Physics, asked by asifshek9499, 1 year ago

A capacitor with a non-uniform dielectric

Answers

Answered by tarushhh
0

Electricity and Magnetism I (P331)

M. R. Shepherd

October 17, 2008

A capacitor with a non-uniform dielectric

Consider a parallel-plate capacitor filled with non-uniform dielectric. The

dielectric is still linear, in other words,

P

=

0

χ

e

E

, but this time

χ

e

varies

as a function of position in the dielectric. Let’s say that

χ

e

varies such that

r

=

r

0

+

ax.

Here

r

0

is a dimensionless constant,

x

is the distance from one plate, and

a

is constant with units 1

/

length. Let’s assume the area of the plates is

A

and their separation is

d

. What is the capacitance?

x=0

x=d

x

Remember the strategy for computing capacitance: put a charge

Q

on the capacitor and then compute

the potential difference between the plates in terms of this charge. The capacitance can then be identified

by comparing this expression to

V

=

Q/C

. In order to compute

V

we first need to find the electric field.

Let’s charge up the plates so that we have a free surface charge density of +

σ

f

on the left plate and

σ

f

on the right plate. We can then compute the electric displacement

D

based on the free charge, and obtain

D

=

σ

f

ˆ

x

.

This then gives

E

=

1

D

=

σ

f

0

(

r

0

+

ax

)

ˆ

x

Now, to obtain the potential, we need to compute the line integral of this field from the negative plate

to the positive plate. We have

V

=

E

·

d

l

=

σ

f

0

0

d

1

r

0

+

ax

dx

=

σ

f

0

a

[ln(

r

0

+

ad

)

ln(

r

0

)]

Let’s rewrite the difference of logs as a log of a ratio and substitute in

σ

f

=

Q/A

. This allows us to identify

the capacitance

V

=

Q

ln(1 +

ad/

r

0

)

A

0

a

=

Q

C

=

C

=

A

0

a

ln(1 +

ad/

r

0

)

.

It is interesting to examine this expression in the limit that the dielectric is uniform. In this case we have

a

0 and

r

=

r

0

. This poses a little problem since, in the expression above, substituting in

a

= 0 we have

0

/

ln(1) = 0

/

0. It is helpful to remember that for small values of

y

we can write

ln(1 +

y

) =

y

y

2

2

+

y

3

3

+

...

This means that for

a

very small we have

C

=

A

0

a

ln(1 +

ad/

r

0

)

A

0

a

(

ad/

r

0

)

(

ad/

r

0

)

2

/

2

=

A

0

(

d/

r

0

)

a

(

d/

r

0

)

2

/

2

.

We can see that in the limit

a

0 this expression reduces to

r

0

0

A/d

, which is exactly what we expect for

a parallel plate capacitor with dielectric constant

r

0

.

Exercise:

Compute the bound charge densities

σ

b

on

both sides of the capacitor. Which side should have the higher bound charge density? In this case,

ρ

b

6

= 0 –

compute it also. Electricity and Magnetism I (P331)

M. R. Shepherd

October 17, 2008

A capacitor with a non-uniform dielectric

Consider a parallel-plate capacitor filled with non-uniform dielectric. The

dielectric is still linear, in other words,

P

=

0

χ

e

E

, but this time

χ

e

varies

as a function of position in the dielectric. Let’s say that

χ

e

varies such that

r

=

r

0

+

ax.

Here

r

0

is a dimensionless constant,

x

is the distance from one plate, and

a

is constant with units 1

/

length. Let’s assume the area of the plates is

A

and their separation is

d

. What is the capacitance?

x=0

x=d

x

Remember the strategy for computing capacitance: put a charge

Q

on the capacitor and then compute

the potential difference between the plates in terms of this charge. The capacitance can then be identified

by comparing this expression to

V

=

Q/C

. In order to compute

V

we first need to find the electric field.

Let’s charge up the plates so that we have a free surface charge density of +

σ

f

on the left plate and

σ

f

on the right plate. We can then compute the electric displacement

D

based on the free charge, and obtain

D

=

σ

f

ˆ

x

.

This then gives

E

=

1

D

=

σ

f

0

(

r

0

+

ax

)

ˆ

x

Now, to obtain the potential, we need to compute the line integral of this field from the negative plate

to the positive plate. We have

V

=

E

·

d

l

=

σ

f

0

0

d

1

r

0

+

ax

dx

=

σ

f

0

a

[ln(

r

0

+

ad

)

ln(

r

0

)]

Let’s rewrite the difference of logs as a log of a ratio and substitute in

σ

f

=

Q/A

. This allows us to identify

the capacitance

V

=

Q

ln(1 +

ad/

r

0

)

A

0

a

=

Q

C

=

C

=

A

0

a

ln(1 +

ad/

r

0

)

.

It is interesting to examine this expression in the limit that the dielectric is uniform. In this case we have

a

0 and

r

=

r

0

. This poses a little problem since, in the expression above, substituting in

a

= 0 we have

0

/

ln(1) = 0

/

0. It is helpful to remember that for small values of

y

we can write

ln(1 +

y

) =

y

y

2

2

+

y

3

3

+

...

This means that for

a

very small we have

C

=

A

0

a

ln(1 +

ad/

r

0

)

A

0

a

(

ad/

r

0

)

(

ad/

r

0

)

2

/

2

=

A

0

(

d/

r

0

)

a

(

d/

r

0

)

2

/

2

.

We can see that in the limit

a

0 this expression reduces to

r

0

0

A/d

, which is exactly what we expect for

a parallel plate capacitor with dielectric constant

r

0

.

Exercise:

Compute the bound charge densities

σ

b

on

both sides of the capacitor. Which side should have the higher bound charge density? In this case,

ρ

b

6

= 0 –

compute it also.

Similar questions