Chapter 10: Potentials and Fields

10.1 The Potential Formulation

Scalar and Vector Potentials
10.1.1 Scalar and Vector Potentials
In the electrodynamics,
In the electrostatics and magnetostatics, 1 B
(i)   E   (iii)   E  
1
(i)   E   (iii)   E  0 0 t
0 E
(ii)   B  0 (iv)   B  0 J  0 0
(ii)   B  0 (iv)   B  0 J t
The electric field and magnetic field can be expressed using How do we express the fields in terms of scalar and vector
potential: potentials?
1 B remains divergenceless, so we can still write, B    A
(iii)   E  0 E  V 2V  
0
Putting this into Faraday’s law (iii) yields,
(ii)   B  0 B  A   (   A )  0 J
 A A
  E   (  A)    ( )    (E  )0
  (  A)  (  A)   2 A  0 J    2 A  0 J t t t
A
E  V
If   A  0. t
1 2

Scalar and Vector Potentials Example 10.1

B   A A Find the charge and current distributions that would give rise
E  V 
t to the potentials.  0 k
1  (ct  | x |) 2 zˆ for |x|<ct
 1 V  0, A   4c
(i)   E    2V  (  A)   for |x|>ct
0 t 0  0
E V 2A Where k is a constant, and c is the speed of light.
(iv)   B  0 J  0 0   (  A)  0 J  0 0( )   0 0 2
t t t 
Solution:    0 (  A)
t
We can further yields. 1  2A  1
J     2 A   0 0 2       A 
 1 0  t  0
 2V  (  A)   
t 0 A A A
 2 2A  V A  x  y  z  0
  x y z
  A   
0 0 2 
     A   0 0    0 J
 t   t   2
 2
2 k  0
2 A  ( 2  2  2 ) Az zˆ  0 zˆ
These two equations contain all the information in Maxwell’s x y z 2c J0
equations. A2
k k
 0 0 2   0 0 0 2c 2 zˆ   0 zˆ
3 t 4c 2c 4
 Example 10.1 (ii) 10.1.2 Gauge Transformations
Since the volume charge density and current density are We have succeeded in reducing six components (E and B)
both zero, where are the electric and magnetic fields from? down to four (V and A). However, V and A are not uniquely
  0 and J  0 determined.
We are free to impose extra conditions on V and A, as long
They might originate from surface charge or surface current.
A k as nothing happens to E and B.
E   0 (ct  | x |) zˆ
t 2
 k  k Suppose we have two sets of potential (V, A) and (V, A),
 x0  0 (ct  x ) 2 yˆ  0 ( ct  x ) yˆ which correspond to the same electric and magnetic fields.
0 k   4c x 2c
B  A   (ct  | x |)2 yˆ  
4c x 
x  0  0 k  k A  A  α and V   V  
(ct  x ) yˆ   0 (ct  x ) yˆ
2
 4c x 2c
B    A    A    α  0  α  
A A  α 
E  V    V      
c
K  nˆ  (H   H  )
t t  t 
There is a surface current K in the yz plane. 1 0 k  
 nˆ  ctyˆ  ktzˆ (   )  0  (  )  k (t )
How do we know? 0 c t t
5 6

Gauge Transformations 10.1.3 Coulomb Gauge and Lorentz Gauge

There are many famous gauges in the literature. We will
α       A   A  
 show the two most popular ones.
    
    k (t )    1
t t  V   V  t  2V  (  A)   
t 0
 2 2A   V 
Conclusion: For any scalar function , we can with impunity   A   
0 0 2 
     A   0 0    0 J
add  to A, provided we simultaneously subtract /t to V.  t   t 
The Coulomb Gauge:   A  0
Such changes in V and A do not affect E and B, and are
1
called gauge transformation. 2V    (Poisson's equation)
0
We have the freedom to choose V and A provided E and B 1  (r, t )
do not affect --- gauge freedom.
V (r , t ) 
4 0  r
d  (setting V  0 at infinity)

V instantaneously reflects all changes in . Really?

