Electric Fields in Matter

PYL100: EM Waves and Quantum Mechanics

Semester I 2018-2019

Prof. Rohit Narula1

1 Department of Physics

The Indian Institute of Technology, Delhi

August 22, 2019

Outline

Preliminaries

Polarization

The Electric Displacement

Linear Dielectrics
References

▶ Introduction to Electrodynamics, David J. Griffiths [IEDJ]

▶ Chapter IV, Electric Fields in Matter
Electric Dipole

▶ An electric dipole consists of two equal and opposite charges

(±q) separated by a distance d.
▶ Its potential at point r is1 ,
( )
1 q q 1 qd cos θ
Vdip. (r) = − ≈
4πϵ0 s+ s− 4πϵ0 r2
when r is relatively large, i.e., r ≫ d.
▶ The potential of a dipole falls off as 12 , as opposed to an
r
isolated point charge, which falls off as 1r .
1
HW: Prove this using the law of cosines, i.e.,
( )2 ( )
s2+ = d2 + r2 − 2 d2 r cos θ, and cos(π − θ) = − cos θ.
The Dipole Moment

▶ The dipole moment is defined as,

p ≡ qd
where d is the displacement vector pointing from the −
charge to the + charge.2
▶ We may recast the far-field potential due to the dipole as,

qd cos θ p · r̂ p cos θ
Vdip. (r) ≈ 2
= 2
=
4πϵ0 r 4πϵ0 r 4πϵ0 r2
where r̂ is directed from the center of the dipole to the
observation point.
2
Opposite to how the electric field lines point.
The Electric Field E due to p

▶ Given that,
p cos θ
Vdip. (r) ≈
4πϵ0 r2
and using E = −∇V expressed in spherical coordinates we
get,
p
Edip. (r) ≈ (2 cos θr̂ + sin θθ̂)
4πϵ0 r3
assuming that p is oriented along ẑ.
▶ Notice that3 Edip. (r) falls off as 13 , whereas a single point
r
charge would’ve fallen off as r12 .

3
Again, Edip. (r) is only valid if r ≫ d.
The Coordinate-Free Far Field Dipole Field Edip.

▶ Without stipulating that the dipole p points in ẑ direction, the

dipole field in the far-field (r ≫ d) is,
1
Edip. (r) ≈ [3(p · r̂)r̂ − p]
4πϵ0 r3
, and is evidently valid for any coordinate system.
▶ HW: Prove it!
Induced Dipoles

▶ Problem: What happens to a neutral atom when it’s placed

in an external electric field Eext. ?
▶ While the atom is electrically neutral as a whole, it consists of
a positively nucleus and a negatively charged electron cloud
surrounding it.
▶ The negatively charged electron cloud is drawn to the Eext.
relative to the positively charged nucleus, and when
equilibrium is reached, the atom is said to be polarized.
Induced Dipoles

▶ Each polarized atom can be seen as a tiny dipole consisting of

charges +q in the nucleus and q at the center of the electron
cloud.
▶ Each such dipole sets up a small induced electric field,
pointing from the positively charged nucleus to the center of
the equally but negatively charged electron cloud.
▶ This induced electric field Eind. , called a polarization field, is
generally weaker than and opposite in direction to Eext. .
Polarizability

Figure: Atomic polarizabilities for different isolated atoms. [IEDJ]

▶ Phenomenologically, each individual atom now has a dipole

moment p proportional to the externally applied electric field,
Eext. 4 ,
p = qd = αEext.
pointing in the same direction as Eext. .
▶ α is known as the atomic polarizability5 .

4
This is an empirical/phenomenological relation valid for relatively small
electric fields.
5
In general, α is a tensor quantity, and thus p may not point exactly in the
direction of E.
Polar Molecules

▶ Thus far we’ve discussed molecules that had no built-in dipole

moment. Their dipole moment was induced due to the
external electric field.
▶ However, there exist molecules like water, H2 O and ethanol
that have a built-in dipole moment.
▶ Problem: How does a polar molecule respond to an external
field Eext. ?
The Torque on p due to Eext.

