A simple method to estimate entropy of atmospheric gases from their action

Ivan R. Kennedy, Harold Geering, Michael T. Rose and Angus N. Crossan

SUNFix Laboratory, Faculty of Agriculture and Environment, University of Sydney,

NSW 2006, Australia

*Correspondence to: [email protected]

A convenient model for estimating the total entropy (ΣSi) of atmospheric gases based on
physical action is proposed. This realistic approach is fully consistent with statistical
mechanics, but uses the properties of translational, rotational and vibrational action to
partition the entropy. When all sources of action are computed as appropriate non-linear
functions, the total input of thermal energy (ΣSiT) required to sustain a chemical system
at specific temperatures (T) and pressures (p) can be estimated, yielding results in close
agreement with published experimental third law values. Thermodynamic properties of
gases including enthalpy, Gibbs energy and Helmholtz energy can be easily calculated
from simple molecular and physical properties. We propose that these values for entropy
are employed both chemically for reactions and physically for computing atmospheric
profiles, the latter based on steady state heat flow equilibrating thermodynamics with
gravity. We also predict that this application of action thermodynamics may soon provide
superior understanding of reaction rate theory, morphogenesis and emergent or self-
organising properties of many natural or evolving systems.

Keywords
Action thermodynamics, statistical mechanics, partition functions, translational entropy,
rotational entropy, vibrational entropy, Gibbs and Helmholtz energies, Sackur Tetrode
equation, adiabatic processes

Introduction
As defined by Clausius (1875), entropy can be considered as measuring the “self
reservoir of heat required to raise the temperature of a system of ideal gas molecules to
the temperature T”. Thus, in agreement with the third law of thermodynamics, at the
absolute zero of temperature Kelvin (K) the entropy should also be zero.

1
Clausius named entropy using the Greek word for ‘in-turning’ or transformation, a
dynamic definition that this paper will show is highly apt. As a property of state
(pressure, temperature and volume), irrespective of the path a system of molecules has
arrived at that condition, entropy is therefore an important feature of atmospheric gases.
Indeed, its capacity to explain how thermal radiation may be absorbed and partitioned
into different degrees of freedom of motion is key information in explaining the warming
potential of greenhouse gases.

The total entropy of an ideal gas molecule can be calculated as the sum of terms,

STotal = St + Sr + Sv + Se + Sn + …. (1)

where the subscripts refer to translational, rotational, vibrational, electronic and nuclear
entropy terms respectively. To estimate the total thermal energy needed to reversibly heat
(i.e. without doing other work) a system of gas molecules we need only multiply by the
temperature T. This thermal energy can be thought of as both kinetic and potential
energy contained in the field of the molecular system (Kennedy, 2000; 2001).
Furthermore, this thermal energy input is physically required to sustain the physical
action of the system (Rose et al., 2008).

This paper seeks to place thermodynamics within easy reach even of non-specialists by
giving a key role to action, a physical property that is realistically the focus of our
interest. Action is related to the vector angular momentum, but is a distinct scalar
quantity since it is independent of direction. Planck (1913) defined molecular action as a
macroscopic property of the product of the generalised space coordinates (r, cm) and the
generalised impulse coordinates (mv, gcm/sec), defining the microscopic state of the
molecule (mvr). Like entropy, action is an extensive or cumulative property, but with
physical dimensions of the integral of energy with time, or of the instantaneous angular
momentum with respect to angular motion; classically, action was considered as the
integral of momentum with distance. As a variable property of conservative systems,
action has been considered to take stationary values, a result sometimes referred to as the

2
principle of least action. In fact, all these viewpoints of action are equivalent. Note that
angular motion is the ratio of circumference to radius and thus is physically
dimensionless, although we still measure it in degrees or radians.

In illustration of the utility of action theory, we advance a unified model that we will
show is valid for calculating the entropy of atmospheric gases. In the range of ambient
temperatures relevant to the Earth’s atmosphere, this realistic model gives results for
absolute entropy closely consistent with previous experimental data. This action model
may help provide an approach to prediction of the rate of global warming based on causal
responses to the increasing greenhouse gas content of the atmosphere, rather than
statistical correlations. In particular, a direct relationship between thermodynamics and
gravity may provide a dynamic view of how the thermal properties of the atmosphere
have significance for warming and climate change.

Methods
Here we show how the classical formulae for estimating translational, rotational and
vibrational entropy using partition functions may be reviewed as physical action.
Although entropy data for atmospheric gases are readily available in standard tables, the
methods developed here illustrate how easily such data can be manipulated to account for
real environmental conditions.

Because of their relevance to the following methods, we state here the partition functions
for ideal gases currently used to calculate the entropies for molecular translation, rotation
and vibration. These functions are given in all standard texts on modern statistical
mechanics, the discipline founded jointly by Ludwig Boltzmann (1896) and J. Willard
Gibbs (1902). The factors governing these functions include absolute temperature (T in
o
K), Boltzmann’s constant (k = 1.3806x10-23 J K-1), Planck’s quantum of action (h =
6.626x10-34 J. sec) and the system volume (V). For ease of use and consistency in
dimensions, all modelling and calculations have been performed in cgs units, before
conversion to SI units where required.

3
Translational partition function

Qt = (2πmkT/h2)3/2V

Here V is taken as the system volume occupied by the molecules (Glasstone, 1951).

Rotational partition function (linear molecule)

Qr = 8πIkT/h2

Rotational partition function (non-linear molecule)

Qr = 8π2(8π3IAIBIC)1/2(kT/h2)3/2

Vibrational partition function (polyatomic molecules)

Qvi = Пi [1 – exp-hνi/kT]-1

hνi
where Пi indicates a product of i functions, for each mode of vibration.

It will be shown that these functions can all be considered as dimensionless statistical
measures of relative molecular action, using Planck’s quantum h or its reduced form per
radian, ħ, as a natural reference unit. Note that there are only three sources of variation in
the partition functions – inertial mass, temperature or the root mean square velocity and
pressure or density.

