SECTION 4

Electric Fields in Matter

This section (based on Chapter 4 of Griffiths) continues with the study of electrostatics but now in the presence
of another material (such as an insulator or dielectric). The topics are:
• Polarization
• The field of a polarized object
• Electric displacement
• Linear dielectrics
Polarization
Conductors and Insulators
Most materials falls into two broad categories as regards their electrical properties:
• Conductors   • Insulators  
These have free charges which move in These have charges which are bound to particular atoms or
response to external electric fields.   molecules.  
When a conductor is placed in an electric When an insulator is placed in an electric field, the atoms and
field, the charges move to cancel out the molecules become dipoles and line up with the field, partially
applied external field, making the net field cancelling the field in the insulator.
equal to 0 inside the conductor.

Unless the electric field is very strong, the electron cloud of a non-metallic atom is only moved a little in the
direction opposite to the electric field. The atom then becomes a dipole with moment p pointing in the direction
of the electric field. If the field isn’t too strong, we have a proportionality:
p =!E
where α is a constant called the polarizability. The polarizability is determined experimentally and it depends on
the properties of the atom.
Electron cloud, -q
E

Nucleus and electron cloud

Nucleus, +q displaced a small distance d
Molecules are more complicated, since they are typically asymmetric and can be categorized as:
Non-polar molecules (p = 0 when E = 0) Polar molecules (p ≠ 0 when E = 0)
+
H
O C O
H
+ O
-
-

+
H

- + -
O C O
O
-

H +
E E

  33  
For molecules, the polarizability is a function of direction of the electric field. In the case of carbon dioxide, the
polarizability is different along the molecule’s axis and perpendicular to it:
p = ! ! E! + !||E||
In general, for asymmetric molecules, the polarizability depends on the field component in each direction:
px = ! xx E x + ! xy E y + ! xz Ez
py = ! yx E x + ! yy E y + ! yz Ez OR p = !E
pz = ! zx E x + ! zy E y + ! zz Ez
where p and E are column vector and α is a 3 × 3 matrix. Hence there is a polarizability matrix (or tensor) in
these cases, instead of a scalar constant.
Polar molecules
Polar molecules, like water, which have built-in dipole moments, experience a torque when they are subjected
to an electric field.
F+
Consider  a  simple  dipole  of  two  charges  q  and  −q  separated  by  distance  d.  
If  the  field  E  is  uniform,  the  two  forces  are  equal  and  opposite  (so  they  
d
cancel  out),  but  they  tend  to  align  the  molecule  with  the  field.    

E                                                                                                      
F!  
The torque is: N = (r+ ! F+ ) + (r" ! F" ) = q(r+ " r" ) ! E = qd ! E
  So a dipole with dipole moment p experiences a torque
N = p!E
For discussion in class: Show that the potential energy of the above dipole in the field E is given by −p.E
Force on an electric dipole
If the electric field is non-uniform, so that it varies from one side of the dipole to the other (which is a rather
small effect typically for molecules), then there is a net force on the dipole:
F = F+ ! F! = q(E+ ! E! ) = q"E
If the dipole is very short (meaning vector d is small), we can approximate the small change in any component
of E (say the x component) by:
!E x = d " #E x
Doing the same for the y and z components gives:
!E = d " #E
and therefore (since p = qd for the dipole moment):  
Polarization of matter
In a material with neutral atoms or nonpolar molecules, when an electric field is applied, a tiny dipole is
induced in each, which on average aligns with the electric field. In a material made of polar molecules, each
dipole tends to align with the electric field.
In either case, the material becomes polarized (i.e., acquires a dipole moment throughout its volume), which can
be expressed as:
Polarization P = dipole moment per unit volume
Next, we look at the electric field or potential produced by the polarized matter.

The field of a polarized object

We have a dipole moment per unit volume in a polarized material. We already know from Section 3 that an
individual dipole gives rise to an electric field. So the polarization of the material gives rise to an electric field.
  34  
To find this electric field, it’s easier to use results for the potential. We start from the potential due to a single
dipole, which is:

re
   
The  potential  outside  a  volume  of  polarized  material  is  found  by  integrating  over  
all  the  dipoles:  
p
     
