Clausius Mossotti Article

Home Search Collections Journals About Contact us My IOPscience
The Clausius-Mossotti equation: an alternative derivation
This content has been downloaded from IOPscience. Please scroll down to see the full text.
1983 Eur. J. Phys. 4 141
(http://iopscience.iop.org/0143-0807/4/3/003)
View the table of contents for this issue, or go to the journal homepage for more
Download details:
IP Address: 134.99.128.41
This content was downloaded on 22/11/2013 at 18:50
Please note that terms and conditions apply.

Eur I. Phls. 4 (1983) 141-143. Printed In Northern Ireland
141
The Clausius-Mossotti equation: an

alternative derivation
J H Hannay
H H Wills Physics Laboratory, University of Bristol, Tyndall Avenue, Bristol BS8 lTL, UK
Received 14 February 1983, in final form 17 August 1983
Abstract The Clausius-Mossotti equation relates the RCume LkquationdeClausius-Mossottirelielacon-

dielectricconstant of a material to the polarisability of stante diklectrique dun milieu & lapolarisabilitC des
its atoms. It finds natural explanation in terms of the atomes qui le constituent. On peut 1interprCter naturel-
(often omitted) 6 function in the electric field ofan lement B partir du terme en fonction 6 qui apparait
ideal dipole. This avoids the subtleties of the rather dans le champ Clectrique dun dip6le idtal, (terme dail-
tricky conventional derivation. leurs souvent omis). On Cvite
ainsi les subtilit& et les
astuces de la dkmonstration habituelle.
The formula for the electric field of an ideal dipole theory of dielectrics. This equation which is taught
p at the origin is (in SI units) in most physics degree courses is usually explained
ina quitedifferentway involving animaginary
cavity in the dielectric whose shape must be taken
as spherical (for quite subtle reasons, often omit-
ted).Actually a precedentforthealternative
explanation has been pointed out to me. Although
the starting point in the form of equation (1) is not
made explicitin his paper, Darwin (1934) clearly
Often the last termis omitted because it is localised had the same idea some 50 years ago in a plasma
onthedipoleitself. Physically it representsthe physics contextandtheapproachhasapparently
infinitely strong backwards field directly between beenuseful in other physical contextsalso (A B
the positive and negative charges. Mathematically it Pippard,privatecommunication).Heretheargu-
arisesfromthesecondderivative of l / r ; if the ment will bepresented in its ownrightwiththe
dipole axis taken as the z axis then E, = ( 4 7 r ~ ~ ) Xp concise notationafforded by the S function.
a(l/r)/az2. Like a(l/r)/ax and a2(l/r)/ay2 this Another derivation which deserves mention, lying
derivativecontains a term f V z ( l / r ) = - 4 ~ / 3 S ( r ) . somewhere between the conventional one and the
It is worth emphasising that the coefficient - p / 3 ~ , present one is that in volume two of the Berkeley
is model independent: if the positive and negative physics course(Purcell1965). A thoroughtreat-
charges are smeared into balls for example to pro- ment of the conventional argument is given in the
duce a field locally different from equation (l),the book by Robinson (1973).
volume integral of the true field over this locality is The model of a dielectric medium in question is
still equal to the corresponding integral of equation an array which may be regular or irregular of ideal
(1) complete with the consequent-p/3so term. (The polarisable atoms in which dipole moments p can
volume integral of the field can be expressed asa be induced by applying an electric field. Each di-
surface integral using the identity JV+dV=J&i dS. pole sits in a local microscopic electric field which is
If thelatter is unchanged, so therefore is the the superposition of the fields,given by equation
former.) (l),of the other dipoles and the external field. In
The case I shallseektomakehere is that the the case of a simple (i.e. Bravais) lattice array this
often neglected 6 function term in the electric field local field will bethesameateachdipoleand
of a dipole can provide a simple and natural expla- the induced moment p is proportional to it through
nation of the Clausius-Mossotti equation in the the atomic polarisability constant a. Otherwise the
0143-08W!83!030141+03$2.25 0 The Institute of Physics & the European Physical Society

