Chapter 34*

Transformation of Electric and Magnetic Fields

A great advantage of the power-force vector is that it enables us to derive

a solution for the Lorentz transformation of the electric field, E, and the
magnetic flux density, or magnetic induction, B. We start with the equatio n

f = q(E + ll X B). (34. 1)

The first sumJnand, qE, describes the eLectricforce, the second one, q(u x B),
gives the magnetic force on a charge q moving at th e velocity u. That (34. 1)
app lies in a relativistic co ntext is a working hypothesis ofthe special theory of
relativity, as pointed out in C hapter 1. This hypothesis is the foundation
of the transformation eq uations for E and B. W ith (33.3), the four-vector
pertaining ro (34.1) reads

I \
Kit = Yuq - £ · u, E + U X B , (34.2)
c

as th e scalar triple product (u x B) · u is zero. Written component by

component, we get

Exux/ C + E1 u1 / C + Ezllz/ C
= y,q
E.v+ lly B z - ltz B y
£1 + Uz B x -
)

Ux B z
Ez + u.vB1 - u1 B x

279
280 Four- lft.ctors nnd Elecrromngnetism
. .
or as a marnx equatto n,

fKO IO E.'( Ey Ez YuC \

K1 q £.,. 0 cBz -cBJ Yuttx
K 2 = -c E, -cBz 0 cBx Yutty '
(34.3)
K3 Ez cB1 -cBx 0 Yu1lz

wh ich can be wrirren mo re concisely as

(34.4)

F is cal led the electromagneticfield-strength tensor, or shorr fieLd-strength

tensor. At chis point, ir is custo mary to inrroduce the tensor formalism,
replacing th e boldface no tati on (F) with two superscripts (F f.l"), or two
su bscripts (F~111 ), or o ne superscripnu1d one subscript (F~), always in Greek.
Since we have now almost completed our course in the special theory of
relativiry, we wi ll dispense with this formalism, noting, however, that it will
be indispensable in the general theo ry of relariviry.
We now write down in matrix form the Lorentz transformation, L ,
wh ich transforms th e S-frame to th e S' -fram e with v = (v, 0, 0), and the
inverse Lorentz transformatio n, L - 1 , wh ich transfo rms the S' -frame to the
S-fram e with - v = ( -v, 0, 0),

I y,, -yvf3v 0 0 \ I Yv Yv f3v 0 0 \

- yll f3t, yll 0 0 L- Yvf3v yll 0 0
L =
0 0 1 0
1
-- 0 0 1 0
.

0 0 0 l 0 0 0 l I
(34.5)

O ne can quickly prove chat the unit matrix is fo rmed by multiplying L

and L - 1,

(34.6)
 34. * Transformation ofElectric and Magnetic Fields 28 1

This enables us to write the transformation of the Minkowski force as

K'l1- = LK 11 , Kt' = L -t K'tt. (34.7)

Any other four-vector can be transform ed in an analogous way.

Let us now return to (34.4). Since the charge, q, and the speed of light,
. . .
c, are mvanan t, we can wnte

K''t = i F' U'tt. (34.8)

c
F' is th e electromagnetic field-strength tensor in the S'-frame. Then it
follows from (34.8), (34.7), and (34.4) that

f!_F'U''' = L f!_ FU'"',

c c
and, with (34.6),

F' U'1.t = L F L- t L U 1'

F' U'1-' = L .F L- t U''J.

(34.9)

Written in components,

0 E.:, E'y E'z 0 0

E.: 0 cB~ -cB'y 0 0
E'y -cB'z 0 cB.~ l 0
E'z cB'y -cB' X
0 0 I

Ex Ez y,,f3v 0 0
0 -cBy y,, 0 0
X
-cBz cBx 0 I 0
cB1 0 0 0 l

The x indicates that the matrix multiplication has to be continued on the

next line, the order of mLJtiplication is irrelevant. Doing so, the right-hand
282 Four-Vectors nnd Electromngnerism

side becomes
0 Ex Yv(E1 - vBz) y, (Ez + vB1 )

Ex 0 cyv (Bz - ~E1 ) -cyv (B1 + ~Ez

y,,(£1 - vBz) -cyv ( Bz - ~£1 ) 0 cB_,.

y,,(Ez + vB1 ) cy,, ( B1 + ~ Ez) -cBx 0

A compariso n co mponenr by co mponenr yields the Lorentz transformations

of E and B,

E'X =Ex

B;= y, ( B+ ~ Ez)
1

B~ = Yv ( Bz - ~ E1 ) . (34. 10)

We would now like ro presenr two examples of how ro use the Lorentz
transformation of the £ -field and B-field.
Let us first return to C hapter 5, namely the tmnsformntion ofa mag-
netic (Loremz) force into an electric (Coulornb) force (Fig. 34.1 ). In that cal-
cularion , v = (v, 0 , 0) was both th e relative velocity between the S-frarne

ions
-\'
r•.
I. s'
I
l

y
I
y a'

"
elect ron~

Fig. 34. 1. Transformation of a Lorenr-L force inro a Coulomb force.