▶ If Eext. is uniform, the figure above suggests a torque N,

N = (r+ × F+ ) + (r− × F− ) = qd × Eext. = p × Eext.

▶ Notice that the torque N is in such a direction as to line up,

p∥E
, i.e., a polar molecule that is free to rotate will swing around
until its dipole moment points in the direction of the applied
field.6
6
Why?
The Force on p due to E

▶ If the field is not uniform, F+ will not exactly balance F− ,

and there will be a net force F on the dipole,
−
F = (F+ + F− ) = q(E+ ext. − Eext. )
Eext. (r + dn̂) − Eext. (r)
= p lim
{z d
d→0
| }
directional derivative
= p [(n̂ · ∇)Eext. ] = (p · ∇)Eext.
assuming the dipole is relatively short, i.e., d → 0.
The Energy of a Dipole

▶ Problem: What’s the energy of a tiny dipole p in an electric

field Eext. ?
▶ For a dipole with charge +q at r + d, and charge −q at r, its
energy in the external field Eext. is7 ,
[ ∫ r+d ]
U = q(Vext. (r + d)) − q(Vext. (r)) ≈ q − Eext. · dl
r

▶ Given that our dipole is relatively tiny, i.e., d → 0 , we may

take Eext. there to be uniform, giving us8 ,

U = −qEext. · d = −p · Eext.

7
How does U differ from the work required to assemble a physical dipole?
8
Can you justify the negative sign?
Polarization
▶ Problem: What happens to a piece of dielectric material
when it’s placed in an electric field?
▶ If the substance consists of nonpolar molecules/atoms, the
field will induce in each of them a tiny dipole moment p ∥ E.
▶ If the material is made up of polar molecules, each
permanent dipole will experience a torque, again tending to
line it up p ∥ E.
▶ We now define a macroscopic measure of polarization for the
entire material (gazillions of atoms),

P ≡ dipole moment per unit volume

called polarization such that,
∑ ∫
ptotal = pind. ≡ P(r′ )dτ ′
all atoms V
Polarization

▶ Instead of dealing with individual dipoles (pind. ), we prefer to

deal with the dipoles collectively,
∑
ptotal = pind.
and therefore the polarization density P.
▶ This is because the field distribution of individual dipoles is
quite complex.
▶ While collectively, they can usually be related to the electric
field Eext. in a simple way, e.g., linear dielectrics (P ∝ Eext. ).
▶ Also, once P is known, the potential V (or, alternatively, the
field Edip. ) due to it becomes easy to calculate, as we will see
next...
Bound Charges

▶ Problem: What’s the potential V(r) produced due to a

collection of dipoles that constitute a polarized medium?
▶ Recall that for a single dipole located at r′ , the far-field
potential is,
p · ŝ
V(r) ≈
4πϵ0 s2

▶ Integrating over all the individual dipoles constituting, and

using the polarization P, gives us,
∫
1 P(r′ ) · ŝ ′
V(r) ≈ dτ
4πϵ0 V s2
′
where dτ is the volume element.
Bound Charges

▶ Using the identity, ( )

′1 ŝ
∇ = 2
s s
we get, ∫ ( )
1 1
V(r) ≈ P · ∇′ dτ ′
4πϵ0 V s

▶ and then invoking the divergence theorem, and integration

by parts,
I ∫
1 1 1 1 ′
V(r) = P · da′ − (∇ · P)dτ ′
4πϵ0 S s 4πϵ0 V s
Bound Charges
I ∫
1 1 1 1 ′
V(r) = P · da′ − (∇ · P)dτ ′
4πϵ S s 4πϵ0 V s
| 0 {z } | {z }
first term second term

The first term is the potential due to a surface charge

density
[ ]
′ C
σb ≡ P · n̂
m2
The second term is the potential due to a volume charge
density
[ ]
′ C
ρb ≡ −∇ · P
m3

