Electric force and electric field

September 16, 2015

The basic problem of mechanics or field theory is to predict the motions of particles or the time evolution
of fields. For particles, this depends on Newton’s second law, and this requires a knowledge of forces. Electric
charges, which can be either positive or negative, exert forces on one another, so we begin by characterizing
the direction and magnitude of those forces.

1 Coulomb’s law and the electric field: point charges

1.1 Coulomb’s law: the force between two charges
For a single point charge Q in the neighborhood of a second charge, q, the force on Q due to q is measured
(by determining the acceleration) to be
1 qQ
F= r̂
4π0 r2
where r̂ is a unit vector in the direction of Q from q and r is the separation of the charges. This is Coulomb’s
law, and it works for any combination of signs of q and Q. The constant
C2
0 = 8.85 × 10−12
N · m2
is the permittivity of free space.
Notice that we may write this equation in any coordinates. For example, suppose Q lies at vector position
x and q lies at x0 . Then

r = |x − x0 |
x − x0
r̂ =
|x − x0 |
where x and x0 may be expressed in any coordinate system:
qQ x − x0
F=
4π0 |x − x0 |3

This is Coulomb’s law.

1.2 Linear superposition

Perhaps the most important property of electromagnetism is that it is linear in the fields, and the first
place we see this is in the additivity of the force due to multiple charges. We again consider the force on
Q at x, but this time in the neighborhood of n different charges, q1 , q2 , . . . , qn at positions x1 , x2 , . . . , xn ,
respectively. Then the force on Q is just the sum of the forces from the individual particles,
n n
X qi Q x − xi X qi Q
F= 3 = r̂i
i=1
4π0 |x − xi | i=1
4π0 ri2

1
1.3 The electric field
A key point here is that the linearity applies to Q, so that if we have various charges Q moving in the
presence of many fixed charges qi , we can define a quantity that gives the force for any Q in a way that
depends only of the fixed charges. We define the electric field, E, produced by the charges qi at any location
x, as the force per unit charge experienced by an additional charge Q at that location, as

F (x)
E (x) =
Q
n
X qi x − xi
=
i=1
4π0 |x − xi |3

In order to avoid any argument that Q alters the configuration of “fixed”charges, we may take the limit of
small test charges,
F (x)
E (x) = lim
Q→0 Q

but it the charges are truly fixed this doesn’t make a difference.

2 Continuous distributions
At the microscopic level, charges are discrete, in units of ±e, the charge of the electron or positron. However,
over macroscopic distances there will be extremely large numbers of electrons and we may treat the charge
distribution as continuous. For example, in one cubic micron of hydrogen gas at standard temperature and
pressure, there are about 2.7 × 108 electrons. A difference of one electron therefore alters the charge by only
parts per billion. For macroscopic volumes, the effect is much smaller. A 1 amp current carries 1 Coulomb
per second, and a single electron contributes only 1.6 × 10−19 C more or less to this. The approximation of
continuous charge distributions is a very good one!
To formulate the equivalent of Coulomb’s law for a continuous distribution of charge, we introduce the
charge density, ρ (x). We can define this as the total charge per unit volume for a volume centered at the
position x, in the limit as the volume becomes “small”. What we mean by small is any size much smaller
than the size over which ρ is changing, but large enough that the volume still contains many charges.
Q in ∆V about x
ρ (x) = lim
∆V →small ∆V
Then, in a volume ∆V , the total charge is ρ (x) ∆V . The continuous idealization is good enough that we
may write this infinitesimally,
dQ = ρ (x) d3 x
With this preparation, we reconsider the electric field, replacing qi with ρd3 x. Let qi be replaced by
qi = q (xi ) = ρ (xi ) ∆V and take the infinitesimal limit
n
X ρ (xi ) ∆Vi x − xi
E (x) = lim
∆V →0
i=1
4π0 |x − xi |3

In this limit, the sum becomes an integral, as the charge positions xi → x0 vary smoothly over all space,
ˆ
1 x − x0 3 0
E (x) = ρ (x0 ) 3d x
4π0 |x − x0 |

If the charges in question are restricted to a surface S or to a curve C, we can define the charge per unit
area, σ (x0 ), or the charge per unit length, λ (x0 ), and the corresponding electric fields,

