Conductors

A conductor (typically, a metal like Cu, Ag etc. or ionic conductors like HCl or NaCl dissolved in water) allows free movement of charges. They have low resistivity 108 m as compared to typical insulators like quartz, glass etc. which have resistivity of the order of 1017 m. However, the property that really distinguishes a metal from insulators or semi-conductors is the fact their temperature coefcient of resistivity is positive while that of semi-conductors is negative. The electric eld inside a conductor is zero. In an equilibrium situation, there cannot be an electric eld inside a conductor as this would cause charges (electrons or ions) to move around. In the presence of an external eld, there is charge separation inside a conductors with opposite charges accumulating on the surface. This creates an internal electric eld which cancels the effect of the external eld in such a way that the net electric eld inside the conductor volume is zero.

Eint

+ + + + + + + +

Eext
Charge density inside a conductor is zero. This follows from Gausss law E = / As E = 0, the charge density = 0. 1
0

(This does not suggest that there is no charge inside, only that the positive and negative charges cancel inside a conductor.) Free charges exist only on the surface of a conductor. Since there is no net charge inside, free charges, if any, have to be on the surface. At the surface of a conductor, the electric eld is normal to the surface. If this were not so, the charges on the surface would move along the surface because of the tangential component of the eld, disturbing equilibrium.

E=0

Induced Charges in a conductor:

The above properties of a conductor inuence the behaviour of a conductor placed in an electric eld. Consider, for instance, what happens when a charge +q is brought near an uncharged conductor. The conductor is placed in the electric eld of the point charge. The eld inside the conductor should, however, be zero. his is achieved by a charge separation within the conductor which creates its own electric eld which will exactly compensate the eld due to the charge +q . The separated charges must necessarily reside on the surface. Another way of looking at what is happening is to think of the free charges in the conductor being attracted towards (or repelled from) the external charge. Thus the surface of the conductor towards the external charge is oppositely charged. To keep the charge neutrality, the surface away from the external charge is similarly charged.

+ + + + + +

++

+q

Example 1 : A charge Q is located in the cavity inside a conducting shell. In addition, a charge 2Q is distributed in the conducting shell. Find the distribution of charge in the shell. What is the electric eld in the region outside the shell.

+ + + + + + + + + + +

+ + + + +Q + + + +

Take a gaussian surface entirely within the conducting shell, completely enclosing the cavity. Everywhere on the gaussian surface E = 0. The ux and therefore, the charge enclosed is zero within the gaussian surface. As the cavity contains a charge Q, the surface of the cavity must have charge Q. As the conductor has distributed charge 2Q, the charge on the outside surface is 3Q. The principle illustrated in the above problem is known as Faradays Cage. If a hollow conducting box is kept in an electric eld, the charges in the cavity are redistributed in such a way that the electric eld inside the cavity is zero. This is used to provide an enclosure for sensitive electronic equpment which must be kept free of external electronic disturbance. Example 2 : Calculate the electric eld outside a conductor carrying a surface charge density .

E + + + + + + ++ + +++

dS + + r + + + + + + + L + + + + + + + + + +

Take a gaussian pillbox in the shape of a cylinder of height h with h/2 inside and h/2 outside the conductor. Lat the cross sectional area be dS normal to the surface. The electric eld is normal to the surface. As the eld inside is zer and there is no tangential component of the eld at the surface, the ux goes out only through the outer cap of the cylider. The charge enclosed is dS and the ux is EdS . The electric eld is normal to the surface. applyinng Gausss law E= . Exercise : Two parallel, innite plates made of material of perfect conductor, carry charges Q1 and Q2 . The plates have nite thickness. Show that the charge densities on the two adjecent inside surfaces are equal and opposite while that on the two outside surfaces are equal. (Hint : Field inside the plates due to four charged surfaces must be zero.) Poissons and Laplaces Equations Differential form of Gausss law, E = Using E = V , E = (V ) = 2 V 4
0

so that 2 V = This is Poisson equation. In cartesian form,

2V 2V 2V + + = 2 2 2 x y z 0 For eld free region, the equation becomes Laplaces equation 2 V = 0 Equipotential surface Equipotential surfaces are dened as surfaces over which the potential is constant V (r) = constant At each point on the surface, the electric eld is perpendicular to the surface since the electric eld, being the gradient of potential, does not have component along a surface of constant potential.

