INDUCED DIPOLES IN DIELECTRICS

ISOTROPIC DIELECTRICS


For isotropic medium (no preferred direction) the average
Faraday found that placing an insulator between the plates



dipole moment per atom is proportional to (in the
of a capacitor (connected to a battery) increased the
same direction).
charge on the plates.


The dipole moments per atom reinforce each other in a
If the insulator filled the space between the plates, the


region with uniform field ( atoms per unit volume).
charge increased by a factor r = dielectric constant or
relative permittivity of that material.


(dipole moment per unit volume)
Dielectric constant of most insulating materials is in the




range 1 to 10 (but can be ). Consider element of volume

with dipole moment

Dielectric medium has no free charges; positive and



negative charges average over the scale of an atom. End faces acquire a charge







such that









Surface charge density














More generally: , where is the



dipole moment perpendicular to face .




External field causes the positive bits to move one way

Faraday’s dielectric acquired surface charges, which had


and the negative bits the other induced dipole.
to be cancelled by charges from the battery.
 INDUCED DIPOLES IN DIELECTRICS
POLARISATION CHARGE DENSITY


3
/
If is uniform, the polarisation charges appear on the 0

.
1
2

0
3

Net polarisation charge due to .

4

surface of the dielectric.


/
/

Stack together many

.
5
7

Similar charges on other faces due to

elements; surface charges
:
<

8
9
1
;

cancel for the interior, which

p
remains electrically neutral.



Alternative derivation:


If is non-uniform, is non-uniform too,
=


Contributions no longer cancel in interior of dielectric. Surface has surface charge
<
?

>
1
;



polarisation charge density p.
density .
8

No free charges




Charge on -face Surface charge density must have

#
$
%
%

!
&
'

"
= come from the interior of the volume.

#
$

%
%
%

!
*


&
'

(
"
@
E

A
A

B
C
B
G

8
>
F
9
1

Charge conservation p






(

+
@
@
@
E

= net positive charge at interface

-
!
A
A
A
A

B
C
B
C
B
C
B

<
<
:
<

>
F

1
1
1

;
;
;

"

#
#
$
#
$
$

%
%
%
%
%
%

!
!
*
,

&
'
&
'

+
+

"
"
-
=
:
<
G

-
!
8
9
1

"

!
p


+

"
-

Polarisation charge per unit volume due to is Q.E.D.

 GAUSS’ THEOREM IN DIELECTRICS
PERMITTIVITIES, SUSCEPTIBILITIES

H
`
a
Dielectric contains:

_
In isotropic medium is parallel to .

I
(1) free charges free ;

b
a

I
(2) polarisation charges p. is also parallel to .

b
a

d
d

c
e
Write r .

H
_
d
Gauss’ theorem: r is a dimensionless constant for the medium.

J
S

P
U
W
= Faraday’s permittivity

K
L
K

N
V

T
I
I

M
Q
free p

R
= dielectric constant.

X
`
a

_
f
d

c
e

divergence Sometimes we use

f

theorem

X
is the susceptibility of the medium.

S
S

P
U
W

K
K

Y
N
Y
[

T
T
I

M
M

Z
g
h

_
d
f

Q
free

R
_

Related via r .
d

U
W

Y
N
[

\
V

Q
I

M
We generally use r when we are interested in the electric
free
f

field, and when we are interested in the properties of the

N
[
]

H
Q
medium.

R
We use a lot, so we define a new field

]
N
[

^
Q
_

R
For homogeneous dielectric can simply replace
j

d
d
d

e
e

r in equations:
n

k
m

H
free
c
i

Field lines of begin and end only on free charges: Poisson’s equation:
d
d
e

Y
]
r

M
free
 p

o
AND


PROPERTIES OF

~
PROPERTIES OF AND (CONT.)

r
s

q
Fixed distribution of determines . If a dielectric


€
medium is introduced to fill all space, the free charges All charges (free and polarisation) are sources of , so

t
must be increased by a factor r to counteract polarisation field lines of can begin and end on polarisation charges.

r
u


€
charges ( fixed varies).
determines the force on a charge at any point. It is

„…

w
‚
ƒ

q
Fixed distribution of free charges determines related to the potential via .

†

x
w
{

y
( free )
We must think of as more fundamental than for

s
If a dielectric medium is introduced, the field is reduced dielectrics.

v
r

t
€
by a factor r ( fixed varies).
At a boundary between dielectrics, where there are no

q
Effective field in dielectric free charges:

s
s

|
…

= (i.e. decreases). (1) must be continuous.