Translational entropy and the Sackur-Tetrode equation

The Sackur-Tetrode equation was published early in the 20th century (Glasstone, 1951),
based on Gibbs’ theory of statistical mechanics (1902).

St = R[ln(2πmkT)3/2V/h3N + 2.5] (2)

4
This equation allows calculation of the translational entropy of N molecules of an ideal
monatomic gas. This is derived from the relationship from the calculus of total entropy
including translational St and internal Sint parts as follows.

S = RT(∂lnQ/∂T)V + RlnQ - klnN! = St + Sint

= RT[(∂lnQtr/∂T) + (dlnQint/dT)] + R[lnQtr + lnQin] – klnN!

Here the factor klnN! allows for the inability to distinguish between N identical
molecules.

For ideal monatomic gases, no internal entropy (from rotation or vibration) at normal
temperatures exists, so the differential of the internal partition function Qint in the above
equation can be ignored.

St = RT[(∂lnQtr/∂T)] + R[lnQtr] – klnN!

Using a solution from the Schrodinger equation that Qt= (2πmkT/h2)3/2V

and the Stirling approximation for lnN! as effectively NlnN - N we have from Moore
(1963).

St = 3/2R + Rln[(2πmkT/h2)3/2V] – Rln(N- 1)

= R[ln(2πmkT)3/2V/(h3N) + 2.5]

Note the close similarity to the translational partition function. This is a well known
result of statistical mechanics, with h being the constant introduced by Planck (1913) as
the quantum of action for radiation. Despite their lack of rest mass, all energy quanta
possess an action of magnitude h and their energy is given by hν, where ν is their
frequency.

5
Here we introduce a revised approach, based on the use of the property of state, action
(@= mrv (Kennedy, 2000; 201). This allows the establishment of the relative action, a
ratio or pure number suitable for logarithmic expression. Some expressions of entropy in
text books include isolated terms such as the logarithm of the temperature lnT. Strictly,
this is invalid, as logarithms can only be taken of pure numbers or dimensionless ratios
and never of quantities with physical dimensions.

For each of the three translational degrees of freedom, the translational action @t can be
derived from the kinetic energy, given each has ½kT of kinetic energy. For three degrees
of freedom we can estimate the action, mvr or Itωt, as follows. The three-dimensional
kinetic energy ½mv2 is 3/2kT which equals ½mrt2ωt2 or ½Itωt2, for motion with polar
coordinates; here the translational angular velocity ωt or dӨ/dt is given in radians per
sec. Thus Itωt2 is equal to 3kT and so Itωt is equal to 3kTIt/ωt or (3kTIt)1/2. The
translational action @t is thus defined as equal to (3kTIt)1/2. It should be noted that 3kT is
a statistical result from the three-dimensional Maxwell distribution equal to twice the
most probable kinetic energy ½mv2 for the root mean square velocity v in three
dimensions; moreover, 50% of molecules have speeds greater than the root-mean square
velocity, which is about 1.09 times the mean speed of the ideal gas molecules (Brown,
1968, Table 8). In the Maxwell distribution, the most probable velocity is slightly less
than either of these speeds.

We can regard the system volume V as containing N cubic cells of volume a3, a cell for
each gas molecule. Then, for rt arbitrarily taken as the mean value of the half-distance
between the centres of any two nearest neighbour gas molecules, a3 is equal to (2rt)3 or
8rt3. Considering a mole of gas at standard temperature and 1 atmosphere pressure
(N=6.022169x1023 molecules in 24465.1 mL at STP), then rt or (V/N)1/3/2 is equal to
1.7188x10-7 cm.

We can then substitute into the Sackur-Tetrode equation (St = Rln[8e5/2rt3(2πmkT)3/2/h3]).

Taking the rt3 term inside the brackets, we have St equal to Rln[8e5/2(2πmrt2kT)3/2/h3].

6
Then, taking mrt2kT equal to kTIt and @t equal to (3kTI)1/2

St = Rln[8(2π/3)3/2e5/2(@t/h)3]

For ħ equal to h/2π, the reduced quantum of action, we have,

St = Rln[e5/2(2/3π)3/2(@t/ħ)3]

If the factor of (2/3π)3/2 were to be incorporated into the action term, the inertial radius
would be decreased to 7.92287x10-8 cm rather than 1.7188x10-7 cm. However, we simply
assume in equation (3) a translational resonance symmetry factor zt of 10.2297 or
(2.1708)3, replacing (2/3π)3/2.

St = Rln[e5/2(@t/ħ)3/zt] (3)

Given the disorderly nature of the spatial distribution of the translating particles (see
Figure 1), the symmetry factor cannot have a precise continuous value, but should be
statistically distributed and fluctuating with time for each molecule (see Figure 1). With
real gases having varying degrees of interaction or binding, the most probable radius may
vary from one gas to another. To maintain consistency with Sackur-Tetrode theory, all
calculations in this paper have employed the geometric result of 1.7188x10-7 cm for rt at
STP, or 1.6994x10-7 cm at the Earth’s global average surface temperature of 288.15 K.
However, it may be possible to make experimental determinations of the dynamic
structure, allowing more accurate estimates of the action. In any case, the sensitivity to
variations is low given its logarithmic nature and any errors will only be a slight
displacement in the entropy value. Furthermore, the absolute value of the translational
symmetry factor zt is rarely of importance because in nearly all cases, differences in
entropy of free energy are being taken, or the system is isothermal. In such cases, the zt
factor disappears.

7
As a result, with suitable choices of the radius and translational symmetry (zt), we can
write the following concise relationship for translational entropy.