We have seen that any charge on a conductor must reside on its surface. These charges would move along the surface if there were a tangential component of the electric eld. The electric eld must therefore be along the normal to the surface of a conductor. The conductor surface is, therefore, an equipotential surface. Electric eld lines are perpendicular to equipotential surfaces (or curves) and point in the direction from higher potential to lower potential. In the region where the electric eld is strong, the equipotentials are closely packed as the gradient is large.

3 kV 2 kV 1 kV 0.5 kV 1 kV 2 kV
P x

= 0

The electric eld strength at the point P may be found by nding the slope of the potential at the point P. If x is the distance between two equipotential curves close to P, V E= x where V is the difference between the two equipotential curves near P. Example 3: Determine the equipotential surface for a point charge. Solution : Let the point charge q be located at the origin. The equation to the equipotential surface is given by V (x, y, z ) = 1 4 0 q x2 + y 2 + z 2 = V0 = constant

Equipotential surfaces (magenta) and field lines (blue) for a positive charge.

Thus the surfaces are concentric spheres with the origin (the location of the charge) as the centre and radii given by R= q 4 0 0

The equipotential surfaces of an electric dipole is shown below.

Electric Field and Equipotential lines for an electric dipole

Example 4 : Determine the equipotential surface of an innite line charge carrying a positive charge density . Solution : Let the line charge be along the z- axis. The potential due to a line charge at a point P is given by V (r ) = ln r 2 0 where r is the distance of the point P from the line charge. Since the line charge along the z-axis, r = x2 + y 2 so that V (r ) = ln(x2 + y 2 ) 4 0

The surface V = constant = V0 is given by ln(x2 + y 2 ) = i.e. x2 + y 2 = e 4 0 V0

4 0 V0

which represent cylinders with axis along the z-axis with radii r = e
zaxis ++++++++++++++++++++++++++++++
2 0 V0

2
As V0 increases, radius becomes smaller. Thus the cylinders are packed closer around the axis, showing that the eld is stronger near the axis.

Example 5 : Consider a charged sphere of radius R containing charge q , completely enclosed by a spherical cavity of inner radius a and outer radius b. Calculate the charge density on all surfaces and potential everywhere. Solution : As eld inside the conductor is zero, by taking a Gaussian surface completely in the region a < r < b, we must have net charge enclosed by such a surface to be zero. To compensate for the charge q that exists on the surface of the inner sphere, the charge on the inside surface of the shell must be q . Since the shell is charge neutral, a charge +q must, therefore, appear on the outside sur- Gaussian surface face of the shell. For r > b, the eld is 1 q r E= 4 0 r 2 The corresponding potential is
r

+q R q b a +q

V (r ) =

1 q 1 q dr = 2 4 0 r 4 0 r

1 q 4 0 b Since the eld between a and b is zero, this is also the potential at all points from r = b to r = a. 1 q V (a r b) = 4 0 b For R < r < a, the potential is given by V (b) = V (r ) =
R 1 q 1 q 2 4 0 b a 4 0 r q q q 1 + = 4 0 b R a

At r = b, the potential is

If the outer surface is grounded, the potential on the shell becomes zero. There is no charge on the outer surface. However the inner surface must have a charge q to keep the eld in the shell zero,
R