In both cases, the primed notation ′ emphasizes that the

relation only applies in the volume in which polarization exists.
Bound Charges
▶ Finally, we may write,
I ∫
1 σb ′ 1 ρb ′
V(r) = da + dτ
4πϵ0 S s 4πϵ0 V s
▶ We’ve concluded that the potential (and, hence also the
associated electric field) of a macroscopic, polarized object in
the far-field is equivalent to that produced by,
▶ a surface charge density σb ≡ P · n̂′
▶ a volume charge density ρb ≡ −∇′ · P
▶ We’re saved from doing the work of summing the
contributions of each of the individual dipoles constituting
the material, and instead …
▶ Instead we first calculate/find these bound charges (both σb ,
and ρb ), and simply calculate the fields they produce using the
familiar formula for the potential due to a charge distribution.
Physical Interpretation of Bound Surface Charges σb

▶ Consider the case of uniform polarization first, i.e.,

∇ · P = 0, e.g., a long, linear string of identical dipoles.
▶ Along the line, the head of one effectively cancels the tail of
its neighbor, but at the ends (aka the ’surface’) there are two
charges left over: plus at the right end and minus at the left.
▶ These are essentially the surface bound charges σb .
▶ We call the net charges at the ends bound charges9 to
emphasize that they cannot be removed, as they are always
bound to their respective positively charged ions.

9
Bound charges may be present even in the absence of an external electric
field, e.g., an electret.
Physical Interpretation of Bound Volume Charges ρb

▶ On the other hand, if the polarization is non-uniform, i.e.,

∇ · P ̸= 0 we get accumulations of bound volume charge
within the material, besides those on the surface due to σb .
▶ A diverging P results in a pileup of negative charge (at that
point) because the higher the ∇ · P the more the source
bound charges would tend to cancel Eext. .
Physical Interpretation of Bound Volume Charges ρb

▶ Overall charge neutrality dictates that overall bound charge

in a volume enclosing a dielectric qbv must be exactly
cancelled by the charge on all its surfaces qbs , i.e.,
qbs + qbv = 0
I ∫
(P · n̂) da + (−∇ · P) dτ = 0
S | {z } V | {z }
σb ρb
The Field Due To A Uniformly Polarized Sphere

▶ Problem: Find the electric field E, and potential V both

inside and outside a uniformly polarized P = Pẑ, solid sphere
of radius R.
The Field Due To A Uniformly Polarized Sphere
1. Using Gauss’ law, and the far-field of a dipole

▶ One way to imagine bound charges in the uniformly

polarized case is that we have…
▶ two slightly displaced uniformly charged spheres: a positive
sphere and a negative sphere.
▶ Without polarization the two are perfectly superimposed and
cancel out completely.
▶ This leftover charge (at either end) is the bound surface
charge σb .
The Field Due To A Uniformly Polarized Sphere
1. Using Gauss’ law, and the far-field of a dipole

▶ The field in the region of overlap between two uniformly

charged spheres is uniform, and given by10
1 qd
E=−
4πϵ0 R3
where q is the total charge of the positive sphere, d is the
vector from the negative → positive center.

10
Work this out by applying Gauss’ law!
The field due to a uniformly polarized sphere
1. Using Gauss’ law, and the far-field of a dipole

▶ We can express this in terms of the uniform polarization P

of the sphere,
4
p = qd = πR3 P
3
as
1
E=− P (inside)
3ϵ0
The field due to a uniformly polarized sphere
1. Using Gauss’ law, and the far-field of a dipole

▶ For points outside, it’s as though all the charge on each

sphere were concentrated at the respective center. We have,
at the far-field, a simple dipole, with potential

1 p · r̂
V=
4πϵ0 r2
with p = 43 πR3 P.
▶ The field outside (far-field) the sphere can be determined by

E = −∇V
The field due to a uniformly polarized sphere
2. Using the method of bound charges

▶ For convenience, we choose the z axis to coincide with the

direction of polarization P.
▶ Since P is given to be uniform, the volume charge density
ρb ≡ −∇ · P = 0
▶ While the surface charge density is
σb ≡ P · n̂ = P ẑ · r̂ = P cos θ
The field due to a uniformly polarized sphere
2. Using the method of bound charges