St = Rln[e5/2(@t/ħ)3/zt]
= 2.5R + 3Rln(@t/ħ) – Rlnzt

Rotational action and entropy

From statistical mechanics (see Moore, 1963), the rotational contribution to the molar
entropy of a diatomic or linear molecule with two-dimensional inertia is given by

Sr = R + Rln(8π2kTIr/σrh2), or Rln[e(8π2kTIr/σrh2)] = Rln[e(2kTIr/ħ)2/σr]

This is two-dimensional only, because there is no significant inertia around the

longitudinal axis of a diatomic or linear molecule like N2 or CO2. The rotational partition
function is 8π2kTIr/h2; clearly, this can also be recast as an action ratio. Here the moment
of inertia Ir is given by (m1m2/m1+m2)rr2 and rr is the average bond length, σr is the
rotational resonance symmetry number (e.g. σr = 2 for O2 and σr = 1 for NO). In this
equation we can recognize the rotational action of a gas molecule, @r equal to (2kTI)1/2 –
derived from the rotational energy equal to ½mrr2ωr2 or Irωr2/2. So Irωr2 equals 2kT and
Irωr is given by (2kTI)1/2 equal to @r, by definition. As a result, using similar notation as
for translational action and entropy, we have

Sr = Rln[e(@r/ħ)2/σr] = R + Rln[(@r/ħ)2/σr] (4)

Thus, just as in the case of translational entropy, the rotational entropy can also be
expressed as a variable of action alone, given that the symmetry factor σr is a constant. It
is shown elsewhere that action is a function of volume and temperature (Kennedy, 2001)
but volume or pressure changes have little or no effect on rotational entropy as long as
the temperature is not too high. All other terms in this equation are constant for a given
gas molecule.

8
For non-linear gas molecules with more than two atoms such as the greenhouse gases,
other than the linear carbon dioxide and nitrous oxide, the rotational entropy Sr is given
from statistical mechanics as,

Sr = Rln[{8π2(8π3IAIBIC)1/2(kT)3/2}/σrh3] + 3/2R,

where IA, IB and IC correspond to the three principal moments of rotational inertia with
respect to three perpendicular axes (see Glasstone, 1951). In terms of action ratios
analogous to those used above, we rearrange this equation to read,

Sr = Rln[π1/2{(8π2kTIA/h2)1/2(8π2kTIB/h2)1/2 (8π2kTIC/h2)1/2}/σr] + 3/2R

Recalling our previous definition of rotational action @r of a diatomic molecule as

(2kTIr)1/2 for each inertial axis of a linear molecule with more than one atom, we can
express the rotational entropy contribution of a non-linear molecule as,

Sr = Rln[{π1/2(@A/ħ)(@B/ħ)(@C/ħ)}/σr] + 3/2R

= Rln[{π1/2e3/2(@A@ B@C/ħ3)}/σr] (5)

where @A, @ B, and @C are the three principal rotational actions for non-linear molecules.
So once again, it is possible to express changes in entropy as a simple function of action
alone, as all other terms in the equation are constant for a given gas molecule. Given that
the product of entropy and temperature ST indicates the thermal energy required, there is
obviously an exact logarithmic relationship between the total energy required to sustain a
system of molecules at a given temperature and the action of each mode of rotation

The rotational symmetry number σr for polyatomic molecules depends on the point group
of the molecule as defined by the Nobel laureate Herzberg (1945), by “A possible
combination of symmetry operations that leaves at least one point unchanged is called a

9
point group”. This is a term derived from crystallography and the characteristic symmetry
number σ of each point group can be shown to be equal to “the number of
indistinguishable positions into which the molecule can be turned by simple rigid
rotations” (Herzberg, 1950). Table 1 adapted from Herzberg (1945) gives the symmetry
number for the more important point groups. Note that methane in the T point group has a
rotational symmetry of 12, indicating how its quantum field indicated by its rotational
entropy is economical for energy in view of its indistinguishable structure regarding its
orientation in space. This situation for methane can be contrasted with a similar
tetrahedral carbon molecule having only one hydrogen atom in its structure, together with
three different halogens such as fluorine, chlorine and bromine. In this case the symmetry
is unity (1.0), so that the energy field has a 12-fold lower frequency of encountering an
identical structure in action space.

Table 1: Symmetry numbers for various point groups

Point group Symmetry No. Point group Symmetry Point group Symmetry No.
σr No. σr σr

C1, Ci, Cs 1 D2, D2d, D2h Vh 4 C∞ 1

C2, C2v, C2h 2 D3, D3d, D3h 6 D∞h 2
C3, C3v, C3h 3 D4, D4d, D4h 8 T, Td 12
C4, C4v, C4h 4 D6, D6d, D6h 12 Oh 24
C6, C6v, C6h 6 S6 3
Modified from Herzberg (1945), p. 508

Quite naturally this 3-dimensional relationship we will see later is related to chemical
potential rather than kinetics, must be exponential. So more energy per molecule is
required to convey information in a molecular field of greater probe-ability,with greater
radial separation of identical molecules.

Vibrational action and entropy

Vibration in molecules between their atoms occurs at a rate usually much more frequent
than rotation, clearly adding to the action of molecules by increasing the length of the
trajectory of their atoms in a given time interval. Vibration is characterized by its
frequency, which does not change appreciably as more energy is added to a system. On
the contrary, the amplitude of vibration increases affecting the mean kinetic and potential
energies and the inertial trajectory of the molecular angular motion. In collision
10
processes, a higher non-equilibrium vibrational energy state resulting from absorption of
a quantum of energy in the infrared will act to equilibrate with its rotational and
translational energies in the micro-wave and radio-wave bands, effectively dissipating
vibrational energy into these modes. This concept will be revisited later when we
consider greenhouse gases and how they can affect the gravitational distribution of the
atmosphere.