A
E  V  unlike electrostatic case.
7 t 8
 The Coulomb Gauge The Lorentz Gauge
Advantage: the scalar potential is particularly simple to  1
 2V  (  A)   
calculate; 1 t 0
2V    (Poisson's equation)  2  A
2
V 
0 
  A   0  0 2       A   0 0    0 J
1  (r, t )  t    t 
V (r , t ) 
4 0  r
d  (setting V  0 at infinity)
V
The Lorentz Gauge:   A  0 0 0
t
 2V 1
Disadvantage: the vector potential will be very difficult to  2V  0 0 2   
calculate for the non-static case. t 0 inhomogeneous
 A
2 wave equations
2A V  2 A   0 0 2    0 J
 2 A   0 J  ( 0 0 2  ( 0 0 )) t
t t
2 1
2  0 0  2 2 V   
The Coulomb gauge is suitable for the static case. t 2 0
2 : the d'Alembertian 2 A   0 J
9 10

The Lorentz Gauge 10.2 Continuous Distributions

Advantage: It treats V and A on an equal footing and is
10.2.1 Retarded Potentials
particularly nice in the context of special relativity. It can be  2V 1
regarded as four-dimensional versions of Poisson’s equation.  V   0 0 2   
2
1
t 0 static case  2V   
0
V and A satisfy the inhomogeneous wave equations, with a  A
2
 2 A   0 0 2    0 J  2 A   0 J
“source” term on the right. t
1
2 V   
0 Four copies of Poisson's equation
2 A   0 J 1  (r)
4 0  r
V (r )  d 

Disadvantage: …  J (r)
A (r )  0  d 
4 r

We will use the Lorentz gauge exclusively.

11 12
 Retarded Potentials V (r, t )
Retarded Potentials
Satisfy the Inhomogeneous Wave Equations
In the nonstatic case, it is not the status of the source right Show that the retarded scalar potentials satisfy the
now that matters, but rather its condition at some earlier time inhomogeneous wave equations.
tr when the “message” left. 1  (r, tr )  2V 1
 d   V   0 0 2   
2
r V (r , t ) 
tr  t  (called the retarded time) 4 0 r t 0
c
Sol: 1   (r, tr )   1 r ( )   (r )
Retarded potentials:
1  (r, tr )
V 
4 0    r   d 
4 0  r2
d 
V (r , t ) 
4 0  r
d  Argument: The light we see now
left each star at the retarded time  f  gf  f g
 J (r, tr ) corresponding to that star’s Using quotient rule:    
A (r , t )  0  d  distance from the earth. g g2
4 r
 1
   (r, tr )  tr   r r  rˆ
This heuristic argument sounds reasonable, but is it tr c
1 1
correct? Yes, we will prove it soon. 1  rˆ  rˆ tr  r   r, ˆ
V   [
4 0 cr r
 2 ]d  c c
13 if r is independet of tr . 14

Retarded Potentials V (r, t ) Retarded Potentials V (r, t )

Satisfy the Inhomogeneous Wave Equations (ii) Satisfy the Inhomogeneous Wave Equations (iii)
1  rˆ  rˆ 1 1    (r , t )
 V   2V 
4 0   [ 
cr r 2
]d  2V  
 1 1
[ 2  4 3 (r )]d   2  d  
4 0 c r c 4 0 r 0
 rˆ  rˆ
1 rˆ rˆ
 [ ]    (  )    (  2 ) 1  1 1 2 2   2V  2V
r 4 0  r tr2 tr2  4 0r
2
cr r c r r d   d   d    2
1 rˆ rˆ rˆ rˆ 4 0 tr2 t
 [     ]  [ 2     2 ]
c r r r r
 1    1  2V  (r, t )
   (r, tr )  tr   r   rˆ and   rˆ V  2 2 
2

tr c c c c t 0
rˆ 1 rˆ 
   2 and   2  4 3 (r )
r r r
 rˆ  rˆ 1   1   1  2V  (r , t )
 [  2 ]  [   2 ]  [ 2  4 3 (r )] V  2 2 
2
cr r c cr r r c c t 0
1  3 
  2  4 (r )
c r 15 16
 Retarded Potentials A(r, t )
The Principle of Causality
Satisfy the Inhomogeneous Wave Equations
Show that the retarded vector potentials satisfy the This proof applies equally well to the advanced potentials.
inhomogeneous wave equations . Advanced potentials:
0 J (r, tr ) 2A  (r, ta )  2V 1
1  2V  0 0  
4  r
A (r , t )  d   2
A      0 J
4 0  r
0 0
t 2 V (r, t )  d  t 2
0
 J (r, ta )  A
2

Sol: A(r, t )  0  d   2 A   0 0 2    0 J
4 r t
 J (r, tr )  r (  J )  J  (r ) r- r
  tr  t  r- r
 r  r2 c ta  t 
c
 A  g (  A)  A  (g ) The advanced potentials violate the most sacred tenet in
Using quotient rule:     
g g2 all physics: the principle of causality.