V (0) = V (R) =
a

E dl =

1 1 1 4 0 R a

Exercise : Determine the equipotential surface of an innite plane with charge density . Laplaces Equation Let us look at Laplaces equation in one dimension. It becomes d2 V =0 dx2 which has the solution V = mx + c The solution shows two important characteristics of the solution of Laplaces equation, which are not immediately obvious in higher dimensions. The rst property is the potential at a point can be expressed asaverage of potentials at neighbouring points. For instance, 1 V (x) = (V (x + x0 ) + V (x x0 )) 2 This also illustrates the second property of the solutions, viz., the solution has no local minimum or maximum. If it did, it would not be possible to express the function as average of values at neigbouring points. To see this consider a function a f (x, y ) = (x2 + y 2 ) 4 in two dimensions, which does not satisfy Laplaces equation as 2 V = 2f 2f + =a x2 y 2

The function has a positive curvature everywhere and there exists a local minimum at x = 0, y = 0. The function looks like the following. 10

V(x,y) 18 17.5 17 16.5 16 15.5 15 14.5 14 13.5 13 12.5 12 11.5 11 10.5 10 9.5 9 8.5 8 7.5 7 6.5 6 5.5 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 -3 -2 -1 -1 0 1 2 y 3

18 16 14 12 10 8 6 4 2 0

x 0

3 -3

-2

Consider, on the other hand, a function V (x, y ) that satises Laplaces equation a V (x, y ) = (x2 y 2 ) 4 The function has no minimum or maximum and looks like the following. It has a saddle point at x = 0, y = 0.

11

V(x,y) 9 8.5 8 7.5 7 6.5 6 5.5 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 -0.5 -1 -1.5 -2 -2.5 -3 -3.5 -4 -4.5 -5 -5.5 -6 -6.5 -7 -7.5 -8 -8.5 -9 -3 -2 -1 -1 0 1 2 y 3

10 8 6 4 2 0 -2 -4 -6 -8 -10

x 0

3 -3

-2

An interesting consequence of Laplace equation is Earnshaw Theorem which states that a charge cannot be held in stable equilibrium only by electrostatic forces. For instance, suppose we position a charge Q exactly at the centre of a cube which has a positive charge q at each of its eight corners. We would expect the charge to be in equilibrium as it is being pulled equally in all directions. However, this will not be a stable equilibrium because at the centre, there being no charge density, Laplace equation is obeyed. Thus there cannot be a minimum of the potential V and hence of potential energy QV of the charge at the centre. Consider again the case of cavity in a conductor. If the interior of the cavity does not contain any charge, Laplace equation is obeyed. Thus the potential has no minimum or maximum inside the cavity. Further, since the boundary of the cavity is an equipotential, the potential inside the cavity is also constant. Uniqueness Theorem : This theorem states that the solution of Laplaces equation is uniquely determined by the values of potential on the boundaries. Suppose V1 and V2 are two potentials which satisfy Laplaces equation in some region with identical coundary conditions, i,e V1 (boundary ) = V2 (boundary ). Consider a function V3 = V1 V2 . This satises Laplaces equation with the condi12

tion V3 (boundary ) = 0. However, as V3 does not have a minimum or a maximum in the region, its value has to be the same value as its value at the boundary, i.e. V3 is constant. Hence V1 = V2 . Laplaces Equations in 3-dimensions We will consider the solutions of Laplaces equations in problems with spherical geometry having azimuthal symmetry. The equation to be solved is 2 V = 1 r 2 r r2 V r + 1 r 2 sin sin V + 1 2V =0 r 2 sin 2

where we have explicitly written down the Laplacian operator in spherical polar coordinates. For problems with azimuthal symmetry, V / = 0 so that we have r r2 V r + 1 sin sin V =0

The equation above is conveniently solved by a technique called separation of variables where we write the function V (r, ) as a product of two functions, one R(r ) which is a function of radial variable r only and the other a function ( ) which is a function of the angle variable alone. Writing V (r, ) = R(r )( ) and dividing throghout by R, we get 1 R r r2 R r + 1 sin sin =0

Since the two terms on the left depend on two independent variables, this equation can be satised only if each of the term equals to constants of opposite sign. We write R 1 r2 R r r 1 sin sin = l(l + 1) = l(l + 1)