Related to the point groups and numbers of atoms in the gas molecule are the number of
vibrational modes; also, the question arises of how many of these modes are degenerate,
that is some different vibrations have identical frequencies because of similarities in
molecular structure. For a non-linear molecule the number of vibrational modes is equal
to 3n-6 where n is the number of atoms. For a linear molecule, the number of possible
modes is one less or 3n-5. In either case, 3n is the total degrees of freedom for motion.
The number 6 in the case of non-linear molecules is the sum of 3 translational and 3
rotational degrees of freedom. The number 5 in the case of linear molecules refers to 3
translational and 2 rotational degrees of freedom.

For both non-degenerate and degenerate modes, the vibrational entropy Svi is given by
Glasstone (1951),

Svi = Rx/(ex – 1) – Rln(1 – e-x), where x = hcνi/kT (6)

Here νi is the wave number equals the number of vibration per sec divided by the velocity
of light in cm per sec. It therefore has the physical dimensions of cm-1. Ultimately, the
total contribution to the vibrational entropy is the sum for all vibrations, taking into
account any degeneracy where more than one mode of vibration has the same frequency.

This equation is derived as follows. According to Moore (1963), the vibrational energy E
is given as,

E = RT2∂lnQvib/∂T = Lhν/2 + Lhνe-hν/kT/(1 - e-hν/kT)

11
Here Lhν/2 is the zero point vibrational energy Eo remaining at absolute zero Kelvin,
where L is Avogradro’s number for the number of molecules in a mole. Thus, taking
hν/kT equal to x, as used above,

(E –Eo)/T = Rxe-x/(1 - e-x)

(Avib – Eo)/T = Rln(1 – e-x) since Avib = -kTlnQvib = Gvib

Here A and G refer to Helmholtz and Gibbs energies.

So Svib = (E – G)/T = Rxe-x/(1 – e-x) - Rln(1 – e-x) as given above.

Furthermore, the vibrational heat capacity is given as

∂E/∂T = Cvib = Rx2e-x /(1 – e-x)2 = Rx2 /(ex + e-x –2)

= Rx2/2(coshx – 1), given (ex + e-x)/2 = coshx

Unlike the cases of translational and rotational entropy, it is not obvious that SvibT
includes the vibrational kinetic energy. Indeed, it is only at temperatures above 470 K
that SvibT exceeds CvibT. Despite this, it is clear that each vibration of frequency νi
contributes its own entropy.

Sν = ΣSνi

In Figure 2, a flow diagram for computing entropy and Gibbs energy is given.

12
Results

Data on molar entropies and absorption wavelengths of greenhouse gases

We have established several concise relationships expressing entropy as a logarithmic
function of an action ratio for translation and for rotation, with an exponential overlay on
rotation by vibration.
St = Rln[e5/2(@t/ħ)3zt] (translation)
Sr = Rln[e(@r/ħ)2/σr] (rotation – diatomic or linear molecule)
Sr = Rln[π1/2e3/2(@A@ B@C/ħ3)/σr] (rotation- polyatomic molecule)
x -x
Svi = Rx/(e – 1) – Rln(1 – e ), where x = hcνi/kT (for each vibrational mode)

For a diatomic molecules like nitrogen or oxygen making up most of the Earth’s
atmosphere we have for an action thermodynamics formulation,

S/N = s = k{ln[e7/2(@t/ ħ)3(@r/ħ)2Qe/(σrzt) ] + hν/kT/(ehν/kT - 1) – ln(1 – e-hν/kT)} (7)

or (given vibration is largely irrelevant for these molecules near 300 K, though
not for carbon dioxide)

S/N = s = k{ln[e7/2(@t/ ħ)3(@r/ħ)2Qe/(σrzt)] (8)

We can compare this to the equivalent equation from statistical mechanics (Hill, 1960).

S/N = s = k{ln[(2πmkT/h2)3/2Ve5/2/N] + ln[(8π2kTIrh2)e/σ] + [(hν/kT)/(ehν/kT- 1) –

ln(1 - ehν/kT)] + lnωe} (9)

It is clear that the translational action ratio @t/ħ will vary as a function of temperature
affecting velocity, but also with volume. In equations (7) and (8), action acts as a
surrogate for the effects of both temperature and volume or density for translational
entropy contained in equation (9). Normally, these variables are considered separately as
shown in (9). At extremely low temperatures near absolute zero, the action ratio will tend
to a minimum and the entropy will tend to zero, as required by the third law of
thermodynamics. Near zero, only vibrational energy remains significant, expressed as the

13
zero point vibrational energy of hν/2 per bond proposed as essential by Planck and
Einstein.

The action-based forms of equations (7) and (8) are much more amenable for modeling
and for understanding than the arbitrary forms of equation (9).

Given the results obtained here, there can be no doubt of the exact co-variation of entropy
with action, accepting that the entropy contains additional terms related to kinetic energy
or enthalpy as well as for symmetry. Causally, this must be based on the need for specific
quantities of field energy to sustain action at a statistically stationary value, as appropriate
for a particular kinetic environment and temperature. The statistical nature of entropy
implicit in Boltzmann’s and Gibbs’ theories must also correspond with the relationship
with action, discussed in Kennedy (2000). This statistical correspondence can be found
in the space-filling dynamic nature of molecules subject to collisions, so that any
complexion of low probability or high pressure involves chemical species occupying a
comparatively small volume per molecule with low translational action as controlled by
temperature; those complexions of higher probability or entropy occupy a comparatively
large volume per molecule with higher translational action. This is also consistent with
Shannon’s information version of entropy, considering information as uncertainty and the
capacity of a message to surprise (Shannon and Weaver, 1949). When an encounter is
less frequent with greater diversity of species, surprise is more likely.

In Tables 2-5 calculations using these equations are indicated for individual contributions
to the molar entropy of some common atmospheric and greenhouse gases. The computed
entropy values at standard temperature and pressure compare very well with the rounded
standard values in reference (Aylward and Findlay, 1974) obtained experimentally.

