversy, P. Bowyer, School of Physics and Astronomy, Southampton, 444, 823824 (2006). (E)
U.K. (2005) (24 p.). (I) 119. Theory of the radiation pressure on dielectric surfaces, R. Loudon, J.
93. Experiments in phenomenological electrodynamics and the electromag- Mod. Opt. 49, 821838 (2002). (I)
netic energy-momentum tensor, I. Brevik, Phys. Rep. 52, 133201 120. Radiation pressure and the linear momentum of the electromagnetic
(1979). (I) field, M. Mansuripur, Opt. Express 12, 53755401 (2004). (A)
94. Possibility of measuring the Abraham force using whispering gallery 121. Radiation pressure and the linear momentum of light in dispersive
modes, I. Brevik and S. A. Ellingsen, Phys. Rev. A 81, 063830 (2010). dielectric media, M. Mansuripur, Opt. Express 13, 22452250
(A) (2005). (A)
95. Electromagnetic momentum conservation in media, I. Brevik and S. 122. Radiation pressure and the linear momentum of the electromagnetic
A. Ellingsen, Ann. Phys. 326, 754769 (2011). (A) field in magnetic media, M. Mansuripur, Opt. Express 15,
96. Photon recoil momentum in dispersive media, G. K. Campbell et al., 1350213517 (2007). (A)
Phys. Rev. Lett. 94, 170403 (2005). (A) 123. Radiation pressure on submerged mirrors: Implications for the mo-
97. Laser manipulation of atoms and particles, S. Chu, Science 253, mentum of light in dielectric media, M. Mansuripur, Opt. Express 15,
861866 (1991). (I) Mechanical force in laser cooling and trapping, 26772682 (2007). (A)
Y. Shimizu and H. Sasada, Am. J. Phys. 66, 960967 (1998). 124. Electromagnetic stress tensor in ponderable media, M. Mansuripur,
Inexpensive optical tweezers for undergraduate laboratories, S. P. Opt. Express 16, 51935198 (2008). (A)
Smith et al., ibid. 67, 2635 (1999). Laser tweezers, K. T. McDonald, 125. Whence the Minkowski momentum?, M. Mansuripur and A. R.
ibid. 68, 486-488 (2000). Zakharian, Opt. Commun. 283, 35573563 (2010). (A)
98. Covariant electrodynamics in a medium. I, J. Cohn, Ann. Phys. 114, 126. Resolution of the Abraham-Minkowski controversy, M. Mansuripur,
467478 (1978). (A) Opt. Commun. 283, 19972005 (2010). (A)
99. Covariant electrodynamics in a medium. II, J. Cohn and N. Hong, 127. Variational formulation of the electrodynamics of fluids and its appli-
Ann. Phys. 125, 231252 (1980). (A) cation to the radiation pressure problem, Z. Mikura, Phys. Rev. A 13,
100. Electromagnetic momentum and radiation pressure derived from the 22652275 (1976). (A)
Fresnel relations, M. E. Crenshaw, Opt. Express 15, 714725 (2007). (A) 128. Momentum of light in a dielectric medium, P. W. Milonni and R. W.
101. Energy-momentum tensor for the electromagnetic field in a dielec- Boyd, Adv. Opt. Photon. 2, 519553 (2010). (I)
tric, M. E. Crenshaw and T. B. Bahder, Opt. Commun. 284, 129. Die Grundgleichungen fur die electromagnetischen Vorgange in
24602465 (2011). (A) bewegten Korpern, H. Minkowski, Nachr. Ges. Wiss. Goettingen,
102. On energy and momentum conservation laws for an electromagnetic Math.-Phys. Kl., 53111 (1908). Reprint in Math. Ann. 68, 472525
field in a medium or at diffraction on a conducting plate, M. V. Davi- (1910).
dovich, Phy.Usp. 53, 595609 (2010). (A) 130. Momentum, pseudomomentum, and wave momentumtoward
103. Uber die im elektromagnetischen Felde auf ruhende Korper aus- resolving the Minkowski-Abraham controversy, D. Nelson, Phys.
geubten ponderomotorischen Krafte, A. Einstein and J. Laub, Ann. Rev. A 44, 39853996 (1991). (A)
Phys. 26, 541550 (1908). [English translation in The Collected 131. The momentum of light in a refracting medium, R. Peierls, Proc. R.
Papers of Albert Einstein, Vol. 2: The Swiss Years: Writings, Soc. London Ser. A 347, 475491 (1976). (A)
19001909, English translation supplement, translated by A. Beck 132. Momentum of an electromagnetic wave in dielectric media, R. N. C.
(Princeton U.P., Princeton, NJ, 1989)]. (A) Pfeifer et al., Rev. Mod. Phys. 79, 11971216 (2007); erratum Rev.
104. Canonical and kinetic forms of the electromagnetic momentum in an Mod. Phys. 81, 443 (2009). (I)
ad hoc quantization scheme for a dispersive dielectric, J. C. Garrison 133. Constraining validity of the Minkowski energy-momentum tensor,
and R. Y. Chiao, Phys. Rev. A 70, 053826 (2004). (A) R. N. C. Pfeifer et al., Phys. Rev. A 79, 023813 (2009). (A)
105. Radiation forces and momenta in dielectric media, J. P. Gordon, 134. Division of the momentum of electromagnetic waves in linear media
Phys. Rev. A 8, 1421 (1973). (A) into electromagnetic and material parts, P. L. Saldanha, Opt. Express
106. Optohydrodynamic theory of particles in a dual-beam optical trap, R. 18, 22582268 (2010). (A)
Gordon et al., Phys. Rev. B 77, 245125 (2008). (A) 135. Observation of a push force on the end face of a nanometer silica fila-
107. Momentum exchange between light and a single atom: Abraham or ment exerted by outgoing light, W. She, J. Yu, and R. Feng, Phys.
Minkowski?, E. A. Hinds and S. M. Barnett, Phys. Rev. Lett. 102, Rev. Lett. 101, 243601 (2008). (A)
050403 (2009). (I) 136. Electromagnetic force density in dissipative isotropic media,
108. Fresnels formulae and the Minkowski momentum, A. Hirose and R. A. Shevchenko and M. Kaivola, J. Phys. B 44, 065403 (2011).
Dick, Can. J. Phys. 87, 407410 (2009); Radiation pressure on a (A)
dielectric surface, A. Hirose, ibid. 88, 247252 (2010). (I) 137. The momentum-energy tensor of the electromagnetic field, D. V.
109. Radiation pressure in a refracting medium, R. V. Jones, Nature (Lon- Skobeltsyn, Sov. Phys. Usp. 16, 381401 (1973). (A)
don), 167, 439440 (1951). (I)
110. The pressure of radiation in a refracting medium, R. V. Jones
and J. C. S. Richards, Proc. R. Soc. London Ser. A 221, 480498
(1954). (I) F. Momentum of Photons
111. The measurement of optical radiation pressure in dispersive media,
R. V. Jones and B. Leslie, Proc. R. Soc. London Ser. A 360, 347363 138. Electromagnetic momenta and forces in dispersive dielectric
(1978). (I) media, D. H. Bradshaw et al. [physics-optics], 19 (2009);
arXiv:0910.077v1. (A)

