CHAPTER SIX

SPECIFICATION OF THE STATE OF A SYSTEM

INTRODUCTION
You should have noticed that a key element in our discussion in the last chapter was the number of different
microstates which corresponds to a given macrostate. We can make progress in our study of statistical
thermodynamics, then, only after we carefully consider the following:

1. We must establish a detailed method of describing the states of a system. This means that
we must have a detailed way of describing the precise microstates in which a system can be
found (or, equivalently, the precise outcome of a single experiment such as the toss of a
die), as well as a precise way to describe the various macrostates of a system.
2. We must describe the system in terms of an ensemble of similar systems (that is, a large
number of similarly performed tosses of the die, one right after another, or an infinitely large
number of equivalent tosses taking place at one instant). This would seem to imply that we
do not concern ourselves with an exact description of the system (such as the exact way in
which a coin is tossed or a single die is thrown
from which we could, in principle,
predict the outcome). Rather, we concern ourselves only with a statistical description of
what we might expect when each experiment is carried out under similar circumstances.
3. We assume that all of the possible microstates are equally likely. There is no real basis for
this assumption. It is an a priori assumption, i.e., one which is assumed in advance and
which is later shown to be true or false by examining predictions based upon this
assumption. Thus far, this assumption seems to work very well!
4. We must use the mathematics of probability and statistics to determine the relative
likelihood of the outcome of the experiments under study. Those systems which have a
large number of microstates corresponding to a given macrostate will be most probable

those with a small number of microstates corresponding to a given macrostate will be least
probable.

The first three of these considerations will be examined in the present chapter. In the following chapter, we will
deal with some of the fundamental concepts of probability and probability calculations. We begin by considering
the way in which we specify the microstates and the macrostates of a particular system.

SPECIFICATION OF THE STATE OF A SYSTEM

For any system which we might consider, no matter how complicated, the motion of individual particles of
the system can be described exactly using the quantum mechanical wave function <(;1 , ;2 , á ,;0 ), where 0 is the
number of degrees of freedom for the molecule under consideration. Thus a system of R vibrating-rotators would
have R ‚ 0 degrees of freedom and thus R ‚ 0 different quantum numbers, since there is a quantum number
associated with each degree of freedom.

Example 1: Consider a system consisting of a single particle, considered fixed, but having a spin
of "# . The projection of the spin of this particle along a fixed axis is restricted to the two possible
projections designated by m œ + "# and m œ
"# ; or “up" and “down". Thus for this simple
system there are only two possible microstates.

Example 2: If we have R particles like the one described in the example above, each would have
2 degrees of freedom. Thus, there would be 2R degrees of freedom in this system, and the state of
the system can be precisely designated by giving the value of 78 for each individual particle 8.
The total magnetic moment of the entire system, however, would correspond to a macrostate of the
system. It should be clear that there may be a large number of microstates which would yield the
same macrostate.
Chapter Six: Specification of the State of a System 2

Example 3: Consider a system consisting of a one-dimensional simple harmonic oscillator whose

position coordinate is given by B. The possible quantum states of this oscillator can be specified
by a quantum number 8 such that the energy of the oscillator can be expressed as
"
E8 œ Ðn  Ñh = (6.1)
#
where = œ Èk/m is the classical angular frequency of oscillation. The quantum number 8 can
assume any integral value 8 œ 0, 1, 2, á .

Although atoms and molecules are properly described in terms of quantum mechanics, a description in
terms of classical mechanics may sometimes be a useful approximation. We would like, therefore, to describe the
specification of the state of a system in terms of classical mechanics. This will give rise to what is called classical
statistical mechanics.
We will start with the simplest case, a single particle in one dimension. We will designate the coordinate of
the particle by ; and the momentum of the particle by : [ œ m;† ]. We will represent this situation geometrically
by drawing Cartesian axes labeled : and ; . Specification of : and ; is then equivalent to specifying a point in this
two-dimensional space, commonly called “phase space" (see the diagram below). As the particle moves along the
; coordinate axis, a path in this “phase space" will be traced out. [Try to determine what the phase-space diagram
of a particle moving at constant speed would look like; what about the phase-space diagram of a simple harmonic
oscillator, a damped harmonic oscillator?]

Fig. 2.1. The classical phase space for a single particle in one dimension.