14
 Table 2a: Entropy of linear gas molecules – translational, rotational
Gas MW It @t/ħ St=Rlne5/2 Bond Ir @r/ħ Sr= 1/λ x= Sv Qe Se ΣS
x1040 = nt x(nt)3/zt] radius x1040 = nr Rlne(nr)2/σ cm-1 h ν/kT
g.cm2 J K-1 x1010 g.cm2 J K-1
cm
H2 2.0000 98.83 104.7569 117.48 74 0.4580 1.8413 12.70 - - - 1 0 130.18
N2 28.0134 139.02 392.9013 150.45 110 14.235 10.2654 41.27 - - - 1 0 191.73
O2 31.9988 158.12 419.0198 152.06 121 19.590 12.0426 43.93 1580 7.63 0.04 3 9.13 205.16
CO 28.0110 138.33 391.9225 150.39 113 14.643 10.4115 47.27 2170 10.47 0.00 1 0 197.67
NO 30.0061 148.54 406.1341 151.28 115 16.555 11.0706 48.29 1904 9.188 0.01 4 11.5 211.12
CO2 44.0099 217.42 491.3535 156.03 244 79.665 24.2846 54.72 See below 2.99 1 0 214.61
N2O 44.0134 215.90 489.6496 155.94 66.9 22.2550 59.90 See below 3.05 0 218.89
Data obtained from Herzberg (1945, 1950) and Aylward and Findlay (1974)

Table 2b: Triatomic gases translational and rotational entropy

Gas MW It x10-40 @t/ħ St IrA IrB IrC @rA/ħ @rB/ħ @rC/ħ σr Sr Point
g.cm2 JK-1 x10-40 g.cm2 x1040 JK-1 group
H2O 18.0154 88.372 313.2678 144.80 1.024 1.920 2.947 2.7533 3.7699 4.6709 2 43.74 C2v
H2S 34.080 167.18 430.8673 152.75 2.667 3.076 5.845 4.4435 4.7721 6.5779 2 52.52 C2v
O3 47.9982 235.45 511.3358 157.02 7.877 62.865 70.900 7.6366 21.5796 22.9104 2 79.94 C2v
SO2 64.0628 314.25 590.7399 160.62 13.807 81.328 95.356 10.1103 24.5377 26.5697 2 84.58 C2v

Table 2c: Triatomic gases vibrational entropy

H2O Wave x= Svi CO2 Wave x= Svi Degen ∑Svi
number hcνi/kT number hcνi/kT
A1 3652 17.6235 <0.0001 σg+ 1388 6.6981 0.0790 1 0.0790
A1 1595 7.6970 0.0329 Π 667 3.2188 1.4547 2 2.9093
B2 3756 18.1254 <0.0001 σu+ 2349 11.3356 0.0012 1 0.0012
Total Total 0.033 Total 2.9895

H2S Wave x= Sv N2O Wave x= Sv

number hcνi/kT number hcνi/kT
A1 2615 12.6193 0.004 ∑ 2224 10.7324 0.0002
A1 1183 5.7088 0.1856 ∑ 1285 6.2011 0.1216
B2 2626 12.6723 <0.0001 Π 589 2.8423 1.4627
Total 0.1860 Π 589 2.8423 1.4627
Total 3.0473
O3 x= SO2 x=
hcνi/kT hcνi/kT
A1 1110 5.3565 0.2504 A1 1151 5.5544 0.2117
A1 705 3.4021 1.2561 A1 518 2.4997 2.5715
B2 1042 5.0284 0.3301 B2 1352 6.5244 0.0919
1.8367 2.8751

Table 3a: Polyatomic molecules

Gas MW Itx1040 @t/ħ St IrA IrB IrC @rA/ħ @rB/ħ @rC/ħ σ Sr Point
g.cm2 J K-1 x1040 g.cm2 x1040 J K-1 group
NH3 17.031 83.543 27.9881 144.10 2.9638 2.9638 4.5176 4.6841 4.6841 5.7830 3 48.36 C3v

Species Wave x= Sv
number hcνi/kT
A1 3337 16.1034 0.0001
A1 950 4.5844 0.4785
E 3447 16.6343 0.0001
E 1627 7.8514 0.0287
0.5074

15
 Table 3b: Polyatomics
Gas MW It x1040 @t/ħ St IrA IrB IrC @rA/ħ @rB/ħ @rC/ħ σ Sr Point
Daltons g.cm2 JK-1 x1040 g.cm2 x1040 JK-1 group
CH4 16.401 78.678 295.6225 143.35 5.27 5.27 5.27 6.2461 6.2461 6.2461 12 42.263 Td
CFCl3 137.37 673.84 865.0408 170.13 340.35 340.35 799.71 50.1970 50.1970 62.433 3 107.59 C3v
CF2Cl2 120.91 593.12 811.5780 168.54 203.73 318.00 375.66 38.8364 48.5206 52.737 2 107.13 C2v
CF3Cl 104.46 512.41 754.3356 166.71 146.32 251.58 251.58 32.9126 43.1566 43.156 3 99.745 C∞v
6
CH4 Wave x= Sv Degen. Sv CFCl3 Wave x= Sv Degen. Sv
number hcνi/kT Number hcνi/kT
A 2914 14.063 0.0001 1 0.0001 A 1085 5.2359 0.2773 1 0.2733
E 1526 7.3640 0.0441 2 0.0882 A 535 2.5818 2.4105 1 2.4105
T 3020 14.575 0.0001 3 0.0002 A 350 1.6890 4.8792 1 4.8792
T 1306 6.3034 0.1113 3 0.3340 E 847 4.0874 0.7208 2 1.4416
Total 0.4225 E 394 1.9013 4.1210 2 8.2420
E 241 1.1630 7.5121 2 15.024
Total 32.275
CF2Cl2 cm-1 x= Sv CF3Cl Band cm-1 x= Sv Degen. Sv
hcνi/kT hcνi/kT
A 1101 5.3131 0.2598 A 1105 5.3324 0.2556 1 0.2556
A 667 3.2188 1.4547 A 781 3.7689 0.9344 1 0.9344
A 458 2.2108 3.2297 A 476 2.2970 3.0163 1 3.0163
A 262 1.2643 6.8968 E 1212 5.8488 0.1646 2 0.3293
A 322 1.5539 5.4387 E 563 2.7169 2.1667 2 4.3334
B 902 4.3528 0.5796 E 350 1.6890 4.8792 2 9.7584
B 437 2.1088 3.4981 Total S 18.627
B 1159 5.5930 0.2048
B 446 2.1523 3.3804
Total S 24.945