Also see Prob. 10.8… Show that the retarded potential No direct physical significance.
satisfy the Lorentz gauge condition.
17 18

Example 10.2 I 2
( ct )  s 2
1
0 for t  0 A( s, t )  ( 0 0 zˆ )  dz
An infinite straight wire carries the current I (t )   4 2
 ( ct )  s 2
s  z2
2

I0 for t  0 How?

I
2
( ct )  s2
Find the resulting electric and magnetic fields.  ( 0 0 zˆ ) ln( s 2  z 2  z )
2 0

Sol: The wire is electrically neutral, so the retarded scalar I ct  (ct )  s 2

2

potential is zero.  ( 0 0 zˆ ) ln( )

2 s

0 J (r, tr ) 0  I (tr )
4  r 4  r
A(r, t )  A( s, t )  d   zˆ dz A 0 I 0 c
E  zˆ
t 2 (ct ) 2  s 2

For t  s/c, the “news” has not yet reached P, and the potential Az ˆ 0 I 0 ct
B  A    ˆ
is zero. s 2 s ( ct )  s
2 2

For t > s/c, only the segment z  (ct ) 2  s 2 contributes.

19 20
 Retarded Fields? 10.2.2 Jefimenko’s Equations
Can we express the electric field and magnetic field using the Retarded potentials:
concept of the retarded potentials? No, but... 1  (r, tr )  J (r, tr )
V (r, t ) 
4 0  r
d  and A(r, t )  0
4  r d 
Retarded potentials: Retarded fields: (wrong)
1  rˆ  rˆ
1  (r, tr ) 1  (r, tr ) V  [  ]d 
V (r, t ) 
4 0  r
d  E(r, t ) 
4 0  r2
rˆd  A 4 0 cr r2
E  V 
 J (r, tr )  J (r, tr )  rˆ t A  0 J (r, t ) t  J
A(r, t )  0 
4 r
d  B (r , t )  0 
4 r2
d  
t
 (
tr 4  r r d ) tr   40  r d 
How to correct this problem? 1  rˆ  rˆ 0 J
E [
4 0 cr r 2
 ]d  
4  r
d 

Jefimenko’s equations. 1  rˆ  rˆ J
4 0  r 2 cr c 2 r
 [   ]d 

The time-dependent generalization of Coulomb’s law.

21 22

Jefimenko’s Equations (ii)

10.3 Point Charges
10.3.1 Lienard-Wiechert Potentials
Retarded potentials:
1  (r, tr )  J (r, t ) What are the retarded potentials of a moving point charge q?
V (r, t ) 
4 0  r
d  and A(r, t )  0
4  r r d 
Consider a point charge q that is moving
0 J (r, tr )  1 1 on a specified trajectory
B   A 
4  
r
d   0  [   J  J   ]d 
4 r r W(t )  position of q at time t.
1 1 rˆ
  J  J  rˆ and ( )   2
c r r
r - w ( tr )
The retarded time is: tr  t 
0 J 1  The time-dependent generalization c
B 
4 r
[ 2  J ]  rˆd 
cr of the Biot-Savart law. W(tr ) the retarded position of the charge.

The separation vector r is the vector from the retarded
These two equations are of limited utility, but they provide position to the field point r 
a satisfying sense of closure to the theory. r  r  W(tr ) r  W(tr ),
23 r is function of tr . 24
 Communication Total Charge
Is it possible that more than one point on the trajectory are 1  (r, tr ) 1 1
“in communication” with r at any particular time t? V (r, t ) 
4 0  r  w(t ) d   4
r 0 r  w (tr ) 
 (r, tr )d 
No, one and only one will contribute. q

Suppose there are two such points, with retarded time t1 and The retardation obliges us to evaluate  at different times for
t2 : different parts of the configuration.
r1  c(t  t1 ) and r2  c(t  t2 ) r1  r2  c(t1  t2 )

This means the average velocity of the particle in the The source in motion lead to a distorted picture of the total
direction of r would have to be c.  violate special relativity. charge.
q No matter how small the
Only one retarded point contributes to the potentials at any   (r, t )d   1  rˆ  v / c
r
charge is.
given moment.
To be proved.