▶ We’ve thus reduced the problem to calculating the field due to

solely a surface charge density, σb = P cos θ!
▶ From the law of cosines,
√
s = R2 + z2 − 2Rz cos θ
The field due to a uniformly polarized sphere
2. Using the method of bound charges

▶ The potential due to a surface charge density σ is,

∫
1 σ ′
V(r) = da
4πϵ0 s
∫ 2π ∫ π
P (cos θ)R2 sin θdθdϕ
= √
4πϵ0 ϕ=0 θ=0 R2 + z2 − 2Rz cos θ
∫ π
πPR2 sin 2θ
= √ √ dθ
2 2
4πϵ0 R + z θ=0 1 − 2Rz 2
R +z 2 cos θ
The field due to a uniformly polarized sphere
2. Using the method of bound charges

▶ We now make use of the result,

∫ ( (√ √ ) (√ √ ))
π
sin 2θ 4 2 1−k− k+1 +k 1−k+ k+1
√ =−
θ=0 1 − k cos θ 3k2

to finally give us,

{
P
3ϵ0 r cos θ r<R
V(r, θ, ϕ) = P R3
3ϵ0 r2 cos θ r≥R
The field due to a uniformly polarized sphere
2. Using bound charges

▶ Recognizing that r cos θ = z, the field inside the sphere is

uniform11 :
P 1
E = −∇V = − ẑ = − P (r < R)
3ϵ0 3ϵ0
exactly as we found in Method 1.
11
Alternatively, use the formula for the gradient in spherical coordinates:
∇f = ∂r
∂f
r̂ + 1r ∂θ
∂f
θ̂ + r sin1 θ ∂ϕ
∂f
ϕ̂
The field due to a uniformly polarized sphere
2. Using the method of bound charges

▶ Outside the sphere, the potential is identical to the far-field

potential of a dipole at the origin,

P R3 1 p · r̂
V= cos θ = (r ≥ R)
3ϵ0 r2 4πϵ0 r2
whose dipole moment is, unsurprisingly, equal to the total
dipole moment of the sphere
4
p = πR3 P
3
Gauss’ Law in the Presence of Dielectrics

▶ Within a dielectric, the total charge density should be written

as,
ρ= ρb + ρf
|{z} |{z}
polarization everything else

where ρf might consist of electrons in a conductor or ions

embedded in the dielectric material or whatever; any charge,
in other words, not a result of polarization!
▶ Using Gauss’ Law, i.e., ∇ · E = ρ
ϵ0 we see that,

ϵ0 ∇ · E = ρ = ρb + ρf = −∇ · P + ρf
Gauss’ Law in the Presence of Dielectrics

▶ We define the electric displacement D as,

D ≡ ϵ0 E + P

▶ and rearrange terms such that,

∇ · (ϵ0 E + P) = ∇ · D = ρf
showing that D obeys its own Gauss’ law12 ,
I
∇ · D = ρf or, D · da = qf

▶ But there’s a catch…

12
The same symmetry-related caveats apply to being able to apply Gauss’
law for D in practice as was the case with E.
Gauss’ Law in the Presence of Dielectrics

▶ Note that even in the electrostatic regime, in general,

∇ × D ̸= 0
since ∇ × P ̸= 0, in general!
▶ Since, in general, ∇ × D ̸= 0, Helmholtz’s theorem tells us
that there’s no Coulomb’s law associated with D, i.e.,
∫
1 ŝ
D(r)̸= ρf (r′ )dτ ′
4π s2
, or, alternatively, knowledge of ρf alone is not sufficient to
determine D.
Gauss’ Law in the Presence of Dielectrics

▶ While Gauss’ law does hold, in general, for D, i.e.,

I
D · da = qf
it’s not useful to actually calculate D.
▶ Also, in general, since ∇ × D ̸= 0, we have no scalar potential
associated with D.
▶ However, ∇ × P = ∇ × D = 0 is guaranteed when
symmetries such as spherical, cylindrical, or plane are
present, and, only in such cases we may use Coulomb’s
law/Gauss’ law to evaluate D given ρf .
Example: D