 34.• Trnnsformation ofElectric and Mngnetic Fields 283

and the S'-frame, and the velocity of the conduction elecuons inside the
copper fi lament, and also the velocity of rhe charge q direcdy above the
filament.
At the position of charge q, i.e. above the fi lament, we can write in the
S-frame

E = (0, 0, 0)

B = (0 , B" 0).

This gives fo r rhe Coulomb fo rce and the Lorentz force

fc = (0, 0, 0)
fL = (0, 0, qvB1 ).

With (34. 10) we write in the S' -frame

E' = (0 , 0, YvvB1 )

B' = (0, y, By, 0).

o the Coul omb and Lorentz. forces are

f( = qE' = (0, 0, YvqvB) (34.11 )

f~ = (0, 0, 0), (34.12)

as in rhe S'-frame the velocity of rhe charge is zero. The Lorentz force
fL is thus turned into a Coul omb force f( , augmented by a facror
of Yv·
We can explain the y -facror from the definition of force as the rime
derivative of momenmm: d p = f dt in the S-fraroe, and d p' = f' dt' in the
S'-frame. In our problem, the momentum vecror increments, d p and d p' ,
are perpendicular to the relative velocity v between the S-and the S' -frames.
In C hapter 20, we postulated that transverse momentum components are
invariant, d p = d p' . This leaves us with cl1e simple equ ation/ dt = f ' dt'.
Knowing that time passes more slowly in the S' -frame than in the S-frame
by a factor of y," we C.:1J1 immediately write dt' = dtjy,,. That is why the
force in the S' -frame has to be larger by a facror of Yv wh ich yields (34. ll ),
! ' = y,J.
284 Four- Vt'ctors nud Elecrromngnt'lism

a
q v
d
q \'
b
l(

Fig. 34.2. Two parcicles of charge q having a munaal disrance of tl

and movi ng ar rhe velociry v parallel ro rhe x-axis.

In a second example, we have two panicles in the S -frame, a and b, each

one carrying a charge q, moving ar rhe mutual distance d parallel to rhe
x-ax is at a velocity of v = (v, 0, O) (Fig. 34.2) . What is rhe magnitude of
the force between rhe two charges?
Forth is problem we can offer two independent solution strategies. The
fi rsr and more diffi cuJr track srarrs our in rhe S'-frame moving, as usual, ar
a velocity of v = (v, 0, 0) relative to the S-frame. In the S'-frame, rt and
b are at rest which, for each charge, makes the magnetic field of the orher
charge disappear. Thus, at rhe posicion of n,
1 q
E' = 0, 4 JU -d2' 0
o

B' = (0, 0, 0}.

With (34.1 0), rhe inverse Lorentz transformation is
q 1
E = ( 0, Yv 47TEo -;[i.' 0

v 1 q
B = ( 0, 0, y,2 4 -' 2 .
c 7TEo a

In the S-frame, the force on particle rt becomes

f = q (E + v X B)
1 q2 2 1 l 2 1 ld2' 0
0, y,, 4 -t2 - y,JJ,, 4 -d2' 0 0, (l - fJ,)y/1 4 '
rre-0 tl rre-0 rre-o
 34."' Tnmiformtttion ofElectric ttnd Mttguetic Fields 285

or

f = 0-
1 L l
-20 (34. I 3)
' Yv 4.rreo d ' ·

ln the S-frame, the force is smaller by a factor of Yv when compared with

th e S' -frame.
T h e second approach invokes the power-force vecror (33.3) and starts
in the S' -frame. The velocity u' of both panicles is zero there, rhus Yu' = I.
The force at a is then given by

I q2
0, 0, 4 J2' 0
.rreo
0

The inverse Lorentz transformatio n gives, with rhe hel p of (33.5),

With K 2 = ytf; , and noting that Yv i= L as bod1 particles are in motio n

in the S-frame, we obtain

and rhus the result (34. 13).

Final ly, we would like to derive Lo rentz transformations of the E-field
and B-field for the more general case of a relative velocity of v = (v.n v1 , Vz)
between the S-fram e and rh e S'-frame. We begin by rewriting (34.1 0), i. e.
the field co mponents for the transfo rmation with v = (v , 0, 0),

B.~ = B.\·

e; = Yv[E + (v x 8)
1 1) B'l = Yv

I
B~ = y, Bz - - 2 ( V X E)z . (34. 14)
c
One identifies the first line as field components parallel to v, and the second
and third lines as field components perpendicular to v . We introduce the
286 Four- ~crors and Electromagnetism

subscripts II and J.., and write

Ell = Eu Bj1 = B11
1
E~ = y, [E.L + (v x B).L] B~ = y,, B.L - 2 (v X E).L
c

= y, (E.L +v X B) (34.1 5)