Volume V In  principle,  we  can  evaluate  this  integral  directly  for  any  polarized  material;  but  
there’s  actually  a  better  way  to  look  at  the  same  calculation.  
First, we know that:  
# 1 & r̂
"!% ( = e2
$ re ' re  
Here the prime on the gradient indicates that we’re taking the derivative with respect to the source coordinates
(re = r –r’ ). So now we have:
1 $1'
V (r) =
4!" 0 V
* P ! #"& ) d# "
% re (  
We can next use integration by parts:
1 , $P' 1 /
V (r) = . *V "! # & ) d! ! + *V ("! # P) d! !1
4!" 0 - % re ( re 0  
Then, applying the divergence theorem to the first term in the bracket gives:
1 & 1 1 )
V (r) = ( " S P ! da# $ "V (%# ! P) d! #+
4!" 0 ' re re *  
We can now re-interpret this result. Notice that the first term looks like the potential of a surface charge, where
the charge per unit area is:
! b = P ! n̂
and the second term looks like the potential of a volume charge, where the charge density is:
!b = !" # P
So we can rewrite the potential in terms of these bound charges that are associated with the polarized material:
1 # #b ! &
V (r) = % !S da" + ! b d! "(
V r
4!" 0 $ re e '  
The bound charges are more than just a mathematical analogy: they correspond to real accumulations of charge
in parts of the polarized material.
Look at two examples:
First,  consider  a  cylinder  with  uniform  polarization  along  the  
P axis.  Since  the  divergence  of  P  will  be  zero  inside,  there  will  be  
no  bound  volume  charge.      
 
However, at the ends, the dipoles are lined up perpendicular to the surface:
There  is  a  negative  net  surface  charge  on  the  left  face  of  the  cylinder,  and  a  
- + - + positive  net  surface  charge  on  the  right  face.  The  total  charge  is  not  changed,  
only  redistributed  (if  the  material  was  neutral,  the  total  charge  is  still  zero).  
 
  35  
Next, consider a sphere whose polarization is radially outwards (along the r direction):
+
+
Taking into account which way the dipoles point, it follows that here will be
+
an excess of negative charge inside the sphere, where the polarization is
- - - large (theoretically it diverges near the centre), while the surface will have
+ - - an excess of positive charge.
- -
- Again, if the material is neutral, the positive and negative charges will
+
cancel overall.
+
 
+

Electric displacement
Total electric field around dielectrics
In the general case where a material is present, we want to calculate the total potential and the electric field that
include the bound charges on a dielectric material as well as any other charges in the region (we will call them
free charges to distinguish them from the induced or bound charge).
We can rewrite Gauss’s law in terms of the total (volume) charge as
! = !b + ! f ! 0 ! " E = ! = !b + ! f = #! " P + ! f
Combining the terms with divergence, we have:
! " (! 0 E + P) = ! f
We can define a new vector D, called the electric displacement, which is the expression in parentheses:
D = !0E + P
Gauss’s law then becomes:

!"D = !f or   !# D ! da = Q f "enc
where Qf-enc is the total free charge in our chosen volume.
This is Gauss’s Law in the presence of dielectrics. We can use it to find the electric displacement in exactly the
same way as we find electric field with the standard Gauss’s law.
Problem 4.15 (from Griffiths) – for discussion in class
A thick spherical shell (inner radius a, outer radius b) is made of dielectric material with a “frozen-in”
polarization
k
P(r) = r̂
r
where k is a constant and r is the distance from the centre. (There is no free charge in the problem). Find the
electric field E in all 3 regions by two different methods:
(a) Locate all bound charge and use Gauss’s law (as in chapter 2) to calculate the field.
(b) Use the modified Gauss’s law (this chapter) to find D, and then deduce E. (The second method is much
quicker).
Note of caution!
From Gauss’s law it looks like the electric displacement D and the electric field E are analogous. However, we
need to be careful because there is no analogue of Coulomb’s law for D. This is because while the divergence of
D looks similar to the divergence of E, the curls are not the same!
The curl of E always vanishes (and this property allowed us to define a scalar potential), but the curl of D can
sometimes be nonzero. In fact:
! " D = ! 0 (! " E) + (! " P) = ! " P

  36  
The outcome is that you cannot always just use Gauss’s law to calculate D (because it gives you the effect
coming from the free charge, but doesn’t tell you about the curl of P).
The rule-of-thumb is that, if the problem has spherical, cylindrical or plane symmetry, the curl of D vanishes
and you can use the usual Gauss’s law methods to solve it: in other situations it is more complicated.
Boundary conditions
Previously we looked at how the electric field changes at a surface where there is a charge per unit area:
! ! 1
Eabove " Ebelow = "
!0
|| ||
Eabove ! Ebelow =0
Recall that the first result came from using Gauss’s law and the second came from the vanishing of curl of E. It
follows that the generalization of the boundary conditions in the case of D is
! !
Dabove " Dbelow =! f
|| || || ||
Dabove ! Dbelow = Pabove ! Pbelow

Linear dielectrics
For many materials, if the field is not too strong, the polarization is proportional to the electric field; these are
called linear dielectrics. If P and E are in the same direction (parallel vectors), we can write
P = !0 " eE
where the proportionality factor χe is called the electric susceptibility of the medium; it is dimensionless.
Here E is the total electric field; it includes parts coming from the free charge and from the bound charge
induced in the dielectric.
We can now find some relationships involving D:
D = ! 0 E + P = ! 0 E + ! 0 " e E = ! 0 (1+ " e )E
This is often written as
D = !E where ε is the permittivity of the material.
The relationship between parameters is
! = ! 0 (1+ " e )
Sometimes it is useful to write so that εr is dimensionless.
 