In order to describe this situation in terms of countable states of the system, it is convenient to subdivide the
ranges of the variables ; and : into arbitrarily small discrete intervals. Phase space is then subdivided into small
cells of equal size and of two-dimensional “volume" (i.e., area)

$ ; $ : œ 2o (6.2)

as shown in Fig. 2.2. Here 2o is some small constant having the dimension of angular momentum. The state of
the system can then be specified by stating that its coordinate lies in some interval between ; and ;  $ ; and that
its momentum lies in some interval between : and :  $ :, i.e., by stating that the representative point (;ß :) lies in
a particular cell of phase space.
Chapter Six: Specification of the State of a System 3

δp

δq
Fig. 2.2. The phase-space of Fig. 2.1 is here shown subdivided into equal cells of
area $ : $ ; œ 2! Þ

The specification of the state of the system clearly becomes more precise as one decreases the size chosen
for the cells into which phase space has been divided, i.e., as one decreases the size of 2o . Classically, 2o can be
chosen as small as we like which means that the number of cells would become larger as 2o became smaller
(eventually approaching infinity). However, we know that there is a quantum mechanical limitation to the size of
2o , because of the uncertainty principle, which states that
h
$: $;   (6.3)
#
(where the equality holds only for a Gaussian wave packet). This means that subdivision of classical phase space
into cells of volume 2o less than ¶ h is really physically meaningless. Therefore, when we write
$ : $ ; œ 2! (6.4)
and understand that 2o must be greater than the order of h we are essentially using a semi-classical approach to
the problem. [It is important to realize that although the minimum allowable volume 2o is fixed by this semi-
classical picture, the nature of the uncertainty principle is still evident. Variations in ; may be large while
variations in : may be small, but the product must still be a constant.]
The extension of these remarks to systems of many particles is straightforward. If we have a system with N
particles, and each can move freely in any of the 3-dimensions, then we have f œ 3N degrees of freedom so that
the set of numbers [;1 , . . . , ;0 ; :" , . . . , ;0 ] can again be regarded as a “point" in a “phase space" of 2f
dimensions, and we again divide the phase space into little cells (of volume $ ;1 . . . $ ;0 $ :1 . . . $ :0 œ 2o0 ,
provided one chooses $ :5 $ ;5 œ 2o ).
Thus, the microstate of a system of particles is simply specified in the following way:

1. Quantum Mechanically: Enumerate in some convenient order, and label with some index
< (where < œ 1, 2, 3, á ), all the possible quantum states of the system. The state of the
system is then described by specifying the particular state < in which the system is found.
2. Classically: After the phase space has been divided into suitable cells of equal size, one
can enumerate these cells in some convenient order and label them with some index <
(where < œ 1, 2, 3, á ). The state of the system is then described by specifying the
particular cell < in which the representative point of the system is located.

The quantum mechanical and classical descriptions are thus very similar, a cell in phase space in the classical
discussion being analogous to a quantum state in the quantum-mechanical discussion.

Example 1: Consider a single particle moving in the B-direction, confined to be located

somewhere between 0 and Bo , and to have a momentum between 0 and :o . A two-dimensional
phase space for this situation is similar to that shown in Fig. 1.1, but restricted to the first quadrant.
The number of available states for this system can be determined by taking the accessible volume
(area in this case) of phase space and dividing by the volume of a single cell, 2o . Thus the number
Chapter Six: Specification of the State of a System 4

a of states available in this case would be

Bo :o
a œ (6.5)
2o
Likewise, for a single particle in three-dimensions, we would have
V< Vp
a œ (6.6)
h$o
where the volume Vr and Vp are the volumes in position space and in momentum space,
respectively, and where $ x$ px œ ho and $ x$ px $ y$ py $ z$ pz œ h$o . In a differential volume of phase
space (denoted by d $ x d $ p) the differential number of states is given by
d $x d $p
da œ (6.7)
h$o
Thus, to find the total number of states, we must add up (or integrate) all the states accessible to
the particle. If the particle is confined within a volume V and we are not concerned about a
particular direction of the momentum but only the magnitude (i.e., if we can write dpx dpy
dpz œ %1p# dp). Then we can write

a œ (  $  p# dp
%1 V
(6.8)
2o

or, in terms of energy

a œ(   % d%
#1V(#m)$/# "/#
(6.9)
2o$

where we assume % œ p# /#m. In general, where we cannot assume this last form for the energy, the
number of states accessible to the system is at least of the form

a œ ( =(%) d% (6.10)

where =(%) is called the “density of states".

Example 2: Consider a one-dimensional harmonic oscillator of mass 7 and spring constant 5 from
the viewpoint of classical mechanics. Denote the displacement of the oscillator by B and the linear
momentum by :. Phase space is then two dimensional. The energy I of the oscillator is given by
p# "
Eœ  5 x# (6.11)
#m #
For constant energy, this equation describes an ellipse in phase space. Now if we specify that the
energy is known to be in the range between I and I  $ I , we have the situation shown in Fig.
2.3.
Chapter Six: Specification of the State of a System 5

δp

E+ δ E
δx 0 δxA

Fig. 2.3. Classical phase space for a one-dimensional harmonic oscillator with energy
between E and E  $ E. The accessible region of phase space consists of the area between the
two ellipses.