We will not attempt to solve these equations but merely quote the results. The solution of the angular equation is in terms of what are known as Legendre Polynomials. Pl (cos ).These are polynomials of degree lin cosine of angle . The rst few polynomials are as follows : 13

P0 P3 P4 P1

P0 (cos ) = 1 P1 (cos ) = cos 1 P2 (cos ) = (3 cos2 1) 2 1 (5 cos3 3 cos ) P3 (cos ) = 2

0.5 Pl(cos )

0 P2

-0.5

-1 -1 -0.5 0 cos 0.5 1

The solution of radial equation is consists of a power series in r and 1/r . The complete solution is

V (r, ) =
i=0

Al r l +

Bl r l+1

Pl (cos )

(A)

We will illustrate the use of these solution by an example. Example 6 : Consider an uncharged conducting sphere in a uniform electric eld and determine the potential at all points in space. Solution :

The sphere, being a conductor, is an equipotential. Let the potential be zero. Far from the sphere, the eld is uniform. Let the eld strength be E0 and be in z-direction, The boundary conditions are : V = 0 at r = R V = E0 z = E0 cos for r R.

+ + + + +

Using Eqn. (A) and substituting the rst boundary condition, we get a relationship between Al and Bl Al R l + Bl =0 Rl+1 14

Thus Bl = Al R2l+1 . Thus, we have

V (r, ) =
i=0

Al r l +

R2l+1 r l+1

Pl (cos )

For r

R, we may neglect the second term in bracket and get

Al r l Pl (cos ) = E0 r cos
i=0

On comparing both sides, we get l = 1 which gives A1 = E0 . Substituting these we get R3 V (r, ) = E0 (r 2 ) cos r The induced charge density is =
0

V |r +R = 3 0 E0 cos r

It can be seen that the charge density is positive in the upper hemisphere and negative in the lower hemisphere. Dielectrics A conductor is characterized by existence of free electrons. These are electrons in the outermost shells of atoms (the valence electrons) which get detatched from the parent atoms during the formation of metallic bonds and move freely in the entire medium in such way that the conductor becomes an equipotential volume. In contrast, in dielectrics (insulators), the outer electrons remain bound to the atoms or molecules to which they belong. Both conductors and dielectric, on the whole, are charge neutral. However, in case of dielectrics, the charge neutrality is satised over much smaller regions (e.g. at molecular level). 2.9.1 Polar and non-polar molecules : A dielectric consists of molecules which remain locally charge neutral. The molecules may be polar or non-polar. In non-polar molecules, the charge centres of positive and negative charges coincide so that the net dipole moment of each molecule is zero. Carbon dioxide molecule is an example of a non-polar molecule.

15

+8e

+6e

+8e

Oxygen atom

Carbon atom

Oxygen atom

Hydrogen atom +e Hydrogen atom +e

+8e

In a polar molecules, the arrangement of atoms is such that the molecule has a permanent dipole moment because of charge separation. Water molecule is an example of a polar molecule.

Oxygen atom

When a non-polar molecule is put in an electric eld, the electric forces cause a small separation of the charges. The molecule thereby acquires an induced dipole moment. A polar molecule, which has a dipole moment in the absence of the electric eld, gets its dipole moment aligned in the direction of the eld. In addition, the magnitude of the dipole moment may also increase because of increased separation of the charges.

16

E=0

E=0 +

A nonpolar molecule in an Electric Field

A polar molecule in an Electric Field

Dielectric in an Electric Field A dielectric consists of molecules which may (polar) or may not (non-polar) have permanent dipole moment. Even in the former case, the dipoles in a dielectric are randomly oriented because dipole energies are at best comparable to thermal energy.