2
2.1 Example 1: Electric field of an infinite line charge
Suppose we have a infinitely long, straight wire with a constant charge density λ. What is the electric field
at all points around the wire?
Because this system has cylindrical symmetry, we may simplify the problem a great deal. Choosing
cylindrical coordinates to match the symmetry, we let the wire lie along the z-axis. Then there can be no
ϕ- or z-dependence of the electric field, E (x) = E (ρ). Moreover, E (ρ) must be radial since there can be
no directional preference for left or right along z or for clockwise or counterclockwise in the ϕ-direction, and
we have E (x) = E (ρ) ρ̂. This reduces our problem to a 1-dimensional integral, and simplifies our general
formula, ˆ
1 x − x0
E (x) = λ (x0 ) 3 dx
0
4π0 |x − x0 |
C

If we let the point x at which we wish to know the field lie at a the point (ρ, 0, 0) then E (x) becomes
ˆ∞
1 ρρ̂ − k̂z 0 0
E (x) = λ 3 dz
4π0 0
ρρ̂ − k̂z 
 
−∞

ˆ∞
λ ρρ̂ − k̂z 0 0
= 3/2 dz
4π0 
2
−∞ ρ2 + (z 0 )

Notice that the integral is over the source charge distribution along the z-axis.
By our initial reasoning, the k̂ term must give zero. This is easy to see. The k̂ term is proportional to
the integral
ˆ∞
z0 0
I=  3/2 dz
2
−∞ ρ2 + (z 0 )

Changing to a new coordinate, z 00 = −z 0 , gives

ˆ
−∞
−z 00 00
I =  3/2 (−dz )
2
∞ ρ2 + (−z 0 )
ˆ∞
z 00 0
= −  3/2 (dz )
2
−∞ ρ2 + (z 00 )
= −I

showing that the integral vanishes. This type of proof is called a parity argument.
We are left with
ˆ∞
λ 1 0
E (x) = ρρ̂ 3/2 dz
4π0 
2
−∞ ρ2 + (z 0 )

Let z 0 = ρ sinh ζ, dz 0 = ρ cosh ζdζ. Then

ˆ∞
λ ρ cosh ζdζ 0
E (x) = ρρ̂
3/2 dz
4π0 ρ2 + ρ2 sinh ζ 2
−∞

3
 ˆ∞
λ ρ2 cosh ζdζ 0
= ρ̂
3/2 dz
4π0 ρ3 2
1 + sinh ζ
−∞
ˆ∞
λ dζ
= ρ̂ dz 0
4π0 ρ cosh2 ζ
−∞
λ ∞
= ρ̂ tanh ζ|−∞
4π0 ρ
λ
= ρ̂
2π0 ρ

E (x) is directed radially away from the wire and falls off as ρ1 .

2.2 Example 2: Electric field of a finite line charge

If instead we have a wire of finite length, 2L, the symmetry argument only still holds at the center of the
wire. Taking the center at z = 0, the electric field is still independent of ϕ, but now depends on both ρ and
z. Our integral must therefore be written with x = ρρ̂ + k̂z, though we may still take ϕ = 0:

ˆL
 
1 ρρ̂ + k̂z − k̂z 0
0
E (x) = λ 3 dz
4π0 0
ρρ̂ + k̂z − k̂z 
 
−L

ˆL
λ ρρ̂ − k̂ (z 0 − z) 0
= 3/2 dz
4π0 
2 0 2
−L ρ + (z − z)

This time we take (z 0 − z) = ρ sinh ζ, and the resulting integral is the same:

ˆ
L−z
λ ρρ̂ − k̂ρ sinh ζ
E (x) =
3/2 ρ cosh ζdζ
4π0 ρ2 + ρ2 sinh2 ζ
−L+z
ˆ
L−z
λ ρ̂ − k̂ sinh ζ
= cosh ζdζ
4π0 ρ cosh3 ζ
−L+z

Notice that because the limits are asymmetrical in z, the parity argument no longer makes the k̂ integral
vanish. Instead, we have

ˆ
L−z z0 =L
sinh ζ 1 
dζ = −
cosh2 ζ cosh ζ z0 =−L
−L+z
z0 =L


1 
= − r 

2 
0
1 + z ρ−z 

z 0 =−L
ρ ρ
= q −q
2 2
ρ2 + (z + L) ρ2 + (z − L)