25 26

Total Charge: a Geometrical Effect Total Charge: a Geometrical Effect (ii)

A train coming towards you looks a little longer than it really
is, because the light you receive from the caboose left earlier
than the light you receive simultaneously from the engine.
caboose
KK:[kǝˈbus]
In general, if the train’s velocity makes an angle  with your
line of sight, the extra distance light from the caboose must
cover is L cos  .

L cos  L  L L
L L  L L   L 
  L  c v 1  v cos  / c
c v 1 v / c
This effect does not distort the dimensions perpendicular to
L
L  Approaching train appears longer. the motion.
1 v / c
The apparent volume  of the train is 
L  
L  A train going away from you looks shorter. related to the actual volume  by 1  rˆ  v / c
1 v / c 27 28
 Lienard-Wiechert Potentials Example 10.3
It follows that Find the potentials of a point charge moving with constant
1  (r, tr ) 1 q velocity. Assume the particle passes through the origin at time
V (r , t ) 
4 0  r
d  
4 0 r (1  rˆ  v / c)
,
t  0.
  (r, tr ) v(tr )  v ( tr ) Sol: The trajectory is: W(t )  vt
A (r , t )  0 
4 r 
d   0  (r, tr )d 
4 r First compute the retarded time: r  W(tr )  r  vtr  c(t  tr )
 qv v
 0  2 V (r , t ) r 2  2r  vtr  v 2tr2  c 2 (t 2  2ttr  tr2 )
4 r (1  rˆ  v / c) c
where  (r, tr )  q (r  r, tr ). (c 2  v 2 )tr2  2(r  v  c 2t )tr  (c 2t 2  r 2 )  0

The famous Lienard-Wiechert potentials for a moving point (c 2t  r  v)  (r  v  c 2t ) 2  (c 2  v 2 )(c 2t 2  r 2 ) Which sign
tr  is correct?
charge.  1 q 1 (c 2  v 2 )
V (r, t )  4 r (1  rˆ  v / c )

0
Consider v =0 tr  t  t 2  (t 2  r 2 / c 2 )  t  r / c
 A (r , t )  v V (r , t )
 We want the minus sign
c2 29 30

Contd.: (c 2t  r  v)  (r  v  c 2t ) 2  (c 2  v 2 )(c 2t 2  r 2 )
tr  10.3.2 The Fields of a Moving Point Charge
(c 2  v 2 )
r  vtr
Using the Lienard-Wiechert potentials we can calculate the
r  c(t  tr ), and rˆ  fields of a moving point charge.
c(t  tr )
  v r  vtr  v  r v2 1 q v
r  r  v / c  c(t  tr ) 1    c ( t  t )  (  tr ) V (r , t )  and A(r, t )  2 V (r, t )
 r 4 0 r (1  rˆ  v / c) c
 c c ( t  tr )  c c
1 A
 ( c 2t  r  v )  ( c 2  v 2 )tr  Find: E  V  and B    A
c t
1 
 (r  v  c 2t )2  ( c 2  v 2 )(c 2t 2  r 2 )  (t )
c The separation vector: r  r  r  r  W(tr ) and v  W r

1 qc The retarded time tr: r  W(tr )  c(t  tr )

V (r, t ) 
4 0 (r  v  c 2t ) 2  (c 2  v 2 )(c 2t 2  r 2 )
0 qcv tr is a function of r and t.
A(r, t ) 
4 (r  v  c 2t ) 2  (c 2  v 2 )(c 2t 2  r 2 )
31 32
 Gradient of the Scalar Potential 1.2.6 Product Rules (II) Chap.1 p.23

1 q   scalar : fg
V   (r  r  v / c) The product rule: 
4 0 (r  r  v / c) 2
vector : fA
d df dg
r  c(t  tr )  ctr ( fg )  g f ( fg )  gf  fg
See Chap.1 p.23 dx dx dx
       ( fA)  f  A  f (  A)   ( fA)  f  A  f (  A)
(r  v )  (r  ) v  ( v  )r  r  (  v )  v  (  r )
      
#1 #2 #3 #4
 scalar : A  B
    
#1 (r  ) v  (rx  ry  rz )v vector : A  B
x y z
dv tr dv tr dv tr ( A  B)  A  (  B)  B  (  A )  ( A  )B  (B  ) A
 (rx  ry  rz )
dt x dtr y dtr z   ( A  B)  B  (  A)  A  (  B) 少見
 r
 a(r  tr )
  ( A  B)  (B  ) A  ( A  )B  A(  B)  B(  A)
acceleration
33 34

       
(r  v )  (r  ) v  ( v  )r  r  (  v )  v  (  r )