+ + +

+
+ +

+ + + +

+
+ +

+ +
+

Randomly oriented dipole in a dielectric (E=0)

+ +

+ +
+ +

+ + +
+

+
+

+ +

Polarised Dipoles in an electric field

When a dielectric is placed in an electric eld the dipoles get partially aligned in the direction of the eld. The charge separation is opposed by a restoring force 17

due to attaraction between the charges until the forces are balanced. Since the dipoles are partially aligned, there is a net dipole moment of the dielectric which opposes the electric eld. However, unlike in the case of the conductors, the net eld is not zero. The opposing dipolar eld reduces the electric eld inside the dielectric. Dielectric Polarization Electric polarization is dened as the dipole moment per unit volume in a dielectric medium. Since the distribution of dipole moment in the medium is not uniform, the polarization P is a function of position. If p(r ) is the sum of the dipole moment vectors in a volume element d located at the position r , p(r) = P (r )d It can be checked that the dimension of P is same as that of electric eld divided by permittivity 0 . Thus the source of polarization eld is also electric charge, except that the charges involved in producing polarization are bound charges. Denoting the local bound charge density by b , one can write P = b The equation above is obtained in a manner that is identical to the way we derived the equation E = / 0 . The absence of the factor o in the equation is because of the dimensional difference between E and P while the minus sign arises because the dipole moment vector (and hence the polarization) is dened to be directed from negative to positive charge as against E which is directed from positive to negative charge. Clearly, if polarization is uniform, the volume density of bound charges is equal to zero. Even in this case, there are surface bound charges given by the normal component of the polarization vector. Summarizing, we have, P = b P n = b We will derive these relations shortly. Free and Bound Charges The charge density of a medium consists of free charges, which represent a surplus or decit of electrons in the medium, and bound charges. The term free charge is 18

used to denote any charge other than that due to polarization effect. For instance, the valence charges in a metal or charges of ions embedded in a dielectric are considered as free charges. The total charge density of a medium is a sum of free and bound charges = f + b Gausss Law takes the form E = Potential due to a dielectric Consider the dielectric to be built up of volume elements d . The dipole moment of the volume element is P d .
0

f + b
0

The potential at a point S, whose position vector with respect to the volume element is r is dV = 1 P r d 4 o r 2

P d

r S

The potential due to the whole volume is V = 1 4 0 P r 1 d = 2 r 4 0 r 1 ( ) = 2 r r 1 P ( )d r volume

volume

where, we have used

Use the vector identity (Af (r )) = A f (r ) + f (r ) A Substituting A = P and f (r ) = 1/r , 19

P 1 1 ( ) = P ( ) + P r r r we get V = 1 4
o

1 P ( )d r 4 vol

vol

1 P d r

The rst integral can be converted to a surface integral using the divergence theorem giving, V = 1 4 1 P dS r 4 0 1 P d r

surf ace

vol

The rst term is the potential that one would expect for a surface charge density b where b = P n where n is the unit vector along outward normal to the surface. The second term is the potential due to a volume charge density b given by b = P The potential due to the dielectric is, therefore, given by V = and the electric eld E = V 1 = 4 0 b r 1 dS + 2 r 4 0 b r d 2 r 1 4 o b dS 1 + r 4 0 b d r

surf ace

vol

surf ace

vol

Electric Displacement Vector D The electri displacement vector D is dened by D=

0E

+P

which has the same dimension as that of P . The equation satised by D is 20

D =

E + P = b = f

which is the differential form of Gausss law for a dielectric medium. Integrating over the dielectric volume, Dd =
volume volume

f d = Qf

where Qf is the free charge enclosed in the volume. The volume integral can be converted to a surface integral using the divergence theorem, which gives D dS = Q f
surf ace

Thus the ux over the vector D over a closed surface is equal to the free charged enclosed by the surface. Example 5: An uncharged spherical dielectric has polarization vector given by P = kr . Find the electric eld both outside and inside the dielectric. Solution : The dielectric has both bound surface charge and volume charge. The surface charge density is b = P n = kR where R is the radius of the sphere. The volume charge density is b = P = k r = 3k . The eld inside the dielectric is given by Gausss law, 4r 2 E = which gives E= rb kr = 30 0 Qencl
0

4r 3 b 30

The eld outside is zero. Example 6 : Consider a spherical dielectric shell of inner radius a and outer radius b. The space in the region between r = a and r = b is lled with a dielectric hasving polarization P = k r . Determine the eld inside and outside the shell. r 21