An interesting feature of the data calculated for Tables 4 and 5, but rarely nechanistically
considered, is the significantly greater entropy related to translational action compared to
rotational and vibrational action for all gas molecules at 298 K. The lower the pressure of
a particular gas the greater is this discrepancy. Indeed, most of the heat required to raise
the temperature of the gas from zero to 298 Kelvin under standard conditions is devoted
to sustaining translational action, with only a small proportion of molecules exhibiting
any vibrational action and entropy at all. If we consider that the relative action states @/ħ
for carbon dioxide at standard temperature and pressure calculated here are 491 for
translation, 24 for rotation and a much lower number for excited vibration, indicating a
very low proportion of molecules excited with infra-red quanta, it is reasonable to
conclude that the size of the quanta associated with rotation and translational action states
have correspondingly lower frequencies. This would appear to place them in the
microwave and the radio-wave range of the electromagnetic spectrum – relatively cold or
dark energy.

16
Thus, although heat is required to generate translational entropy, could its actual form be
considered as gravitational work, although for sub-orbital molecules with quanta of
relatively high frequency. If so, we could consider translational action and entropy as
indicating the quantity of heat required to reach T degrees Kelvin in the gravitational
field, including that normally regarded as pressure-volume work.

Table 4: Summary of total entropy terms

Gas St Sr Sv S (A&F) IR spectrum Wavelength m (3n-6)
JK-1 JK-1 JK-1 JK-1 JK-1
H2O 144.80 43.74 0.033 188.6 189 2.662, 2.738, 6.270
CO2 155.94 54.715 2.99 213.6 214 4.257, 7.2046, 14.993
H2S 152.81 52.54 0.19 205.5 206 3.808, 3.824, 8.453
N2O 155.94 59.80 3.05 218.9 220 4.446, 7.782, 16.978
O3 157.02 79.94 1.84 238.8 239 9.009, 9.597, 14.184
SO2 160.62 84.58 2.875 248.1 248 7.396, 8.688, 19.305
NH3 144.10 48.36 0.507 193.0 192 2.901, 2.997, 6.146, 10.526
CH4 143.36 42.26 0.423 186.0 186 3.311 (2), 3.432 (3), 6.553 , 7.657 (3)
CFCl3 170.14 107.59 32.275 310.0 310 9.217, 11.806(2), 18.692, 25.381(2), 28.571, 41.494(2)
CF2Cl2 168.545 107.137 24.945 300.6 301 8.628, 9.083, 11.086, 14.999, 21.834, 22.421, 22.883, 31.056, 38.168
CF3Cl 166.721 99.749 18.627 285.1 286 8.251(2), 9.050, 12.804, 17.762(2), 21.008, 28.571(2)
O2 162.07 43.93 0.035 206.0 205 6.329
CO 150.31 47.19 0.0025 197.5 198 4.608
NO 162.69 48.39 0.0087 211.6 211 5.252
H2 117.48 12.70 - 130.2 131 -
N2 150.45 41.27 - 191.7 192 -
A 154.84 - - 154.8 -

We must regard the heat that was required to melt and vaporise the carbon dioxide
molecules as having performed configurational work, either on separating the molecules
or pressure-volume work of lifting the atmosphere. This work-heat consumption
identified by Clausius (1875) explains the fact that not all the heating included in the
entropy function contributes to sensible heat just raising the temperature as increased
kinetic energy, consistent with the heat capacity of each molecule. Similar conclusions
can be drawn for the melting of ice to water and its subsequent vaporization. Such
interpretations regarding inter-conversions of heat and work require further investigation
and this may be facilitated using the quantum features of the action approach.

Considering water’s phase changes

17
Of all the atmospheric gases considered here, only water exists in the atmosphere on
Earth as gas, liquid and solid. This erratic cycle is very apt for illustrating significant
changes in entropy states associated with changes in phase. Most of the permanent gases
in the atmosphere only exist as vapors. As a result, their changes in entropy refer only to
changes in kinetic and potential energy corresponding to changes in enthalpy and in free
energy as a response to changes in temperature and pressure respectively. However,
water has highly significant changes in action and entropy in the atmospheric weather
cycle, with corresponding consumption or release of heat.

The total thermal capacity to bring a mole of water to vapor at 298.15 K and 1
atmosphere pressure (were this possible) is 56.2 kJ. Of this, only 7.4 kJ can be attributed
to its heat capacity as increased kinetic energy per mole over the temperature range, with
5.8 kJ required for melting and 44.0 kJ to vaporisation at 298 K. Almost 90% of the solar
heat absorbed by water as vapor in the atmosphere is available for release in the two
phase transitions of forming snow or hail. Although these facts are well known and the
major warming possible during atmospheric condensation of water vapor is understood,
this could also be a fruitful area for further investigation using the action-entropy theory.

It is of interest from Table 4 that water has only minor vibrational entropy – lower even
than that of oxygen, mainly at a very similar short infrared wavelength (6.270 and 6.329
μm) – potentially allowing overlap for emission or absorption. On this basis, there might
seem to be only a weak case to consider water a major greenhouse gas, although this is
customary. In fact, the converse is true as its low actual excitation of vibration,
particularly by the two shorter wavelengths for water is that there remains a high
population of water molecules able to be excited by radiation from the Earth’s surface.
However, the reversible latent heat of vaporisation of water released when it condenses in
clouds or at dewpoint is also an important factor for heat transfer in the atmosphere.