‡
ˆ

(2) must be continuous.

‰
Š

(3) must be continuous.

w
w

q
€

Free charges are the only source of . Field lines of

Learn this.
can only begin and end on free charges.
 BOUNDARY CONDITIONS IN DIELECTRICS

 
¡
BEHAVIOUR OF AND AT BOUNDARIES

‹
No free charges.

‹
For consider Gaussian
‘pill-box’ in plane of the boundary.

’
£
¤

¢
¢

“
”


Œ
Œ




‘
Gauss



Lines of are continuous Lines of can end at

Ž
Ž
“refracted” at boundary. polarisation charges.

•
•
•

‘
‘


—
—
—

–
˜
is continuous
«
«

£
¦
¦

¢
¨©ª
¨©ª

¥
¬

§
§
­
­

B.C. for
«
«

¤
®
®

™
¯
¯

¢
ª
ª

‹
¥
°
°

§
§
­
­

For consider loop shown B.C. for

£
¤

±
±

¬
²

r
«
«

´
´

³
¨©
¨©

±
±

§
§
­
­

š
’

“
”
“œ


™
™




›
Stokes

›››



Ž
Ž
«

³
´

¨©

±
¬

r constant





‘
‘

ž
ž
ž

–
˜
is continuous
¢

N.B. Not really like refraction!

‹
For , the potential is single-valued, so must be
continuous at the boundary.
 POLARISATION AND ELECTRIC FIELD

µ
RELATIONSHIP BETWEEN AND
Ã

Â
Ä

Example:(b) Thin slab of dielectric perpendicular to .

¸
½

·
º»
(1) is uniform (symmetry)

¹
¼
Related by ,

½
where is the electric field within dielectric. (2) is continuous
É
É
Ì

¾
¾
È
È

Ä
Ä
Ä

·
¹
But in out in general. in

Ì
É

È
Í

Ê
Ä
Question: Block of dielectric is inserted into an electric field (3) in

½
¸
½

¼
¼
. What is in terms of ?
Î
Ð

É
É

Ï
Ë

Ç
È
È
Í

Ä
Ä
Ä

in

·
É

Good question! Answer depends on the shape of the

Ä

Ì
É
É

Ç
È
Ñ

Ê
Ê

Ä
Ä

Ï
Ï

Ë
Ë

Í
Í

dielectric. in

·
Å
Ã

¼
Â
Ä

Example: (a) Long thin rod parallel to So the relation between and is dependent on the
shape of the dielectric.
Â

These two examples are the extreme cases:

Í
Î
Ð

Ì
É
Ó
Ï

Ô
Ô

È
Ò

Ê
Ä
Ä

Ï
Ë

In general:

¾
½
¾
Ò
Í

¹
À

¼
Must have in ( continuous)

¸
½

Á
Õ

º»
Ï
Ö

¹
¼
¼
Ò
Ê

e.g. sphere
Õ

Ï
×

Ò
Ê

cylinder
 DIELECTRIC SPHERE IN UNIFORM FIELD

DIELECTRIC SPHERE IN UNIFORM FIELD II

Ø
Solve via potential theory

Ù
ÜÝ

Ú
Û
õ

ô
ö
is continuous at the boundary.

û
ü
þ

õ


ù
ù
ù

÷
÷

ö
ú

ø
þ

Ù
Þ
( outside)
ý

Ø
r r
þ
ÿ

Guess that in and are uniform inside the sphere (very

difficult if not. . . )


õ



ù
ù

ä
ú

Ý
ß
áâã

Ú
Û
,in r in



in in



ç
ä

áâã


õ




ä
÷

Ý
ß
ö
ú
ú

áâã

Ú
å

Û
è


,out

à
out




÷
ú
ø

uniform dipole Using gives

in

Ø
à
é







Potential continuous at boundary







ú
ú

ë
ì

ç
r

ä
ä

ß
ß
æ

ê
áâã
áâã

é
é

Ú
å

Û
Û
è

é
in






ç
ø










r
÷
÷

ú
ú

ä
ß
ß

å
Û
ù
ù

in

í
OK for all if

é
in r r

Ø


î



continuous yields the same condition.






