Fundamental Distributions in Statistical Mechanics
Author(s): Frank C. Andrews and A. C. Andrews
Source: Transactions of the Kansas Academy of Science (1903-) , Winter, 1963, Vol. 66,
No. 4 (Winter, 1963), pp. 600-605
Published by: Kansas Academy of Science
Stable URL: https://www.jstor.org/stable/3626810
REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/3626810?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Kansas Academy of Science is collaborating with JSTOR to digitize, preserve and extend access
to Transactions of the Kansas Academy of Science (1903-)
This content downloaded from
132.77.58.62 on Fri, 31 Mar 2023 14:34:31 UTC
All use subject to https://about.jstor.org/terms
Fundamental Distributions in Statistical Mechanics
FRANK C. ANDREWS and A. C. ANDREWS
Introduction
It is a matter of taste which formulation of Fermi-Dirac, Bo
Einstein, and grand canonical distribution functions one prefers. Gi
simply postulates the form of the grand canonical distribution. In
book,2 Tolman considers this postulate justified by the small fluctuat
predicted and by the agreement of calculated thermodynamic quant
with their known values. He then uses the grand canonical distribut
to obtain the Fermi-Dirac and Bose-Einstein distributions for non-inter-
acting particles. Fowler and Guggenheim3 rationalize the grand canonical
distribution by a steepest descent integration, but actually justify the result
in the same way as Tolman. They obtain the Bose-Einstein and Fermi-
Dirac distribution by the steepest descent technique of Darwin and
Fowler.4 Schridinger5 employs both steepest descents and the com-
monly used combinatorial method of Boltzmann. Landau and Lifshitz
start from an assumption about the probabilistic interpretation of the
theromdynamic entropy. An expansion of the entropy yields an approx-
imation giving the grand canonical distribution. From this they easily
obtain the Fermi-Dirac and Bose-Einstein distributions.
In 1940, Tolman7 considered that the grand canonical distribution
had not yet been formulated on a sound and understandable basis. Use
of extremum methods concerned him for a number of reasons. He sought
a formulation which showed clearly just what physical conditions were
necessary for a grand canonical ensemble to represent satisfactorily a
system open to exchange of matter with its surroundings. He stated that
a grand canonical distribution must be used in order to obtain exact
Fermi-Dirac and Bose-Einstein results, but his formulation of the grand
canonical distribution was complicated enough that his approach and this
general problem have since been more of less neglected.
Canonical Ensemble
In this paper, Tolman's problem is reconsidered, starting from the
canonical distribution for the probability, PN(i), of the ith quantum
state of an N-particle system:
e-PEN(i)
PN(i) = a e-EN(i) - (1)
QN
Transactions of the Kansas Academy of Science, Vol. 66, No. 4, 1963.
This issue was published February 25, 1964.
[600]
This content downloaded from
132.77.58.62 on Fri, 31 Mar 2023 14:34:31 UTC
All use subject to https://about.jstor.org/terms
Fundamental Distributions in Statistical Mechanics 601
This probability distribution has been recently discussed by the authors.8
Here, QN is the partition function or reciprocal of the normalizing con
stant, a, for the N-particle system,
QN -- e- EN(i) , (2)
and f may be shown to be 1/kT. Use of Eq. (1), regardless of how it
is formulated, to represent a closed system in equilibrium with a heat
bath reflects two necessary and sufficient conditions: 1) The only prop-
erty of the ith quantum state on which its probability may depend is the
energy of the state. 2) If the energy of the N-particle state may be
written as a sum of small contributions of various physical origins,
ENO(i) =- E , (3)
then P, factors into a product of pr
tions. That is, nothing explicitly cou
contributions, though several may b
pendence of their E's on the same
feature imaginable which might coup
butions explicitly is a restriction on
role of the heat bath is to serve as a s
ating this restriction from considerat
If the system is an ideal gas of ind
are N contributions to the sum, Eq.
written.
ENx - v n, ev, (4)
where n, is the number of particles in the vt
a particle ("occupation number of particle sta
of a particle in state v. Use of Eq. (4) inst
that one cannot label the particles or distingu
quantum state.
Fermi-Dirac Statistics
For a system composed of fermions, the occupation numbers must be
either zero or unity. The ensemble average of the occupation number for
particle state v, in, is the same as the probability that particle state v be
occupied:
e-PEN( i)
n-y - Q PN
%, (i) = : . (5)
This content downloaded from
132.77.58.62 on Fri, 31 Mar 2023 14:34:31 UTC
All use subject to https://about.jstor.org/terms
602 Transactions Kansas Academy of Science
The sum is over all N-particle quantum states which have particle state
v occupied. If the notation QN' and QNO be used for the sums over all
N-particle states having particle state v occupied or unoccupied respec-
tively, then Eq. (5) becomes
QN' QN'
n, QN
-- QNo + QN'
QNO
QN'
n-= .(6)
1N-10
The ratio in the denominator of Eq. (6) depends on choice of par
state v only through the restriction o . For large numbers of fermions
at absolute zero or for large or small numbers at moderate temperat
this dependence is negligible. Then nii is indeed [a exp(pEv)
for a Fermi-Dirac gas.