It is also apparent that the actual vibrational entropy of methane is only about one-sixth
that exhibited by carbon dioxide and nitrous oxide. Presumably the low vibrational
entropy can also be related to a higher residual absorptivity of methane, but it is only

18
slightly statistically enhanced by being poorly excited at this temperature. Only a very
low proportion of methane as well as water molecules are excited by infrared radiation at
equilibrium under ambient temperature conditions, as shown by their low vibrational
entropies at 298.15 K.

Shown in Tables 3a,3b and 4, the organo-halogens such as Freon 11 (CFCl3) that have
been withdrawn from use under the Montreal Protocol have an exceptionally large
vibrational contribution to entropy. Replacing the hydrogen atoms of methane with these
two halogen atoms also significantly increases both the rotational and vibrational
entropy; this should lessen absorptivity in the longer infrared region significantly since
this is relatively excited at 298 K as a result of longer bond lengths and greater ease of
dissociation of atoms. According to Glasstone (1951), the formula for calculating
vibrational entropy strictly applies only to divalent molecules. However, this cautionary
note may not be required. By applying the formula to each bond separately and
summating as shown in Table 3b, including any degeneracy, the agreement with
experimentally determined entropies using the Third law approach is just as good as for
other molecules where only translational and rotational contributions are significant.

Nitric oxide (NO), although not a greenhouse gas with only one vibrational line in the
short-wave infrared (5.252μm), is included for comparison, as are CO (4.608 μm) and O2
(6.329 μm) (Tables 2, 4). For nitric oxide (NO) a large discrepancy in total entropy
between the data calculated here from translation, rotation and vibration would occur if
the electronic (Qe = 4) term was neglected, a result of its free radical nature containing
unpaired electrons; these add Rln22 or 11.53 extra entropy units per mole, giving a total
value of 211.1, in agreement with the Aylward and Findlay value (1974).

The Sackur-Tetrode equation includes a term for the electronic partition function (Qe) .
In the case of O2, the ground state electronic partition function Qe is 3 at STP because this
molecule has two unpaired electrons that can have their two spins oriented three ways
with respect to the nuclear spin – both up, both down and oppositely. Because they can
be distinguished, the three different oxygen species have three times the volume per

19
particle, affecting their action because of the greater radial separation than if only a single
species existed. This gives an additional electronic entropy contribution of Se = RlnQe or
Rln3.

However, the fact that a mole of oxygen will contain one-third of a mole of each species
must be also be considered in estimating the total entropy. This Qe factor is included in
the Sackur-Tetrode equation as St=Rln[Qee5/2(V/N)(2πmkT)3/2/(h3)], or in the action form
of the equation as Rln[e5/2(@t/ħ)3Qe/zt], which is equal to RlnQe – Rlnzt + 2.5R +
3Rln(@t/ħ). Thus asymmetry (Qe) increases entropy by increasing spatial distance
between molecular interactions and symmetry (zt) decreases it by reducing spatial
distances– less field energy is needed to sustain a symmetrical molecule than an
asymmetrical one.

Because of the statistical variation in quanta for rotational and translational fine structure,
the actual vibrational spectra are not confined to these spectral lines but distributed
around these wavelengths. The spectra may be sharpened by cooling the gas and this is
usually done when testing the theory with data.

The data in Tables 2-4 are calculated for standard conditions of temperature and pressure.
To adjust these results to the actual gas pressures in the atmosphere at the same
temperature, only the translational action and entropy will vary. In terms of centimeter-
gram-second (cgs) units, the pressure, is equal to kTa-3 or kT/8r3 at 1.013x105 pascals,
being the product of the mass of air per square cm of the earth’s surface (ca. 1 kg) and the
acceleration of gravity (9.807 m.sec-2).

It should be understood that entropy is a dimensionless number corresponding to its role

in probability, given that it expresses a total thermal capacity per degree of temperature –
a ratio of extensive and intensive measures of energy. It is also instructive to be aware
that the product of entropy and absolute temperature (ST) is always a significant multiple
of the kinetic energy since that is merely one of its components – to its kinetic energy
must be added the sustaining field energy corresponding to decreases in free energy from

20
absolute zero while heating the molecules, absorbing any heat that becomes latent during
this process and in doing any work such as breaking H-bonded aggregated structures or
pressure-volume work against the atmosphere. In this connection, much of the magnitude
of ST is generated together with increased enthalpy during phase changes when parent
solid or liquid matter is melting or vaporizing. Gibbs energy does not change when these
reversible processes occur isothermally. The chemical potential of the liquid water is
equilibrated with that of the vapour at the boiling temperature, the increase in the
enthalpy on vaporization being effectively an increase in internal entropies associated
with increasing the internal vibration and rotation of the de-clustered water molecules.
Such increases in internal action and entropy are actually increases in enthalpy.

Any decrease in the density of a chemical substance such as the expansion of a solid,
liquid or gas will also increase ST as its Gibbs free energy decreases. For example, liquid
water gradually changes its state during heating from large H-bonded clusters of about 30
water molecules just above freezing to fewer than half that number per cluster just below
boiling temperature (Kennedy, 2001), above which the clusters completely dissociate.
The variable action of these flickering clusters between zero Celsius and 100 C could be
calculated and the changes in entropy estimated.

Some results calculated for the actual sea level pressures (ppmv) of all atmospheric gases
at the standard temperature of 298.15 K are given in Table 5. Entropy values are given
per mole of each substance, using Boltzmann’s constant k multiplied by Avogadro’s
number N as the unit value and the product [entropy x temperature, ST] estimated for
each as a proportion of the total.