▶ Problem: A long, straight wire, carrying uniform line charge

λ, is surrounded by rubber insulation out to a radius a. Find
the electric displacement D.
▶ Since we do have cylindrical symmetry, we may use Gauss’
Law for evaluating D,

D(2πsL) = λL
and thus,
λ
D= ŝ
2πs

▶ Notice that this formula holds both within the insulation and
outside it.
▶ Can we determine E inside the rubber yet?
Boundary Conditions for D [END OF LEC. 6]

▶ Using the same procedure of constructing a Gaussian pillbox

as for the normal component E⊥ we get,

Dabove
⊥ − Dbelow
⊥ = σf
while the continuity of the tangential component E∥ implies,

Dabove
∥ − Dbelow
∥ = Pabove
∥ − Pbelow
∥
Revisiting the Uniformly Polarized Sphere

▶ Problem: For the uniformly polarized sphere with P = Px̂,

what’s D just outside the sphere?
▶ We’d seen earlier that,

Eoutside ̸= 0

▶ Indeed, just outside the sphere we have vacuum, and hence

zero polarization, i.e.,

Poutside = 0
Revisiting the Uniformly Polarized Sphere

▶ Using the definition of D, we find that

Doutside = ϵ0 Eoutside + Poutside
̸= 0
despite there being no free charge ρf anywhere!
▶ However,
I
D · da = qf = 0 or, ∇ · D = ρf = 0
still holds for any surface everywhere, yet is consistent with
D ̸= 0.
▶ Bottom line: Even if ρf = 0, in general, we cannot conclude
that D = 0.
Gauss’ Law in the Presence of Dielectrics
revisited
▶ Issue: Some of you had pointed out during the lecture that in the
presence of a dielectric, the total charge density ρ neglects the
contribution of the surface charge density σb , and should instead be
written as,
ρ = ρb + ρbs + ρf
∑
= ρb + σbi δ(η − ηi ) + ρf (!!?)
i
where ηi represents the direction normal to the surface charge σbi .
▶ This issue arises because in the expression for the potential due to a
macroscopic polarization P, i.e.,
∫ I
1 1 1 1
V(r) = (−∇′ · P)dτ ′ + P · da′
4πϵ0 V s 4πϵ0 S s
we had not properly accounted for the fact that P is a vector
function and should be defined for all space.
▶ While, V(r) above is OK, since the limits of integration were
restricted to the periphery of the dielectric, it creates trouble later
on…
Gauss’ Law in the Presence of Dielectrics
revisited

▶ A more mathematically sound way to write the potential

due to a polarization vector field P is:
∫
1 1
V(r) = (−∇′ · P)dτ ′
4πϵ0 all space s
where the ”surface term” vanishes due to the assumption of
the finite volume of the dielectric.
▶ Note that in this representation, the volume charge density
ρb = −∇ · P actually contains the contribution of the bound
surface charge density σb term, as would be made clear by the
next example.
A ’Uniformly’ Polarized Wire

▶ Problem: Consider a square wire of length L, and sides a with a

’uniform’ polarization density P0 = const. along its axis ẑ. Find
the bound charges.
▶ We can write the three-dimensional polarization vector field13
representing the wire as,

P = P0 [H(x) − H(x − a)] [H(y) − H(y − a)] [H(z) − H(z − a)] ẑ

where H(x) is the Heaviside step function.