ð

ù
ù
ù

÷
÷
÷

ú
ú
ú
ú
ú




ß
ß
ù


in

Ú
Ú

î
Û
ñ
ä
Check: r

à


ù


ç
ä

ã
ó
( r )

ä
ä

ß
ß

ò
ò
æ

ã
ã

ó
ó

Ú
å

Û
Û
í
(consistent)

é
in
 DIELECTRIC CYLINDER IN UNIFORM FIELD

DIELECTRIC SPHERE IN UNIFORM FIELD III

$



#



"



!

%
Solve via potential theory.
-

&
)
*+,

'
(

in in
/
-

,
*+

&
)
*+,

'
.

(


out
0
0
3

1


uniform dipole ( )
%

Boundary conditions:
&
)

5

'

(1) (or ) continuous at ;

6

5

'

(2) continuous at .
;


"
3

)
)


5
<

'
.
(

Figure shows field lines (thick, blue) and equipotentials

8

in





solve. . .

"
3

)
)

9
:
5

'
.
=

(thin, red) for a dielectric sphere with r .

r in
%

Can do this for other geometries too. . .

 Hello world agai
Field lines of > and equipotentials Field lines of > and equipotentials
@ @
? ?
A A
for a oblate spheroid with r . for a prolate spheroid with r .
DIELECTRIC SPHEROIDS IN UNIFORM FIELD

B
These figures illustrate the fact that, in general, the polarisation inside
a dielectric is not parallel to the external electric field.
C C C C
K
T R R
T T
J K J
T
H H
H H
V V

of charge
G G

take image
P NO M L P NO M L
E
F
F
I
S
E

For region 1:
F

at
D D
E

at
W W
K F
R
Q
T T
H G
F
H G
K
R E

situation by images.
X S X S
D
S

Point charge distance

T T
I
D

IMAGES IN SEMI-INFINITE DIELECTRICS

E
D
T R
F
Y S Y S

It is possible to model this

For region 2: use potential

U
T T

from boundary of two dielectrics.

What are the fields in the media?

 IMAGES IN SEMI-INFINITE DIELECTRICS II
IMAGES IN SEMI-INFINITE DIELECTRICS III

Hello world againag

. {
|

and equipotentials
~

z
}

\
\

[
]

Z
Z
For For


_
`

^
b
b

a
c
e
(1) Continuity of (or ) at
for

i
i
i

f
g
g
g

d
(A)
Field lines of

^
b
b

k
c
e
(2) Continuity of at

i
i

g
q

`
d

e
l
w

e
e
e
v
e

r
u

m
h
h

no
q
s
t

p
r
u
i

q
g
g

`
d

c
l
w

e
e
e
v
e

r
u

m
h
h

no
q
s
t

p
i
i
i

r
u

^
o
g
g
o
g

c
e

l
. 

B.C. satisfied if (B)

{

and equipotentials
~

^
}

from (A) and (B)

|

x
{

o
o
o

c
c
e

i
i
i

f
g
g
y
g
g

d
d

h
h

o
o
o
o
for

c
e
c
e
Field lines of
 MORE IMAGES IN DIELECTRICS ENERGY DENSITY IN DIELECTRICS

‰
In electrostatics the total energy is given by

. †
Ž

ˆ
Œ


Š
’

‘

‡
‹

free

. 
„
…
and †

ƒ
‰

Hello world againag

The above expression is still valid in the presence of linear

‚
“
•


dielectrics. ( )

and equipotentials
€
‰
Gauss:
Ž

Œ


Š
’–
“

‹
—

Field lines of
for a sphere with
You can’t do this by images — it needs
an expansion in both
Œ
š
œ

Ÿ


›
“
’
“
–
’

˜


‹
™
—
—


 
 
•

divergence
¡

defn. of
theorem
 

¢
Ž

Œ
Œ

œ



£


’–
’
•
“

˜


‹
™
—
—

.




„

ƒ
‰

The surface term tends to zero for a localised charge


¥

and equipotentials
distribution: over a large surface, radius :

€
¢


Œ

Œ

¦
¥
’
¥
“
¥
¥
«

¨
¨
¨

©
ª

”
§
”
”
”

.
Œ
•
“

‰
—

Field lines of Energy localised in fields: energy density

for a cylinder with
This problem can be done by the method
of images — try it!
 ENERGY IN DIELECTRICS — NON-EXAMINABLE

LOCAL FIELDS IN A DIELECTRIC MEDIUM

¬
If the response of the dielectric is not linear we have to be
more careful.