The last conversion is possible as a cross product is always a vector pointing

at right angles ro the vecro rs from which ir is formed, which makes one
of rhe indices J.. superA uous. (34.1 5) now holds for any arbitrary v; rh e
limitation to v = (v, 0, 0) is overcome. Following the prescriptions of
(J 8.4) and ( 18.5), we write En and E.L as projections of E on v,
(v · E)v
£ 11 = (34. 16)
v2
(v · E)v
E .L = E - E 11 = E - . (34. 17)
IJ
2

Co rresponding results hold for B11 and B.L. This enables us ro wri te
I I I (v . E)v (v · E)v
E = En + E .L = + y,, E- + v x B
v2 v2

I I I ( v . B)v (v · B)v 1
B = B 11 + B.L =
112
+ Y" B-
v2
- - v X
c2
E

(v·mv
E = y,(E + v x B) - (y" - I )
1

v2
l (v · B)v
Bl = y, B - -v X E - (y, -1 ) 2 . (34.18)
c2 v

For v/ c --+ 0, we obtain the non-relativistic approx imation,

E1 = E + v x B

I I
B = B - 2 v x E. (34.19)
c
 34."' Transformation ofEl~ctric and Magnetic Fi~lds 287

It is inreresting ro nore that the electric and magnetic fields interrningle at

any velocity, no matter how small. This asto nishes us so much more as the
appearance of rhe E'-field in the S' -frame is due ro length co nrracrio n, as
was expl ored in C hapter 5.
From the E- and B-fi elds, two Lorentz invariant quantities can be
fo rmed wh ich are caJied the invariants ofthe electrom11gneticfield,

E' · B' = E · B (34 .20)

£ '2 _ c2B'2 = £ 2 _ c2B2. (34 .21 )

W ithout a loss ofgenerali ty, we can prove (34.20) and (34.2 1) wirh (34. J0):

E' · B' = E'XB'

X
+ E'JJ
B' + E' B' t;Z

= E.>:Bx + y: ( E B + ...;.c E Ez -
1 1 1 ·vB1 Bz - ; EzBz
c

+Y?' EzBz - : E1 Ez + vB1 Bz -

c c
E1 B1 ) v:
= E.>:Bx + y:[E B1 1 ( 1- ~) + EzBz ( 1 - ~)
= l:.~>:Bx + l:.~ B1 + EzBz = E · B,
Q.E.D., and
£'2 _ c2B'2 = £ '2
X
+ E'2
J
+ E '2 _
t;
c2B'2 _ c2B'2 _ c2B'2
X J Z

= .£.'! + Y,7(E} - 2vE1 Bz + v 2Bi + El + 2vEzBy + v 2B'j )

2 2
- c2 B~.2 - c2Y,,2 ( B12 v
+ 2---zEzB + v 2 + Bz2 -
--;;Ez v E Bz + --;;E
2- V 2)
. c 1 c c2 1 c 1

= £ .>2: _ c2B2
x
+ y2,, [E2
1
( 1 _ ~)
c2 + £ z2 ( 1 _ !C)
c2

+ B'j (v2 - c2) + s; (v 2 - c2) ]

= £2_ c2B2 + £ 2 + E 2 _ c2B2 _ c2B2 = £ 2 _ c2B2

X X J Z J Z '

Q.E.D.
288 Four- vectors and Electromagnetism

These somewhat tedious calculations enable us to infer some interesting

consequences, namely:

• lf E and B are at right angles to each other (E · B = 0), they are so

in aU inertial reference frames. Electromagnetic radiatio n serves as an
example.
• If E and B form an acute angle or an obtuse angle, i.e. if E · B =
IE II BI cos<p > 0 or E · B = IE II BI cos<p < 0, respectively, they do so
in all other inertial reference frames. A transformation into reference
frames with E = 0 and/or B = 0 is impossible. ln other wo rds: If a
Lorentz transformation is to turn E or B into zero, E and B have to
stand at right angles to each other.
• If lEI = ci BI, or £ 2 - c2B 2 = 0, rhen this is also true for all inertial
reference frames. Electromagnetic radiation may again be cited as an
example.
• If there is an £ -field bu t no B-fi eld in the 5-frame, a transformation to
a pure B'-fidd in rhe 5'-frame is impossible and vice versa. If it were
possible, we would have

or

wh ich, because of the sign, can only be true for a vanishing £ -field or
B-field .
• If in any inertial frame lEI > ci BI h olds, i.e. E 2 - c2B2 > 0, then this
must also be rrue for any other inertial reference frame. Then, as shown
before, the fields can be transformed to a pure E-field if E·B = 0 holds.
Correspondingly, for E 2 - c2 B 2 < 0, the fiel ds can be transformed to
a pure B-field forE· B = 0.