#2 ( v  )r  ( v  )r  ( v  ) W(tr )  v  ( v x
 vy  v z ) W ( tr )       
x y z #1 #2 #3 #4
dW tr dW tr dW tr  
 v  (vx  vy  vz )  v(1  ( v  tr ))  a(r  tr )  v(1  ( v  tr ))  r   a  tr   v   v  tr 
dtr x dtr y dtr z 
 v  (r  a  v 2 )tr
   vz v y v v v v  r 1 1  
tr     r   (r  r )1/2  
1  
#3 r  (  v )  r  (  )xˆ  ( x  z ) yˆ  ( y  x )zˆ    1/2 (r  r )
 y z z x x y  c c c 2c ( r  r )
1    
  1/2 2 r  (  r )  (r  )r 
  v t v t v t v t v t v t 
= 
 r  ( z r  y r )xˆ  ( x r  z r ) yˆ  ( y r  x r ) zˆ 
 tr y tr z tr z tr x tr x tr y  2c ( r  r )
   
 r   a  tr   r  (  r )  r  ( v  tr )
where      
(r  )r  (r  )(r  W(tr ))  r  v(r  tr )
  ( z  Wz )  ( y  Wy )  ( x  Wx )  1   
 (  ) ˆ
x  (  tr     1/2 r  ( v  tr )  r  v(r  tr )
 y z z c(r  r )
#4 v  (  r )  v   

  ( z  Wz ) ˆ  ( y  Wy )  ( x  Wx ) ˆ  1   r
 )y  (  )z    r  (r  v )tr )  tr =   W(tr ) is function of tr .
  x  x  y  cr cr  r  v
 v   v  tr  P.14, tr 
1 1
r   rˆ because r is independet of tr .
35 c c 36
 1 qc    Curl of the Vector Potential
V   3 (rc  r  v ) v  (c  v  r  a)r 
2 2

4 0 (rc  r  v )
1 1
A  2
  (Vv )  2 V (  v )  v  V 
c c
Similar calculations 1 q r   
  3 r  (c  v ) v  (r  a) v  (r  u)a 
2 2
  c 4 0 (r  u)
(rc  r  v )(  v  r  a / c) 
A 1 qc  r 2  1 q r   1
     3 r  ( c  v )u  r  ( u  a)   rˆ  E
2 2
t 4 0 (rc  r  v )3   (c  v  r  a) v
2
 c 4 0 (r  u) c
 c 
 
where r  v  r  u.
A q r 
E  V    3 ( c  v )u  r  (u  a) 
2 2

t 4 0 (r  u)
1 The magnetic field of a point charge is always
B  rˆ  E perpendicular to the electric field, and to the
where u  crˆ  v c
vector from the retarded point.

37 38

Generalized Coulomb Field Example 10.4

Calculate the electric and magnetic fields of a point charge
moving with constant velocity.
 
  Solution:
q r   
E  ( c  v )u 
2 2
r  ( u  a)
4 0 (r  u)3      q r
 velocity field acceleration field  E  3 (c  v )u, since a  0.
2 2

 radiation field  4 0 (r  u)

u  crˆ  v

v  0 and a  0  ru  cr  rv  c(r  vtr )  c(t  tr ) v  c(r  vt );
 
E
q r
( c 3 )rˆ 
q 1
rˆ  r  u  cr  r  v  Rc 1  v 2 sin 2  / c 2 (Prob. 10.16)
4 0 ( cr ) 3
4 0 r 2 where  is the angle between R and v.

q 1  v2 / c2 Rˆ
E , where R  r  vt
4 0 (1  v 2 sin 2  / c 2 )3/2 R 2
39 40
 Fields of a Moving Point Charge Homework of Chap.10

q 1  v2 / c2 Rˆ
E ,
4 0 (1  v 2 sin 2  / c 2 )3/2 R 2
where R  r  vt
Obtain the same result by using the
Lorentz transformation. Chap.12

Problem 10.16 Show that the scalar potential of a point charge moving
with constant velocity  Eq. 10.49  can be written more simply as?
1 q
V (r , t )  , 10.51
1 1 4 0 R 1  v 2 sin 2  / c 2
B  (rˆ  E)  2 ( v  E) where R  r  vt is the vector from the present ! position of the particle
c c to thefield point r, and  is the angle between R and v.
Note that for nonrelativistic velocities (v 2  c 2 ),
1 q
V  r, t  
41
4 0 R 42

Homework of Chap.10

43