½
Medium composed of non-polar molecules

¬
We can calculate the work done in a small changea (i.e. having no permanent dipole moment).

­
­
­

®
³

±
²

¯
½
free . For a continuous medium

¬
¾
Â
Ä
Â

À
À
À

¿
¿

Á
Á

If the sources are localised, we can write (use divergence

°
r

­
­
­

®
´
¶

¯
µ
Â
theorem again) .
is the macroscopic field in the dielectric.

·
¹
º

´
­ ¶
´
¶

¬
is the average polarisation per unit volume.

µ
µ

¸
If medium is linear .

¬
Medium composed of molecules, which become polarised
If the medium is non-dissipative there is a well-defined

¶
´
in the electric field.
relationship between and and process is reversible.
Â

½
Å
Æ

¬
Molecular polarisation is local ,
Â

Otherwise, some of the work done polarising the medium

¶
´
where local is the electric field at the moecule ;
is dissipated as heat, and the relation between and
Æ

is the polarisability
Ê

depends on the history — the medium displays

Æ
È
À
É

¿
Á

(e.g. for conducting sphere).

hysteresis.
½

¬
How do we relate the microscopic polarisability to the
We will return to this when studying ferromagnetism.
macroscopic electric field and dielectric constant?

»¼
a
We use to indicate that the work done is not necessarily a perfect
differential (see later in Thermodynamics).
 MICROSCOPIC VIEW OF DIELECTRICS

Ë
First attempt: suppose the field at the molecule local is LOCAL FIELDS IN DIELECTRICS

Ì
the same as the average field in the bulk.

Ë
ã
molecules per unit volume
Each molecule contributes
ä

Ï
Í
Ï
Í
Ì
Í
Ì

Î
Ñ
Ñ

Ð
Ð
Ð
to the average .
local

Ò
Ì
Í
Ì
Ì

Î
Ó
Ñ
Ó
Ó

Ð
Ð

Ô
Ô
r

Í
ä
å

Ñ
ã

Ö
Ù

Î
We have used this average to calculate the field and

Ó
Ð

×
Ó
r

Ô
hence the polarisation of the molecule.

Ë
ã
This formula does not agree very well with experiment! We have allowed each molecule to polarise itself! Oops. . .

Ú
Data for C0 at 100 C.
å

ã
We must compensate the local field local to remove the
ä

Ü
à
Ü
à

ß
ß

Ý
Ý

Þ
Þ
r r effect of the molecule on .

Ü
à

Ý
Pressure density density
r
ã

1 7.51 7.49 Idea: (due to Lorentz) spilt the field into 3 parts:

10 7.96 7.69 (1) the molecule itself;

29 8.70 7.70 (2) its surroundings;
97 9.18 7.62 (3) the bulk medium.

Ë
What went wrong? Why is the other formula
(Clausius-Mossotti) better?
 CLAUSIUS-MOSSOTTI EQUATION

LOCAL FIELDS IN DIELECTRICS II

é

è
Want to calculate the field in inside

æ
sphere due to uniformly polarised
Lorentz showed that the contribution to local from
medium outside
molecules in a sphere surrounding the molecule of interest
cancelled exactly.

æ
è
Lorentz considered several
Can do this by subtracting the field
plausible cases (cubic lattice,
of a uniformly polarised sphere
è

random arrangements. . . ).
Let us calculate this directly,

æ
Can you see why his result is reasonable? using potential theory.
ñ

ê
é
îïð

æ
ë
ì

Does it have to be a sphere, rather than any other shape? in in constant

õ
ñ

ò
ö
îï
ð

óô

(Ans: YES)
ë
ú

dipole

æ
óø
í

out
ù

Local field is due to contribution

ö

è
í
ô
ë
ë
ì

of molecules outside the sphere,

÷

Boundary Condition at
ø

in
ù

far enough away to be considered

é
ü

macroscopically. Note sign: in opposes .

ç
ç

æ
è

inside the sphere is increased over outside by the Field inside original sphere surrounded by polarisation:
effect of polarisation charges on the inside of the sphere.
ü

ý
ý
þ

ë
÷

local
ù
 CLAUSIUS-MOSSOTTI EQUATION II

ÿ
Field


(1)


local

ÿ
Field doesn’t depend on how big the sphere was!
(Good. . . )

ÿ
Now use





local (2)

and






r (3)


( r is the bulk permittivity.)






((2) in (1))















(using (3))



r










r

Clausius-Mossotti equation