13
A vector field must be defined for all space.
A ’Uniformly’ Polarized Wire
▶ The negative of the divergence is,
ρb = −∇ · P = 0 − P0 [H(x) − H(x − a)] [H(y) − H(y − a)] [δ(z) − δ(z − a)]

▶ The first term is zero (only for this particular example), and
constitutes the bound volume charge density.
▶ While the second term above is really two surface charges
σb = P0 [H(x) − H(x − a)] [H(y) − H(y − a)] lying at z = 0 and
z = L as,
− σb δ(z) + σb δ(z − a)

▶ Thus, once you write P properly as a vector function,

ρb = −∇ · P, it actually includes the contribution of both the
volume and the surface charges!
▶ Thus, the relationship,
ρ = ρf + ρb (✓)
is perfectly valid, and does not neglect the surface charge!
Susceptibility
▶ For many substances, the polarization is proportional to the
field, provided E is not too strong:
P = ϵ0 χe E

▶ χe is called the electric susceptibility of the medium (ϵ0 has

been extracted to make χe dimensionless).
▶ The susceptibility is a material property, and substances that
obey this constitutive relation14 are called linear dielectrics.
▶ For such linear materials,

D = ϵ0 E + P = ϵ0 (1 + χe ) = ϵE
where ϵ is called the permittivity of the material, while
ϵr ≡ 1 + χe
is known as the relative permittivity or dielectric constant.
14
A constitutive relation is a relation between two physical quantities that is
specific to a material or substance, and approximates the response of that
material to external stimuli.
Common Dielectric Constants

▶ Notice the variation of nearly 5 orders of magnitude!

▶ What explains the wide variation in dielectric constant
between materials?
HW: The Bar Electret

▶ Problem: [Griffiths P.11] A short cylinder, of radius a and

length L, carries a “frozen-in” uniform polarization P, parallel
to its axis15 . Find the bound charge, and sketch the electric
field E, and the displacement field D.
▶ L >> a
▶ L << a
▶ L≈a

15
An electret is the electrostatic analog of a permanent magnet.
Energy in Dielectric Systems

▶ Problem: What’s the work required to charge up a dielectric

material?
▶ Claim: We can simply use the expression derived earlier for
the work required to assemble a set of charges, i.e.,
∫ ∫
1 ϵ0
W= ρVdτ = |E|2 dτ
2 2
▶ Q: Is this view correct?
▶ Ans: It’s close but not quite, because the work required to
charge up a dielectric is not only of an electrostatic nature,
there’s also a change in the internal energy such as bond
twisting, and stretching (think of springs) holding the
electrons and their ionic cores together, which we must
account for.
Energy in Dielectric Systems

▶ Suppose you’re given a fixed piece of dielectric having an

exisiting macroscopic free charge density qf , potential V, and
fields E and D.
▶ We now bring a tiny amount of incremental free charge δqf ,
to the dielectric.
▶ In response to δqf , the polarization P (and thus E, and D) of
the dielectric will change, and therefore the
bound charge distribution ρb till equilibrium is achieved.
Energy in Dielectric Systems
▶ The incremental work done in bringing a tiny amount of
free charge δρf from ∞ to the dielectric material which
provides a potential V is,
δW = (δqf )(V − 0)
which can be rewritten as,
∫
δW = (δρf )Vdτ
all space

▶ From Gauss’ Law ∇ · D = ρf , and for a small change δ,

δ(∇ · D) = (∇ · δD) = δρf
and thus, ∫
δW = (∇ · δD)Vdτ
all space

▶ Using the vector identity,

∇ · [(δD)V] = [∇ · (δD)]V + δD · (∇V)
…
Energy in Dielectric Systems
▶ …and then using integration by parts we get,
∫ ∫
δW = ∇ · [(δD)V]dτ + (δD) · Edτ
all space all space

▶ Applying the divergence theorem to the first term gives a surface

integral,
I :0 ∫
  
δW = [(δD)V]da + (δD) · Edτ
space
all all space

▶ The first term/surface integral then vanishes over all space, i.e.,
even for point charges, VδD ∝ 1r r12 ∝ r13 , thus going to zero faster
than r12 , and hence the surface integral over all space goes to zero.
▶ Therefore, the incremental work done is,
∫
δW = (δD) · E dτ (applies to any material)
all space
Energy in Linear Dielectric Systems
▶ Now, if the medium is a linear dielectric, i.e., D = ϵE,
1 1
δ(D · E) = δ(ϵE2 ) = ϵ(δE) · E = (δD) · E
2 2
▶ We then get,
∫
1
δWlinear = δ(D · E) dτ
2 all space