Solution : The charge densities are,

outer surf ace b =P n =

k b k a

inner surf ace b = P n =

k r2 For r < a, no charge is included, hence the eld is zero. For a < r < b, the charges enclosed by a Gaussian surface are the surface bound charges on the inner surface and the volume charge within the region. Thus b = P =
r

encl

= 4b +
a

2 in

b 4r 2 dr
r

= 4a

2k

a a = 4ka + 4k (a r ) = 4kr

(k/r 2 )4r 2 dr

Thus E = (k/ 0 r ) r . For a Gaussian surface outside, the total charge enclosed can be similarly calculated to be zero, so that eld is zero. Example 7 :Electric Field Due to Uniformly polarized sphere : Since the polarization is uniform, the bound charge density is zero. Only on the surface, there are bound charges. We have b = P n = P cos where is the angle between the direction of the external eld (zdirection) and a point on the sphere.
P ^ z ^ n

This is, once again, a problem with azimuthal symmetry with no charges inside or outside the sphere. Hence Laplaces equation is satised both in the interior of

22

the sphere and outside.

V (r, ) =
l=0

Al r l +

Bl r l+1

Pl (cos )

For r < R, the second term must vanish since the potential cannot become innity at the origin. Similarly, for r > R, the rst tem must vanish as the potential must be well dened at large distances. For r < R, V (r, ) = Al r l Pl (cos ) and, for r > R Bl Pl (cos ) r l+1 At r = R, the potential is continuous. Hence, V (r, ) = Bl = Al R2l+1 At r = R, while the tangential component of the eld is continuous, the normal component has a discontinuity, above below En = n En
0

Using E = V , V above V below = r r 0 Thus,

l=0

P cos Bl Al r l1 Pl (cos ) |r=R = l +2 r 0

Comparing both sides, we see that only l = 1 term is non-zero. We get, 2B1 P + A1 = 3 R 0 Using B1 = A1 R3 , we get A1 = P/3 V (r, ) =
0

and B1 = P R3 /3

Finally, we get

P r cos for r < R 30 P R3 = cos for r > R 3 0 r3 23

The electric eld inside the sphere is uniform and is equal to V = P/3 0 z . Outside the sphere, the potential has the same form as that of a giant dipole with dipole moment equal to volume of the sphere times the polarization vector, located at the centre, because, V = P R3 cos 3 0 r2 3p/4R3 R3 = cos 30 r2 1 p cos = 4 0 r 2 1 pr = 4 0 r 2

Constitutive Relation Electric displacement vector D helps us to calculate elds in the presence of a dielectric. This is possible only if a relationship between E and D is known. For a weak to moderate eld strength, the electric polarization P is found to be directly proportional to the external electric eld E . We dene Electric Susceptibility through P = 0 E so that D = = +P o (1 + )E =
0E

0 rE

= E

where r = 1 + is called the relative permittivity or the dielectric constant and is the permittivity of the medium. Using differential form of Gausss law for D, we get f 1 E = D = Thus the electric eld produced in the medium has the same form as that in free space, except that the eld strength is reduced by a factor equal to the dielectric constant . Capacitance lled with Dielectric If a material of dielectric constant is inserted between the plates of a capacitor, 24

the eld E is reduced by a factor . The potential between the plates also reduces by the same factor . Thus the capacitance C= Q

+Q Q

increases by a factor . Example: A parallel plate capacitor with plate separation 3.54mm and area 2m 2 is initially charged to a potential difference of 1000 volts. The charging batteries are then disconnected. A dielectric sheet with the same thickness as that of the separation between the plates and having a dielectric constant of 2 is then inserted between the capacitor plates. Determine (a) the capacitance , (b) potential difference across the capacitor plates, (c) surface charge density (d) the electric eld and (e) displacement vector , before and after the insertion of the dielectric . Solution : (a) The capacitance before insertion of the dielectric is C=
0