The number of cells (or states) which are available to the system are the number of cells contained
between the two ellipses corresponding to the energies I and I  $ I . Thus there are many
different states available to the system for which there corresponds only one macrostate
(designated by the energy I of the system). Notice also that a given interval $ B corresponds to a
larger number of cells lying between the two ellipses when B ¶ A (the maximum amplitude) than
when B ¶ 0! In fact if you plot the number of accessible states available to the system as a
function of position B, it is clear that this is a classical, and not a quantum mechanical description
of the system!

STATISTICAL ENSEMBLES
In principle, the problem of a system consisting of many particles is completely deterministic. Classically,
if we know the momentum and position of every particle at time >o , then by solving Newton's equations of motion
for the well defined electric forces of interaction, we can determine the position and momentum of every particle
at a later time >. Likewise, for a quantum system, if we know the initial wave function for the system at time >o we
can use Schrödinger's equation to determine the wave function at any other time >, and therefore the probability of
finding a particle at some position B with a particular value of momentum :. However, we typically do not know
this information precisely for all particles which are involved. And even, if we did, the detailed information
which we would gain from solving the equations (if we could even do it) would be enormous. It would be very
difficult even to determine trends in the mountain of data that would be generated. We therefore proceed to
discuss the system in terms of probability concepts. Thus, we do not consider a single, well-defined system, but
rather a collection, or “ensemble", of essentially identical systems. [Now what we mean by essentially identical is
that for certain macroscopic parameters, the systems are essentially identical, whereas each individual system in
the ensemble may actually be slightly different on the microscopic level.]

Example: Consider a system of three fixed particles, each having spin "# , so that each spin can be
pointed either “up" or “down". Since the magnetic moment . t of the particle is associated with the
t
t œ .o S), the potential energy of each of the particles is determined by the projection of the
spin (.
spin along the axis of an applied magnetic field according to

ti † Ht œ
|.ti | |H|
Ui (H) œ
. t sin) œ
.o H mi (6.12)
where 73 œ „ "# . Thus the state of the whole system can be specified by determining the quantum
numbers m" , m# , and 73 . We list the possible states and the associated potential energies in the
following table.
Chapter Six: Specification of the State of a System 6

State Quantum Numbers Total magnetic Total

index r m1 , m2 , m3 moment energy
__________________________________________________________________________
1    3 .o
3 .o H

2  
.o
.o H
3 
 .o
.o H
4
  .o
.o H

5 


.o .o H
6


.o .o H
7


.o .o H

8



3 .o 3 .o H

One usually has some additional information about the system, such as the energy of the
system. This information restricts the possible microstates of the system to those which
correspond to this energy. Thus, a knowledge of the macrostate of the system limits us to those
microstates which are accessible to the system. For example, if the system above is known to have
an energy equal to
.o H, then the only allowed microstates correspond to states 2, 3, and 4. But
we do not know in which of these states the system may actually be!

FUNDAMENTAL POSTULATE
For an isolated system which cannot exchange energy with its surroundings, the total energy of the system
must be a constant (at least after having been isolated for a long period of time). This means that there are only a
certain number of microstates accessible to the system. Each of these microstates, then, make up the initial
ensemble of states for our problem. We assume that there is no reason why any one of these microstates is
preferred to another, so a priori we assume that all microstates are equally probable! This is the fundamental
assumption of statistical thermodynamics. Specifically stated this postulate says that any isolated system in
equilibrium is equally likely to be in any of its accessible microstates.
Chapter Six: Specification of the State of a System 7

Problems

6.1 Consider an electron moving in one dimension, which is constrained to be in a small region, 10
* m long,
and whose speed must be less than 10( m/s (i.e., the velocity is between 10( and 10
( m/s). Roughly how
many different states are available to this electron? Use the semi-classical approximation where
29 œ h œ " ‚ "!
$% J † s.

6.2 Consider a proton moving in three dimensions whose motion is confined to be within a nucleus (a sphere of
radius 10
"% m) and whose momentum must have magnitude less than :9 œ "!
"* kg m/s. Roughly how
many different states are available to this proton? Use the semi-classical approximation where
29 œ h œ " ‚ "!
$% J † s. [Hint: The volume of a sphere of radius :9 is 41:9$ /3.]