▶ Finally, we buildup the free charge density up from zero to the

final configuration: ρf , which corresponds to integrating16
from zero field up to the final field D (E)
∫ ∫ D
1
Wlinear = δ(D · E) dτ
2 all space 0
giving us
∫
1
Wlinear = D · E dτ
2 all space
16
Such an integration is known as a functional integration, in which the
domain of an integral is no longer a region of space, but a space of functions..
Example: metal sphere surrounded by a linear dielectric

▶ Problem: A metal sphere of radius a carries a charge Q. It’s

surrounded, to radius b, by linear dielectric of permittivity ϵ.
▶ 1. Find the potential V at the center (relative to ∞),
2. the bound charges ρb , and σb ,
3. the total energy stored in the configuration WT , and,
4. the energy stored in the ”springs” WS inside the dielectric.
Example: metal sphere surrounded by a linear dielectric
D and E fields

H
▶ Using Gauss’ law for D, i.e, D · da = qf , to obtain17 D, we find,
{
0 r≤a
D= Q
4πr2 r̂ r>a

▶ …and thus using D = ϵE,



0 r≤a
Q
E= 2 r̂ a<r<b
 4πϵr
 Q
4πϵ0 r2 r̂ r>b
17
We can only do this because we have spherical symmetry in this problem
Example: metal sphere surrounded by a linear dielectric
potential V at center

▶ Calculating the potential at the center r = 0

∫ 0 ∫ b (
) ∫ a( ) ∫ 0
Q Q
V=− E · dl = − dr − dr − (0)dr
∞ ∞ 4πϵ0 r2 b 4πϵr2 a
( )
Q 1 1 1
= + −
4π ϵ0 b ϵa ϵb

▶ As for P, since,
ϵ0 χe Q
P = ϵ0 χe E = r̂ (a < r < b)
4πϵr2
Example: metal sphere surrounded by a linear dielectric
bound charges

▶ The bound volume charge density is,

ρb = −∇ · P = 0

▶ While the surface bound charge density is,

{
− ϵ0 χe Q2 inner surface
σb = P · n̂ = ϵ0 χ4πϵa
eQ
4πϵb2
outer surface
▶ Note that σb (r = a) is negative, which is natural, since the
positively charged metal sphere would induce negative charges
adjacent to it.
Example: metal sphere surrounded by a linear dielectric
total stored energy

▶ In order to calculate the total stored energy in the configuration

we employ,
∫
1
WT = D · E dτ
2 all space
[∫ ( ) ( ) ∫ ∞( ) ( ) ]
b
1 Q Q Q Q
= r̂ · r̂ dτ + r̂ · r̂ dτ
2 a 4πr2 4πϵr2 b 4πr2 4πϵ0 r2
[ ∫ ∫ ]
Q2 1 b 1 1 ∞ 1
= (4π) dr + dr
32π 2 ϵ a r2 ϵ0 b r2
[ ( ) ( )]
Q2 1 1 1 1 1
= − +
8π ϵ a b ϵ0 b
Example: metal sphere surrounded by a linear dielectric
[END OF LEC. 7]
’spring’ energy
▶ For the stored electrostatic energy in the configuration we invoke,
∫
ϵ0
WE = |E|2 dτ
2 all space
[∫ ( )2 ∫ ∞( )2 ]
b
ϵ0 Q Q
= dτ + dτ
2 a 4πϵr2 b 4πϵ0 r2
[ ∫ ∫ ]
Q 2 ϵ0 1 b 1 1 ∞ 1
= (4π) 2 dr + 2 dr
32π 2 ϵ a r2 ϵ0 b r2
[ ( ) ( )]
Q2 ϵ0 1 1 1 1
= − +
8π ϵ2 a b ϵ0 b

▶ The energy stored in the ”springs” is then

( )
Q2 (ϵ − ϵ0 ) 1 1
WS = WT − WE = −
8πϵ2 a b