‘MSci 4261 ELECTROMAGNETISM
LECTURE NOTES IL
2.1 Electricity, Magnetism and Matter
Jn the previous notes, the equations eoveming electromagnetism in a vacuum, with source tems, were
discussed, These equations are just four in number, describing, the div and au of the dectric and magnetic
fidds. We will have more to say about these equations later in the course. At this point we tum to the
consideration of electromagnetic fidds in matter. This is in general a very complex situation, with matter
exhibiting a variety of dectric and magnetic behaviour, such as dickectricity, ferroclectricty, piezoelectricity,
paramagnetism, diamagnetism, ferromagnetism, superconductivity, etc. However, under conditions whee
the fields are not strong and with matter which is not in an unusual state, one can simplify the discussion
considerably.
2.2 Polarisation and the Electric Displacement Field
Matter which is not a perfect conductor is able to support an dectric fidd in its interior without flow
of current. The fidd however acts on the charged nucki and dectrons, causing polarisation. simple mola
of this involves atoms at fixed sites acquiring electric dipole moments. We can define the polarisation field
P(r) of the insulator to be such that the dipole moment d in any finite volume V is given by
a= [ pa = [ro-pyars [x @-as)
Vv Vv Is
(one can cheek the socond equality for each vector component using integration Ly pauts). Physically then
dik the chettic dipole moment per unit volume. If we now choose the volume V’ to enclose all matter, with
the boundary of V ouside all matter, then P vanishes on the bounding surface ~
Pls=0,
and thes:
a=- [ ry-Pya.
However, by definition the dipole moment is given by the integral d = ['rpp(x)dV where pp is the bound
charge, ie in the matter sample. Thus we conclude that,
V-P=-p.
‘Tus the macroscopic density of bound charge ps giving the polarisation is given by po =
may also be an additional ‘free’ charge density py. Maxwell's equation for E is then
1 1
V-E= Fm+ pp) = GOV P+)
Defining the new electric displacement field D by
D(r) = @E(r) + P(r)
wwe have the equation
V-D=py.
F the external field FE & not too large to break down the structure of the insulator, then typically it is
found that the polarisation field induced is proportional to the electric field, ie
P(r) = coxeB(r)
where the constant Xe & called the electric susceptibility. (We are assuming that the matter is uniform and
‘isotropic. Insulators of this type are called dielectrics. Thus, for didectrics,
D=coGE
where ¢, = 1+ Xe is called the relative permittivity or dielectric constant,
12.3 Magnetisation and the H Field
‘There is an analogous discussion to that above for magnetic fields. Matter placed in an external magnetic
fidd will react, for instance via diamagnetism, where oppesing atomic magnetic moments are incluoed by the
external magnetic field.
Here we will again assume that there can be defined a smooth averaged macroscopic field, here called
the magnetisation field M(x) such that the total magnetic manent m in a volume V & given by
ma [ Me) avy
so that m is the magnetic moment. per unit valume. First note the vector identity.
x (VX M)=2M+V(eeM) = (VM = (VM
(this is straightforwanily proved by considering each component separately). ‘Then
[Mavs (}excexm0-Ve-my +9 e-VM)av
f A
‘The second term an the right-hand side of the above equation gives zero - as each component gives a total
derivative, which integrates to a boundary integral, which is zero since we choose V so that there is no matter
at or outside the boundary and hence M = 0 there. A similar argument, applies when one combines the last.
two terms on the right-hand side of the above equation; again this is most easily seen for each component,
separately:
‘Thus
m=3 [ (rx(vxm)
‘However, by definition the magnetic moment is given by
m 3h (ex Jn) av
where J is the current. Thus we can identify the axrent distribution comesponding to the magnetisation as
Jp=VxM.
If there are additional free currents J, then Maxwell's equation for J x B becomes
Vx B= (Sy + Jo) = oy +V Tune,
Jam fw °
where $ is the suface bounding the volume V, with unit outward normal vector n = (m,r2,n3). We will
aclopt the Hinslein summation comention, whereby if an expression contairs an index repeated twice, it is
‘assumed that this index is summed over in the expression. An example # the index 0 in the above equation,
‘wher it is assumed that the expression is to be summed with } running from one to thiee.
‘Thus we finally come to the result
d
SP n+Pema= fh Tuned.
5‘The left-hand side is the a-component. of the rate of change of total mamentum, including that of the
ectromagnetic field Pj. The right-hand side is the e-companent of the flow par unit area of momentum,
across the surface S into the volume V, ie the force per unit area acress the surface, acting on the fields and
matter.
Finally, we ramark that angular momentum can also be treated in a similar way. The angular momentum
dersity in the dectromagnetic fidd is given by
Lem =x Pens
where Py is the momentum in the fie discussed above, The flux of angular momentum is given by
M=Txr,
where T is the energy-momentum tenscr given above (se Jackson, Problem 6.10, for more discussion of this
notation). Conservation of angular momentum is expressed by the equation
Gillin +Len)+V-M=0,
where Lm is the mechanical angular momentum, and this equation follows by use of Maxwell's equations,
Just as for the equations expressing conservation of energy and momentum which we have discussed earlier.
2.7 The Clausius-Massotti Relation
As an application of the ideas of polarisation etc to matter, and its lation with micrescopic quantities,
wwe now derive the Clausius-Mossotti relation, expressing malecular polarisability in terms of the dielectric
constant, We consiler the matter under consideration to be made from molecules, and under the influence
of an applied external macroscopic dectric fidd , the matter will be polarised, The polarisation of neigh-
bouring molecules will give rise to an internal fidd B; at each molecular site. This internal field will itself be
made from the difference of the actual contribution Enear of the molecuiks close to the given site, and the
averagp contribution from those molecules Bp due to the overall polarisation
EB) =Enar—Ep
(we will not use the near field here).
‘Now, it was noted earlier that the potential & for a charge distribution p is given by
le)
90 = Be | eet
where the volume V’ encloses the distribution, Taking V to be a sphere of radius R, it is possible to show
that (see, for example, Jackson, Section 4.1)
- od
Eav=-2,
a BoP
where p is the dipdle mement of the duanze distribution with respect to the center of the sphere. Now,
P= 4rF®P, if we take R small enough so that P is corstant inside the sphere. Thus the average electric
field inside the sphere is given by
5; =e /B=-a?.
° = TB ly Be
1
= goP + Eran
Bis gaP+
‘We can define the molecular polarisability mot by the relation < pmot >= €0?/mot(E + E;), where
< Pmoi > is the average dipale moment of the molecules. Notice that E+E; is the applied field at the
Whence
6molecule, We also have P = NPmat where NV is the average mmber of molecules per unit volume, Recall
also that € = eo(1+e)« Assuming that the near field vanishes, we thus dertuce that
Using P =y