Bose-Einstein Statistics
For an ideal gas of bosons, the ensemble average of the occupation
number of particle state v is the ensemble average of the number of
particles in that state:
n-=QN'+
n
QN" QN"
1.- 2 +3 -- ***
QN QN QN
QN' + 2 QN" + 3 QNt"' +- + -
QN + QN' + QN" + QN" + "'"
Q x"
QNf +1 + 2 QN2if++ QNQ'
3 QN3Q +2Q
This content downloaded from
132.77.58.62 on Fri, 31 Mar 2023 14:34:31 UTC
All use subject to https://about.jstor.org/terms
Fundamental Distributions in Statistical Mechanics 603
The reader may quickly verify that the last form of -in is an identity.
In the numerator and the second series of the denominator, exp (- p ev)
may be cancelled in each term to yield
Q N' Q '[-1 Q--" ' QN-i"
QN-1Q QI? Q
Q0o 01 Q_0o QN-O
Q-o 1+ 2- +- 3 - -- -... QN' 1 + -2 -t-3 -}-+"
Except for the difference between QN and Q-1,, the series in the paren-
theses are identical and may be cancelled. This difference might be
noticeable in some of the last terms, but when N is large the effect on
the sums is negligible:
QN'
QNO - QN'
n= --(7)
IQN-10
This is completely analogous to Eq. (6). Only when the temperature is
so low that all the particles are crowded into a handful of lowest energy
levels could the ratio in the denominator depend on choice of state 'v.
Thus n, is given by [a exp(pflv) - 1]-1 for a Bose-Einstein gas.
Grand Canonical Ensemble
In formulating the grand canonical ensemble, the total number of
particles contained in the system is allowed to vary. The system may be
thought of as a well defined portion of a larger composite consisting of
system plus particle bath. The composite contains N particles and is in
equilibrium with a heat bath, so it may be represented by a canonical
distribution:
e-PEN
P=
QN (8)
This content downloaded from
132.77.58.62 on Fri, 31 Mar 2023 14:34:31 UTC
All use subject to https://about.jstor.org/terms
604 Transactions Kansas Academy of Science
The distinction between system and particle bath is clear enough that it is
meaningful to ask how many particles are in the system. Furthermore,
the total energy of the composite must be expressible as the energy, E1, of
the n particles in the system in some state or other plus the energy, E, of
the(N - n) particles in the bath:
E - E + e. (9)
The probability, P(n, Ei), that the system contains n particles in a
given n-particle quantum state which has energy E, is simply proportional
to the sum of terms of the form of Eq. (8) over all states of the com-
posite which show n particles in the system in their appropriate state.
It makes no difference in what state the (N - n) particles of the bath
are:
e-3(E+?n) QN-n
P (n, En) oc N ,- )- e-E3 . (10)
QN QV
The summation in Eq. (10) is over
of the bath. Equation (10) may be
Pe -1 0 N-2 .Q-3 S
P (n, E~) ac e-n"? . . . Q
Qv Qv-1 Q_-2 QN-
The n ratios multiplying the expo
negligible amount if the system is v
bath. The result is
P (n, En) )c e-PEn A,
P (n, en) - A e p n-En, (11)
where A has been written as exp(p#3). This is the usu
distribution.
ACKNOWLEDGEMENTS: The University of Wis
of this work was supported by National Science Found
G20725. The Kansas State University portion was
tional Institutes of Health grant E(1354) (C8).
This content downloaded from
132.77.58.62 on Fri, 31 Mar 2023 14:34:31 UTC
All use subject to https://about.jstor.org/terms
Fundamental Distributions in Statistical Mechanics 605
References
J. W. GIBBs, Statistical Mechanics (Yale University Press, New Haven, 1902),
p. 190.
2 R. C. TOLMAN, The Principles of Statistical Mechanics (Oxford University Press,
London, 1938), Sec. 114.
" R. FOWLER and E. A. GUGGENHEIM, Statistical Thermodynamics (Cambridge
University Press, Cambridge, 1939), Chap. VI.
4 DARWIN and FOWLER, Proc. Camb. Phil. Soc. 21, 391 (1922).
S E. SCHR6DINGER, Statistical Thermodynamics (Cambridge University Press, Cam-
bridge, 2d ed., 1952).
o L. D. LANDAU and E. M. LIFSHITZ, Statistical Physics (Pergamon Press, London,
1958), Sec. 35.
7 R. C. TOLMAN, Phys. Rev. 57, 1160 (1940).
S F. C. ANDREWS and A. C. ANDREWS, Trans. Kansas Acad. Sc. 65, 247 (1962).
-Chemistry Department, University of Wisconsin, Madison;
Chemistry Department, Kansas State University, Manhattan
This content downloaded from
132.77.58.62 on Fri, 31 Mar 2023 14:34:31 UTC
All use subject to https://about.jstor.org/terms