2 A = 8.85 1012 = 5 109 F d 3.54 103

After the insertion the capacitance doubles and becomes 108 F. (b) Potential difference between the plates before insertion is given to be 1000 V. On introducing the dielectric it becomes half, i.e. 500 V. (c) The charge on each capacitor plate was Q = CV = 5 106 coulomb, giving a surface charge density of 2.5 106 C/m2 . The free charge density remains the same on introduction of the dielectric. (d) The electric eld strength E is given by E= = 2.8 105 volts/meter
0

The electric eld strength is reduced to 1.4 105 volt/meter on insertion. (e) The displacement vector remains the same in both cases as the free charge density is not altered. It is given by D = = 2.5 106 C/m2 . Example : The parallel plates of a capacitor of plate dimensions a b and separation d are 25

charged to a potential difference and battery is disconnected. A dielectric slab of relative permittivity is inserted between the plates of the capacitor such that the left hand edge of the slab is at a distance x from the left most edge of the capacitor. Calculate (a) the capacitance and (b) the force on the dielectric.
yaxis a b d

xaxis x Force

zaxis

Solution : Since the battery is disconnected, the potential difference between the plates will change while the charge remains the same. Since the capacitance of the part of the capacitor occupied by the dielectric is increased by a factor , the effective capacitance is due to two capacitances in parallel , C=
0

b [x + (a x)] d

The energy stored in the capacitor is U= 1 Q2 Q2 d = 2C 2 b 0 x + (a x)

Let F be the force we need to apply in the x-direction to keep the dielectric in place. For an innitisimal increment dx of x, we have to do an amount of work F dx , which will increase the energy strored in the eld by dU , so that F = dU dx

26

the differentiation is to be done, keeping the charge Q constant. Thus F = dQ2 1 2b 0 [x + (a x)]2

Since > 1, F is positive. This means the electric eld pulls the dielectric inward so that an external agency has to apply an outward force to keep the dielectric in position. Since the initial potential difference is given by Q/C , one can express the force in terms of this potential F =
0b

2d

2 ( 1)

This is the force that the external agency has to apply to keep the left edge of the dielectric at x. The force with which the capacitor pulls the dielectric in has the same magnitude. Example 22 : In the above example, what would be the force if the battery remained connected ? Solution : If the battery remained connected Q does not remain the same, the potential does. The battery must do work to keep the potential constant. It may be realised that the force exerted on the dielectric in a particular position depends on the charge distribution (of both free and bound charges) existing in that position and the force is independent of whether the battery stays connected or is disconnected. However, in order to calculate the force with battery remaining connected, one must, explicitly take into account the work done by the battery in computing the total energy of the system. The total energy U now has two parts, one the work done by the external agency F dx and the other the work done by the battery, viz., dQ where dQ is the extra charge supplied by the battery to keep the potential constant. Thus U = F dx + dQ which gives F = Since is constant, we have 1 2 c 2 Q = C U = 27 dQ dU dx dx

Using these 1 dC dC 1 dC F = 2 2 = 2 2 dx dx 2 dx (Note that if the work done by the battery were negnected, the direction of F will be wrong, though, because we have used linear dielectrics, the magnitude, accidentally, turns out to be correct !) In the previous example, we have seen that C= giving b0 dC = (1 ) dx d which is negative. Thus F is positive, as before, F = b 0 2 ( 1) d b0 [x + (a x)] d

Example : The space between the plates of a parallel plate capacitor is lled with two different dielectrics, as shown. Find the effective capacitance.

1 2

d1 d2

Solution :

28

Take a Gaussian pill-box as shown. We have D dS = f ree = 0 as there are no free charges inside the dielectric. Contribution to the integral comes only from the faces of the pill-box parallel to the plates and dS1 = dS2 . Hence, D1 = D 2 = where is the surface density of free charges. be the potential difference between the upper plate and the interface between the dielectric and 2 that between the interface and the lower plate. We have = 1 + 2 = E 1 d1 + E 2 d2 D1 D2 = d1 + d2 1 0 2 0 d1 d2 = + 1 0 2 0 Thus the effective capacitance is given by C= Q = = A

1 2

d1 d2
dz

dS 1

Let 1
dS 2

d1 1

d2 2

A0 d2 + 2 C1 C2 = C1 + C 2
d1 1

where C1 and C2 are the capacitances for parallel plate capacitors with one type of dielectric with separations d1 and d2 between the plates respectively.