4
while only the limits of the ρ̂ integral change,
ˆ
L−z
1 z 0 =L
dζ = tanh ζ|z0 =−L
cosh2 ζ
−L+z

 z0 =L

z 0 −z


ρ 
= r 

2 
z 0 −z
1+

ρ 
z 0 =−L
L+z z−L
= q −q
2 2
ρ2 + (L + z) ρ2 + (L − z)

Combining these, the electric field is:

ˆ
L−z
λ ρ̂ − k̂ sinh ζ
E (x) = cosh ζdζ
4π0 ρ cosh3 ζ
−L+z
    
λ   L+z z−L ρ ρ
= ρ̂ q −q  − k̂  q −q 
4π0 ρ 2 2 2 2
ρ2 + (L + z) ρ2 + (L − z) 2
ρ + (z + L) 2
ρ + (z − L)

In the limit as L → ∞ we recover the previous result:

    
λ   L+z z−L ρ ρ
E (x) = lim ρ̂ q −q  − k̂  q −q 
L→∞ 4π0 ρ 2 2 2 2 2 2 2 2
ρ + (L + z) ρ + (L − z) ρ + (z + L) ρ + (z − L)
   
λ L+z z−L  ρ ρ 
= lim ρ̂ − − k̂ −
L→∞ 4π0 ρ L L L L
λ
= [2ρ̂]
4π0 ρ
p
There is another interesting limit. If the point of interest lies at a distance ρ2 + z 2  L then the
wire will start to look like a point charge with Q = 2λL. To check that this holds, we need to expand each
denominator in a Taylor series,
−1/2
L2
 
1 1 2zL
= p 1+ +
ρ2 + z 2 ρ2 + z 2
q
ρ2 + (L + z)
2 ρ2 + z 2

1 L2
 
1 zL
≈ p 1− 2 −
ρ2 + z 2 ρ + z2 2 ρ2 + z 2
1 L2
 
1 1 zL
≈ p 1+ 2 −
ρ + z2 2 ρ2 + z 2
q
ρ2 + (L − z)
2 ρ2 + z 2

Then
    
λ   L+z z−L ρ ρ
E (x) = ρ̂ q −q  − k̂  q −q 
4π0 ρ 2 2 2 2
ρ2 + (L + z) ρ2 + (L − z) ρ2 + (z + L) ρ2 + (z − L)
1 L2 1 L2
    
λ zL zL
= ρ̂ (L + z) 1 − 2 − − (z − L) 1 + −
ρ + z2 2 ρ2 + z 2 ρ2 + z 2 2 ρ2 + z 2
p
4π0 ρ ρ2 + z 2

5
 1 L2 1 L2
    
λ zL zL
− k̂ ρ 1 − 2 − − ρ 1 + −
ρ + z2 2 ρ2 + z 2 ρ2 + z 2 2 ρ2 + z 2
p
4π0 ρ ρ2 + z 2
zL (L + z) 1 (L + z) L2 zL (z − L) 1 (z − L) L2
 
λ
= ρ̂ L + z − − − (z − L) − +
ρ2 + z 2 2 ρ2 + z 2 ρ2 + z 2 2 ρ2 + z 2
p
4π0 ρ ρ2 + z 2
1 ρL2 1 ρL2
 
λ ρzL ρzL
− k̂ ρ − 2 2
− 2 2
−ρ− 2 2
+
2 ρ2 + z 2
p
4π0 ρ ρ2 + z 2 ρ +z 2ρ +z ρ +z
L2
   
λ 2zL (L + z) λ 2ρzL
= ρ̂ 2L − −L 2 + k̂
ρ2 + z 2 ρ + z2 ρ2 + z 2
p p
4π0 ρ ρ2 + z 2 4π0 ρ ρ2 + z 2
z2
    
2Lλ ρz
= ρ̂ 1 − 2 + k̂
ρ + z2 ρ2 + z 2
p
4π0 ρ ρ2 + z 2
2Lλ 
2

= 3/2
ρ̂ρ + ρz k̂
4π0 ρ (ρ2 + z 2 )
2Lλ  
= 3/2
ρρ̂ + z k̂
4π0 (ρ2 + z 2 )

Since we may write r = ρρ̂ + z k̂ and Q = 2Lλ, this is exactly

Q
E (x) = r̂
4π0 r2
as we predicted.

2.3 Example 3: Infinite, parallel wires

As a third example, we compute the electric field due to a pair of infinite, charged,
parallel wires. One, with
charge per unit length λ, lies parallel to the z-axis passing through (x, y) = d2 , 0 , while the second with
charge per unit length −λ, is parallel but passing through (x, y) = − d2 , 0 . The electric field from the first

is found above to be
λ
E1 (x) = ρˆ
2π0 ρ1 1
but we must offset the radial distance and direction. Changing to Cartesian coordinates, we may write
 
λ x − d2 î + y ĵ


E1 (x) = 
2 
2π0 x − d2 + y 2
q
2
where the radial distance from the wire is ρ1 = x − d2 + y 2 , and the direction is ρ1 = + (x − d) î + y ĵ.
The electric field from the second wire is similar, λ → −λ and replacing x − d2 with x + d2 ,
 
λ x + d2 î + y ĵ


E2 (x) = − 
2 
2π0 x + d2 + y 2

and, using superposition, the total electric field is simply the sum of these

E (x) = E1 (x) + E2 (x)

!
x − d2 î + y ĵ x + d2 î + y ĵ



λ
=
2 −
2
2π0 x − d2 + y 2 x + d2 + y 2

6
Superposition makes this complicated field easy to write.
p a simpler form if we approximate it for large distances from the wire. Then with ρ1 , ρ2  d,
This takes
we set ρ = x2 + y 2  d, and expand
!
x − d2 î + y ĵ x + d2 î + y ĵ



λ
E (x) =
2 −
2
2π0 x − d2 + y 2 x + d2 + y 2
!
x − d2 î + y ĵ x + d2 î + y ĵ



λ
= −
2π0 ρ2 − xd ρ2 + xd
!
x − d2 î + y ĵ x + d2 î + y ĵ



λ
= −
2π0 ρ2 1 − xdρ2 1 + xdρ2

Then, expanding the denominator in a Taylor series we have

 −1  
xd xd
1± 2 ≈ 1∓ 2
ρ ρ
so
!
d d



λ x− 2 î + y ĵ x+ 2 î + y ĵ
E (x) = xd
−
2π0 ρ2 1− ρ2 1 + xd
ρ2
       
λ d xd d xd
≈ x− î + y ĵ 1+ 2 − x+ î + y ĵ 1− 2
2π0 ρ2 2 ρ 2 ρ
d x2 d xd2 d x2 d xd2
         
λ xd xd
= x− + 2 − 2 − x+ − 2 − 2 î + y 1 + 2 − y 1 − 2 ĵ
2π0 ρ2 2 ρ 2ρ 2 ρ 2ρ ρ ρ
2x2 d
  
λ 2xyd
= 2
−d + 2
î + 2 ĵ
2π0 ρ ρ ρ

Combining terms

2x2
  
λd 2xy
E (x) = −1 + 2 î + 2 ĵ
2π0 ρ2 ρ ρ
λd  
−1 + 2 cos2 ϕ î + 2 sin ϕ cos ϕĵ


= 2
2π0 ρ
λd  
= î cos 2ϕ + ĵ sin 2ϕ
2π0 ρ2
Sketch some of the flux lines and you will see loops from one wire to the other.

3 Exercises (required)
3.1 Exercise 2.1: Force between two electrons
An electron has a charge of 1.6 × 10−19 Coulomb. Find the force between two electrons at a distance of
10−10 meter. Find the initial acceleration resulting from the electrostatic repulsion.

3.2 Exercise 2.2: Electric field from point particles

Find the electric field produced by a system of three charges: 2q at the origin, 3q at position x = 2î + 3ĵ,
and −q at x = 5k̂.

7
3.3 Exercise 2.3: Infinite plane
Find, by direct integration, the electric field at a point a distance z above and infinite plane of constant
charge density σ, located at z = 0.

3.4 Exercise 2.4: Disk

A thin circular disk lies in the xy plane at the origin. It has radius R and carries a total charge Q.

1. Find the electric field at any distance z along the z-axis.

2. Show that your answer gives the expected result for z  R.