Then we have
 
and εr is called the relative permittivity (or the dielectric constant).

Some common linear dielectrics, !e " (C2/N.m2) "r

compared to vacuum: Vacuum 0 8.85x10-12 1
 
Air 5.4x10-4 8.85x10-12 1.00054
Salt 4.9 5.22x10-11 5.9
Ice 98 8.67x10-10 99

Fields in linear dielectrics

Inside a linear dielectric, the curl of P (and also of D) must vanish, since P is proportional to E. This argument
does not hold at the boundary of a linear dielectric, and the curl of P may be nonzero there.
Nevertheless, there are cases where we can take advantage of the lack of curl of P inside the linear dielectric: if
the dielectric is much larger than the region we’re interested in, or completely fills the space we’re considering,
we can use all the simplifications (because the surface becomes unimportant).

  37  
In the case that our space is entirely filled with a homogenous linear dielectric, we have:
!"D = !f and !"D = 0
These are exactly the equations for electric field in a vacuum, except for a factor of ε0. It means that D can be
found from the free charges, just as though the dielectric were not there. So, if we define Evac as the electric
field that the same free charges would produce in the absence of any dielectric, then:
D = ! 0 E vac
and
1 1
E = D = E vac
! !r
For example, a point charge embedded in a large dielectric has a field:
1 q
E= r̂
4!" r 2
This modification to Coulomb’s law means that the field is reduced compared to the vacuum case; the point
charge is now “shielded” by induced charges in the polarized dielectric.
Most gases, liquids and amorphous solids (e.g., glasses) are approximately isotropic in their properties; so P is
in the same direction as E and they have a single scalar constant susceptibility.
Crystalline solids, on the other hand, may polarize more easily in one direction (usually a crystal axis of
symmetry) than another. In that case, the polarization is given by:
P = !0 " e E
where χe is now a susceptibility tensor (or 3×3 matrix) for the crystal, by analogy with the polarizability tensor
for an asymmetric molecule, discussed earlier.
Boundary conditions with linear dielectrics
The bound charge density inside a linear dielectric is easily found:
$ " ' $ ! '
!b = !" # P = !" # &! 0 e D ) = ! & e ) " f
% ! ( % 1+ ! e (
i.e., it is just proportional to the free charge.
In particular, the bound (or induced) charge density vanishes if the local free charge density is zero (no
embedded charges in the dielectric). Then the only charge is on the surface of the dielectric (sort of like in the
case of a conductor).
We could go ahead by using Laplace’s equation to solve for the potential inside the dielectric – but first we
would need the new boundary conditions to be able to solve Laplace’s equation.
For the electric field, assuming there are now free charges σf at the surface, we have:
! !
!above Eabove " ! below Ebelow =" f
(Note the permeability factors).
In terms of the potential we have
!V !V
!above above " ! below below = " f
!n !n
And, as before, the potential must be continuous across the boundary:
Vabove = Vbelow

Energy in linear dielectrics

First we need to know how the capacitance changes when the capacitor is filled with a linear dielectric instead
of vacuum. It is easy to see that the new capacitance is
C = !r Cvac
(We use the definition C = Q/V and the fact that E, and hence V, are scaled by a factor of 1/εr).
  38  
Since the energy of a capacitor (for any given voltage V) is W = 12 CV 2 it follows that W also is also increased
by a factor of εr
In terms of the electric field the energy becomes
!0 2 !0
W =
2
!! E
r d" =
2
! D " E d!
Forces on dielectrics
We now show that dielectrics are attracted into regions of high electric fields; but it can be complicated to
calculate the force on a dielectric.

F We can take a shortcut by considering a

d simple example, i.e., slab of dielectric
x
partially inside a parallel plate capacitor.
l  
dielectric
(Note: by considering the slab part way out, we are avoiding any difficulties with the fringing fields near the
edges).
To find the force F we consider the work done dW when x is increased infinitesimally to x + dx. We will assume
that the total charge on each plate stays constant at Q and −Q.
For the energy stored in the capacitor, we can write
1 Q2
W=
2 C
So, for the infinitesimal change:
" 1 Q2 % Q2
!Fdx = dW = d $ ' = ! dC
#2 C & 2C 2
In terms of the voltage V = Q/C, this becomes:
V 2 dC
F=
2 dx
We need an expression for C. The basic result for a simple parallel plate capacitor is that the capacitance is
εA/d, where A is the area of the plates.
In our example, area A varies as a function of x : we have one capacitor with area wx filled with vacuum and
another with area w(l − x) filled with dielectric, where w is the other dimension of the plate, so
! wx ! w(l ! x)
C= 0 + r
d d
This gives
dC ! 0 w ! 0!r w ! 0 (1! !r )w !! 0 " e w
= ! = =
dx d d d d
So the force is related to the susceptibility of the dielectric by
!0 " ew 2
F =! V
2d

  39