142 Hannay
J H
localfield will vary from site to site but in the (YE

model the induced moments are required to be p = 1 -(na/3E,,)
equal, having a common p given by a times the
mean of the local electric fields on the sites. (One so from equation (2) we obtain
can imagine this constraint being imposed by
mechanical connections between the dipoles.) nolie,,
-l=
The object is to derive an expression for the 1- ( n a i 3 e J
macroscopically observable dielectric constant E of
the medium in terms of the microscopic atomic the Clausius-Mossotti (or Lorentz-Lorenz) rela-
polarisability cr. The starting point is the definition tion.
of F To complete the description thestandardre-
np = ( F - l)coE ( 2 ) marks on the validity of the equation (e.g. Ashcroft
and Mermin 1976) translate into the present tan-
where n is the number of atoms per unit volume. guage as follows. Fora regular array of dipoles
The electric field E is the field in the medium, that appropriateto a crystalline solid dielectric it is
is, the spatial average of the microscopic electric quite possible that, by virtue of the regularity, the
field. There is no need to consider how this field field inwhich each dipole sits isnot equal to the
decomposesintointernal and externalcontribu- de-spiked spatial average field ( E + np/3s,,), so that
tions, only their resultant is involved. the C-M equation is wrong. For a simple lattice r;
If n is sufficiently small, forexample if the of dipoles with one at the origin the requirement
dielectric is a dilute gas, then there is no difficulty; for equality is
one merely substitutes p = crE in equation (2). This
is not correct for dense dielectrics however for the
rather surprising reason that the electric field that
the dipoles see (on average) is larger than E. This is
the phenomenon whose explanation we seek, for it
leads to the Clausius-Mossotti relationship between =iLV,V,- 1 (6)
F and a. Now the electric field varies markedly on a W,, 47Wl lql
microscopic scale from place to place due to the
fields given by equation (1) from the dipoles them- where V is the volume of the unit cell and the
selves complete with S function spikes onthe primes indicate that the integral o n the left ex-
dipole points, and E is by definition its spatial aver- cludes the S spike at the origin, and the sum on the
age. So the assertion is that the dipoles somehow right excludes the term with r, = 0. Each side of this
contrive to find themselves in spots of higher than equation is only conditionally convergent (which is
average field. This claim becomes especially where all the subtlety of the conventional deriva-
paradoxical in the case of an irregular array of tion arises because correct conditions at infinity
dipoles whose positions arerandom, for then an must beconsidered), but their difference is abso-
extra test dipole inserted at a random position must lutely convergent. It is not zero for general lattice
at first sight feel, o n average, exactly the spatial symmetry, but is so for cubic symmetry since then
average field E. the tensor which p , multiplies is isotropic but easily
The resolution of this paradoxand the single seen to be traceless andthereforezero. In this
central proposal of the Clausius-Mossotti relation cubic case therefore the C-M equation is valid.
is that the average field experienced by the dipoles For an irregular array the question is whether the
is after all still just a spatial average electric field, mean of the fields in which the dipoles sit is equal
but it is the spatial average excluding the 6 spikes to the de-spiked spatial mean ( E + np/3&,,).If so the
on the dipoles themselves. In theexample,the C-M equation is an exact consequence of the
position of the extra dipole was random except that model,otherwisenot. The criterion is expressed
(implicitly) it was not inserted exactly on top of algebraically by astraightforward modification of
anotherone. Dipoles do not feel their own or equation (6) and it is satisfied, by the same argu-
each others S spike electric fields, but apart from ment as was used for the regular cubic array if the
that they are indeed subject to the typical field in irregular array is statistically isotropic as in a liquid
the medium. That is the assumption of Clausius- for example.
Mossotti and the consequent equation follows im- Finally it should be mentioned that to go beyond
mediately. The spatialaverage field duetothe 6 the mean-field model that yields the C-M equation
spikes is -np/3e,, so the mean field felt by a dipole is, not unexpectedly, quite a subtle matter and an
is E +np/3eO and the polarisation p is therefore extensive literature beginning with Onsagers reac-
given by tion theory (Onsager 1936) has developed on this
P = a ( E + npi3ao)
(3) subject. While the modified viewpoint adopted
above might streamline such elaborations it would
or not of course be expected to yield anything new.
The Clausius-Mossotti equation : an alternative derivation 143
References
Ashcroft N W and Mermin N D 1976 Solid State Purcell E M 1965 Electricity and Magnetism, Berkeley
Physics (New York: Holt-Saunders) Physics Course, v01 2 (New York: McGraw-Hill)
Darwin C G 1934 R o c . Roy. Soc. A146 17 Robinson F N H 1973 Macroscopic Electromagnetism
Onsager L 1936 J. Am. Chem.Soc. 58 1486 (Oxford: Pergamon)

Clausius Mossotti Article

Uploaded by

Copyright:

Available Formats

Clausius Mossotti Article

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Clausius Mossotti Article

Uploaded by

Copyright:

Available Formats

Home Search Collections Journals About Contact us My IOPscience

The Clausius-Mossotti equation: an alternative derivation

1983 Eur. J. Phys. 4 141

Please note that terms and conditions apply.

The Clausius-Mossotti equation: an

Received 14 February 1983, in final form 17 August 1983

Abstract The Clausius-Mossotti equation relates the RCume LkquationdeClausius-Mossottirelielacon-

0143-08W!83!030141+03$2.25 0 The Institute of Physics & the European Physical Society

localfield will vary from site to site but in the (YE

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.