17 Am. J. Phys., Vol. 80, No. 1, January 2012 David J. Griffiths 17

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
139. Phenomenological photons and the uncertainty principle, I. Brevik, 143. Zur Theorie des Gesetzes der Energieverteilung im Normalspectrum,
Eur. J. Phys. 2, 3743 (1981). (I) M. Planck, Verh. Dtsch. Phys. Ges. 2, 237245 (1900). [English trans-
140. A quantum theory of the scattering of X-rays by light elements, A. H. lation in The Old Quantum Theory, D. ter Haar (Pergamon, Oxford,
Compton, Phys. Rev. 21, 483502 (1923). (I) The scattering of X-rays 1967), pp. 8290]. (I)
as particles, A. H. Compton, Am. J. Phys. 29, 817820 (1961). (E)
141. Uber einen die Erzeugung und Verwandlung des Lichtes betreffenden
heuristischen Gesischtspunkt, A. Einstein, Ann. Phys. 17, 132148
(1905). [Translation by A. B. Arons and M. B. Peppard, Am. J. Phys. ACKNOWLEDGMENTS
33, 367374 (1965)]. (I)
142. On diffraction within a dielectric medium as an example of the Min- The author thanks Rex Adelberger, Timothy Boyer,
kowski formulation of optical momentum, M. J. Padgett, Opt. Express Vladimir Hnizdo, Peter Milonni, and Mark Semon for useful
16, 2086420868 (2008). (I) suggestions.

18 Am. J. Phys., Vol. 80, No. 1, January 2012 David J. Griffiths 18

Downloaded 21 May 2013 to 193.205.213.166. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission