UNIT 14 THE PARTITION FUNCTION
Structure
14.1 Introduction
Objectives
14.2 The Partition Function of an
ldcal Monatomic Gas
Thermodynamic Functions
Gibbs Paradox
14.3 The Sackur-Tetrode Formula
14.4 Diatomic Gases
Rotational and Vibrational Partition Funcions
Heat Capacity of Hydrogen
14.5 Summary
14.6 Terminal Qucstions
14.7 Solutions and Answers
14.1 INTRODUCTION
In the prcvious unit, we obtained the distribution function for a clássical Maxwell
Bolzmann idcal monatomic gas. Then we introduced the concept of partition function and
lcarnt lo express various thermodynamic functions in terms of the partition function. It is,
therefore, important to know to compute the parition function for athermodynamic
system of interest. In Sec. 14.2, you will learn to compute the partition function of an ideal
monatomic gas. Thc expressions for thermodynamic functions willalso be derived. Gibbs
showed that the expression for entropy so derived suffers from a serious flaw; cntropy
fails to cxhibit its extensive character. You will learn to resolve this so-called Gibbs
paradox in Scc. 14.3. This leads us to the Sackur-Tetrode equation. From Unit 9 of Block
3, you would recall that unlike a monatomic gas, diatomic gases have rotational and
vibrational degrees of freedom. These contribute to their total energy and hence heat
capacitics. You will learn to evaluate rotational and vibrational partition functions with
particularrcference to hydrogen.
Objectives
A•cr studying this unit, you should be able to :
and work out all the
compute the partition function of an ideal monatomic gas
thermodynamic functions
pointout the flaw in the expression for entropy
Gibbs paradox
derive the Sackur-Tetrode formula and resolve the
capacities of diatomic gases
calculate rotational and vibrational contributions to heat
capacity of hydrogen.
explain the observcd temperature variation of heat
FUNCTION OF AN IDEAL
14.2 THE PARTITION
MONATOMIC GAS
.
each of mass mand occupying
Consider an ideal monatomic gas consisting of N particles, translational. That is, the
energy of the system is wholly
a volume V. This mcans that the forccs are absent.
potential energy is zero sincc intermolccular
The encrgy ol a particle in thc ith coll is givcn by
and the single particle partition unction is
2m (14.1)
If encrgy is a continuous variable, we can
rcwritc it as
(14.2)
Intgration over the space variables gives Vso that
(14.3)
Yu wi note that all thrcc intcgrals are
them l us idcntical and it is sulficicnt to evaluatc any onc of
considcr
By putting p, =- in the first
dp:. (14.4)
and we can write integral, you can casily show that both integrals arc alike
(14.5)
If wc now introduce a change of variable by writing
=r
2m
we find that
P. dp, = dr
or
dp, = m1/2 dr
and thc intcgral
(14.6)
is a standard gamma function of ordcr (1/2). Its
valuc is So that
1/2
0 (14.7)
You willobtain he same exprcssion for , and I, .
Hence, Eq. (14.2) gives
3/2
(2 mr (14.8)
B
This is the partition funcçon in -space for individual particles. Since we are discussing a
monatomic gas consisting of Nparticles, the N-particle partition function
where E, =t +E; +...t Ey. Thus
Within thc framework of classical statisics, we assume that these particlcs are
distinguishable and indepcndent. So we can rewrite it in a compact form as
where Zis given by Eq. (14.1). Hence
3N/2
2mr (14.9)
h3N B
Having got an expression for the partition funcion for a N-particle ideal gas, we can now
calculate the various thermodynamic functions.
4.2.1 Thermodynamic Functions
You will recallthat all thcrmodynamic functions are rclatcd to the partition funcion via
Eqs. (13.36)- (13.39). So to procccd with calculation of thesc functions, we first take
natural log on both sides of Eq. (14.9). This gives
In Zy = Nin V+ 2 NJn 2Tm 3N
-In ß (14.10)
h 2
From Eq. (13.36), we recall that U = In ZN Hence, internal energy of a
N-particle gaseous system is given by
3N 3
U=
2ßB Nkg
2 T (14.11)
The hcat capacity at constant volumc is
(14.12)
You arc familiar with this result from Unit 9, Block 3 of this course.
We can calculate the average pressurc cxertod by the gas using the relation
From Eq. (14.10), we have
p=
-#)
Nky T 2U
(14.13)
which is the familiar idcal gas cquation,You willnote that ncither thermodynamics nor
kinctictheoryof gascs cnables us to establish its Cxact form. It means that natural
anls explanationof molccular chaos lics in statistical argumcnts, which arc morc profound.
The Hclmholv Irce cncrgy cxprcssed in terms of Z is
F= - NhpTIn Z
(14.14)
From the previous unit, you would rocallthat cntropy and partition function are conncctcd
by the relation
On substituing for In Zand Ufrom Eqs. (14.8) and (14.11), we get
S- Nk¡lln V- 3Inh+n(2nnky)+in T]+
(14.15)
This is hc classical cxpression for thc cntropy of an ideal
considcr it in dctail a liulc later. To get a littlc practice, youmonatomic
gas. We will
should solve thc following
SAQ.
SAQ 1
Obtain thc cntropy and pressure from Hclmholtz cncrgy
The degcncracy paramctcr A is given by
A= Z (2mkg T)y2
nh3 (:4.16,
You will note that at high tempcratures and/or low densities, A will be large. We can
relate it to the deBroglie wavelength aB and intcr-paricle distance by noting that
with P -= kaT. Hence,
2m
h
V2mkgI (14.17)
Also, the mean inter-particle distance ro is given by
1/3
Hence,
A=
-()
When A>>1, we have aB << Io. That is, the thermal wavelcngth is very
(14.18)
-compared to the inter-particle distance and classical rcsults hold goodsmall
at high
temperatures and/or low densities.
The chemical potential u is given by
= kgTinA =kgTln (2Tmk,Tr2 (14.19)
nh3
Let us nowre-examine the expression for
entropy closcly.
14.2.2 Gibbs Paradox
From Eq. (14.15), we note thatcntropy of an The Parition Punction
the functional depcndenceon volume and idcal gas depends on V, T and N. Morcover,
tempçrature is the same
thermodynmic considerations. But we note that asT ’ 0K,S’-, This is from as obtaincd
physically meaningful and not
S’0as T’ ).However,contradicts
you
the third law of
thermodynamics (which states that
have expected dhis result because should not be unduely concerned. In fact, you should
classical
temperatures. The explanation has genesis statistics
in quantum
is a good description only at high
detail in the next unit. mechanics. You will lcarn this in we You wouldrecall that in
characterize variableshermodynamics
as intensive and
A
more serious objcction against Eq. (14.15) is its extensive. Intensive variables do not
bchave as an cxtensive quantity. Let us incrcasc implication that entropy does not depend on the number of particdes or the
both
contains a term N In V. Hence, S will not incrcasc in theV and N by a factor . Eq. (14.15) size of thc sysem. On the oher hand,
clearly understood by considering same proportion. This can be extensive paramcters depend on the size of
intermixing of two idcal gases. the sysem, ie get multiplied by the factor
the size of the system undeer
Thecntropy of a system of Nparticles consideration
occupying a volume Vis given by increases.
S= Nkalln V- 3In h+In (2mka) + In 2 Nks
2 T]+21
= Nk [In V+ InT +o)
(14.20)
with
G = 2 In (2r mkg) - In h? +
Lct us now consider two idcal gases
contained in two chambers of volumes V, and V, and
scparatcdby a rigid partition as shown in Fig. 14.1. Suppose that these gases are in
cquilibrium at Temperature T. Then the entropy of each gas is given by
S, = N, kp [In V, +In T+ o)
2
nd
iS2 = Ng ky[In V, +In
2 T.+ o]
Fig. 14.1: a) The system divided by a partition into 1 and 2.
b)Asystem of N-particdes in a volume V.
So, the total initial entropy of these gases is
S, = S,+ S, = NkglIn V,+In 7T+Gi]
+Ngka [In V, +In T+G,l
Now we remove the partition and these gases mix by diffusing into one another. We can
regard it as frec expansion of cach gas to volume V= V+ V. Then final entropy of the
system is
2 T+o]
S, = N,k, [In V+In
+N,ka[In V+inT+ ol
2
would fccall that difTusion is an irrcvorcil.1.
From thc Unit S, Block 2 of this coursc, you
incrcasc or dcrease? It should increasc. To
nics
proccss. Do youcxpcct the cntropy to -S;.
differencc AS = S,
discover that it is actually so, lct us computc thc
S, - s, AS - (N, + N,)An inVn7
(14.21)
Obviously, ASis grcatcr than zero. For the spccial case,
N, = N= andV = V, =
we tind that
AS = NA, ln 2 (14.21a)
This is in comformity with thermodynamic resulis.
Le us now consider that the same gas is put in the two chambers shown in Fig. 14.1. We
expect that removal or subscquent insertion of partition is acompletcly reversible
process.It should not influence the macroscopic behaviour of the gas and distribution of
particlcsover accessible microstates. Thercforc, the cntropy of mixing should be zero.
But this contradicts Eq. (14.21) since its derivation does not depcnd on the identity of the
gases. That is, even for self-mixing, Eq. (14.21) gives the same increase in entropy which
iscertainly not tenable. It implics that Sdepends on the history of the system and cannot
be a function of the thermodynamic state only. That is, we can manage to change entropy
by factors extraneous to the system. This is known as the Gibbs paradox.
It we carefully re-cxamine our analysis, we can idcntify the root causc of this rouble in
that wc have treated all N particles as distinguishable. This amounts to tacitly assuming
that interchange of two particles leads to a physically distinct state of the gas. This, of
course, is not correct and Eq. (14.21) over-estimates the number of accessible states. You
would recall that the number of permutations of Nparticles among themselves is N!. For
identical particles, these permutations lead to the same physical situation. On accounting
for these, we get acorreCt expression for entropy. We will consider it now.
14.3 THE SACKUR-TETRODE FORMULA
The corrcct way to treat a system of indistinguishable particles is to use quantum
slatistics.You will lcarn this in the next unit. However, for the prescnt, we use an ad-hoc
procedure suggestcd by Gibbs. This is known as correct Boltzmann counting.
We begin by looking at Eg. (13.20) for the thermodynamic probability W. Treating the
paricles as indistinguishable amounts to dividing Wby N!. Hcnce,
1
W =
Nn,!
Inthe partiion function for an idcal gas made up of Nindistinguishable particles, it is
reflcctcd in the form
Z, = (ZN) correct = ZN
N!
3N/2
yN(2tm ka I
N! (14.22)
SAQ2
Verily Eq. (14.22).
Taking logarithm on both sidcs, we get
The Purtitlon Function
In N!
Fr large N, wc can use Sirling's approximation (In N! = N In N
N) to obtain
In Z§ =
Since U and p depcnd on the dcrivativc of Z, the
expressions unchanged. Howevcr, for cntropy, thepresence of the factor N! lcaves thcir
expression is modificd to
(14.23)
with
3/2
(14.23a)
Hencc,
S= Nkg
lan (14.24)
whcre Bis dcBroglie wavclength associated with the
This equation is known as gaseous particles at temperature T.
Sackur-Tetrode equation. It has been established
expcrimcntally that this is the corrcct cxpression for the entropy of an ideal monoatomic
gas at high tempcratures. The constant h, which can be
arbitrarily small in classical
statistics must,of course, be put cqual to Planck's constant.
However, you must now
convincc yoursclf by solving the following SAQthat Sackur-Tetrode
Gibbs paradox. equation is free from
SAQ3
Show that Eq. (14.24) is frce fromn Gibbs paradox.
The Helmholtz energy is now given by
F=U-7S =Nk¡T -Nk¡i
T nn
_52
(14.25)
14.4 DIATOMIC GASES
Considcr a diatomic molccule like HCI. It may be trcated as a two
hcld by inter-atomic forces along the line joining the particles. Let particle (atom) system
m, and m, (assumed to be point-like) and separated by a the masses of atoms be
distance r. This is known as the
dumb-bcllmodel of a molccule and is depicted in Fig. 14.2.
------m,
Fig. 14.2 Dumb-bell model of a diatomic molecule
Lcius choosc x-aa is along thc line joining the masses. The
momcnts of incrtia about the
3
classmate
Date
Pege
rade chang
he
asmblíes Qnd thess
wh
div(s-v) 3dimRas te foM
quanh
ahen is Subaihte fot
alssata
Data
Paga
Comonad
eneA the
Can be
yutm Consideing
m m
the
hatial
m
-V
mowentm
edassnute
Dato,
Pane.
Siuce
No
No
no Lioiile nan Comditin