29

Example : A capacitor consists of an inner conducting sphere of radius R and an outer conducting shell of radius 2R. The space between the spheres is lled with two different linear dielectrics, one with a dielectric constant from r = R to r = 1.5R and the other with dielectric constant 2 from r = 1.5R to r = 2R. The outer shell has a charge Q while the inner conductor has a charge +Q. Determine the electric eld for r > 0 and nd the effective capacitance. Solution :

The electric eld is radially symmetric and may be obtained by applying Gausss law for the displacement vector D dS = 4r D = Q
2 f ree

+ +

+ + +

+ + + R 1.5R +

For R <

where Qf ree is the free charge enclosed within a sphere of radius r . For r < R, the eld is zero as the free charges are only on the surface of the inner cylinder.

r < 1.5R, the electric eld is E= and for 1.5R < r < 2R, Q D = 0 4 0 r 2

2R

D Q = 2 0 8 0 r 2 For r > 2R, the eld is zero. The elds are radial with the ineer sphere at a higher potential. The potential difference is calculated by taking the taking the E=

30

line integral of the electric eld along any radial line.

2R

= =
R

E dl =
R 1.5R

Edr
2R 1.5R

Q dr + 4 0 r 2
0

Q dr 8 0 r 2

5Q = 48 The effective capacitance is

C=

48 Q = 5

Example : A parallel plate capacitor has charge densities on its plates which are separated by a disance d. The space between the capacitor plates is lled with a linear but inhomogeneous dielectric. The dielectric constant varies with distance from the positive plate linearly from a value 1 to a value 2 at the negative plate. Determine the effective capacitance.
2

As the dielectric is linear, D=

0 E

1+

x E d

dielectric constant

As the insertion of dielectric does not affect free charges, the displacement vector D is remains the same as it would in the absence of the dielectric. Thus D = . electric eld E is given by

Thus the
++++++ ____ d

0 distance from positive plate x

E =

a 0 (x + d)

The eld close to x = d is given by E = / 2 0 , which shows that adjacent to the negative plate there is a positive charge density /2. To nd the effective 31

capacitance, we nd the potential difference between the plates by integrating the electric eld d d d d dx = = ln 2 Edx = 0 0 0 0 d+x so that A A0 Q = = C= d ln 2 The polarization P is given by P =D
0E

x x+d

The volume density of bound charges, given by P = b is found as follows : b = d d x = dx x + d (x + d)2

The bound charge density on the surface, given by n P = P , has a value /2 on the dielectric adjacent to the negative plate (x = d). As the dielectric is charge neutral, this requires a net volume charge of /2 in the dielectric. This can be veried by integrating over the volume charge density b given above. Exercise : A parallel plate capacitor of plate area S and separation d, contains a dielectric of thickness d/2 and of dielectric constant 2, resting on the negatve plate.

+ d/2 d

A potential difference of is maintained between the plates. Calculate the electric eld in the region between the plates and the density of bound charges on the surface of the dielectric. [Ans. eld in empty region = 4/3d, within dielectric = 2/3d, bound charge density = 2 0 /3d] 32

Exercise : The permittivity of a medium lling the space between the plates of a spherical capacitor with raddi a and b (b > a) is given by = 2 4
0 0

a r (a + b)/2 (a + b)/2 r b

Find the capacitance of the capacitor, distribution of surface bound charges and 1 1 1 1 , the total bound charges in the dielectric. [Ans. C = 8 0 a a + b 2b bound charges on dielectric surface with radii a, (a + b)/2 and b are respectively /2, 3a2 /(a + b)2 and 3a2 /4b2 ]

33