The total entropic energy as ΣST in air at sea level is about 2.4 MJ per cubic metre. It is
obvious that the very dilute gases like nitrous oxide and methane have a relatively large
translational entropy compared to the major gases and therefore need more heat per
molecule to bring them to this temperature and pressure. The majority of the heat
required (ST) to raise the atmosphere is absorbed into the fields of only three different
molecules – nitrogen, oxygen and water. Given the reversible phase changes available to

21
water, most of its maximum entropic energy is made available during condensation as
part of the hydrological cycle. Roderick et al. (2013) have estimated that for a warming
of 2.8 K, the atmospheric content of water would increase from an equivalent liquid
column of 30 mm to 35.9 mm, or 7% per degree K of warming. According to Table 5, by
proportion alone, this would amount to 4,241 J per cubic metre of air at the surface of
extra heat required. However, an exact calculation would need to consider the
diminution of the translational entropy per molecule as a result of its increased
concentration.

Table 5: Summary of total entropy and entropy-temperature terms in the real

atmosphere at 298.15 K
Gas Pressure (atm) St Sr Sv S S/ J/Mole of J per m3
Total mole air/K
STP
H2O 0.00775 185.29 43.74 0.033 229.1 188.6 529.37278 21,637.477187
CO2 0.000397 215.51 54.72 2.99 273.2 213.6 32.337468 1,321.755203
H2S 0.0000000002 338.49 52.52 0.19 391.2 205.5 0.0000023 0.000003
N2O 0.000000325 280.24 59.90 3.05 342.2 218.9 0.0331588 1.355326
O3 0.0000000266 302.13 79.94 1.84 383.9 238.8 0.0032865 0.134332
SO2 3 x 10-10 343.01 84.58 2.875 430.5 248.1 0.0000039 0.000159
NH3 5x10-10 322.23 48.36 0.507 371.1 193.0 0.0000553 0.002260
CH4 0.0000017 272.21 42.26 0.423 314.9 186.0 0.1596086 | 6.523810
CFCl3 0.00000000026 350.00 107.59 32.28 489.9 310.0 0.0000380 0.000006
CF2Cl2 0.00000000055 345.90 107.14 24.945 478.0 300.6 0.0000784 0.000001
CF3Cl 0.0000000001 353.08 99.75 18.627 471.5 285.1 0.0000141 0.000002
O2 0.2095 165.05 43.93 0.035 209.0 205.1 13,054.6468 533,593.023660
CO 0.00000015 279.58 47.19 0.0025 326.8 197.5 0.0146153 0.597382
NO 3x10-10 333.56 48.39 0.0087 382.0 199.6 0.0000034 0.000139
H2 0.0000005 238.11 12.70 - 250.8 130.2 0.0373880 1.528190
N2 0.78084 152.45 41.27 - 193.7 191.7 45,094.8022 1,843,195.930604
A 0.00934 193.70 - - 193.7 154.8 539.400466 0.0442339
Total 2,399,758.372498

In Table 5, this operation is illustrated for the densest surface layer of the atmosphere
only. If the temperature and pressure of the gas is known, a similar calculation can easily
be repeated at all altitudes in the troposphere or the stratosphere and the results integrated
to give the total heat capacity of the atmosphere, a task that others are invited to perform.
For such calculations it is convenient to use suitable computer programs. Fully annotated
outlines of these programs are available on request to the corresponding author.

22
Discussion

Phase space as action space

As shown in the equations above, the mean molecular Gibbs energy can always be
expressed as a variable of the relative action ratio @/ħ alone. Entropic energy (sT) also
includes kinetic energy and the capability for pressure-volume work that are considerd as
enthalpy. This extends to vibrational and electronic states although their contribution to
atmospheric gases near the surface of the Earth at ambient temperatures is usually
relatively small, as shown in the tables. Using the action model, the negative relationship
between free energy and the entropy is more clearly revealed. Paradoxically, “free
energy” is not real energy, but actually denotes its absence; it is better viewed as a
system’s action potential showing its capacity to accept thermal energy by increasing its
action and sustaining field energy, building more complex and diverse structures whilst
doing external work such as expanding the atmosphere or lifting weights in a
gravitational field. To the extent that cooling gravitational work is done that increases the
free energy, heat may re-emerge later if reverse work is done on the molecules of the
system. This reversibility is the essence of the second law of thermodynamics in action
theory. We also consider that referring to this process of increasing entropy as one of
increasing disorder is misleading (Kennedy, 2001). Diversity would be more appropriate
term.

Given that the relative action or mean quantum number @/ħ can be expressed simply as a
function of the particle’s mass, its radial separation and the square root of the temperature
affecting velocity, all of the paradoxes regarding entropy such as its lack of change
during the mixing of equal volumes of identical gases versus the change when two
distinguishable gases are mixed at the same final pressure and temperature are easily
resolved. Each of the distinguishable molecules now occupy twice the space as formerly,
increasing their action accordingly, whereas identical gases must remain in the same
space with the same action as before mixing.

23
In principle, the suggestion to calculate entropy from the logarithm of the translational
and rotational action (including its modification by vibration) is not new. Gibbs identified
the significance of action in his classical text (1902), describing it as the extension in
phase (Vpq), claiming that “the quantity …. which corresponds to entropy is log V, the
quantity V (not volume) being defined as the extension in phase”. So we can conclude
that, according to Gibbs even before Planck identified his quantum of action, any equi-
potential contour in phase space of equal translational action (Vp x Vq = mv x r) would
also correspond to states of equal translational entropy. In effect, changes in the
momentum mv and a linear coordinate r would lead to no change in their product action
and its logarithm, entropy. We can now recognise such contours as adiabatics differing
by a minimum of Planck’s quantum of action h, giving a scale for estimating maximum
uncertainty in momentum or position.

No claim is made here that the action model is inherently more accurate than the classical
methods of statistical mechanics or that the formulae given here for ideal gases apply
without corrections under all conditions of temperature or pressure. But the results are
easily obtained from primary data and are surprisingly accurate, even for vibrational
entropy. This suggests that the action method will have strong heuristic value – not only
for climate science but also for theoretical and experimental purposes in all branches of
chemistry and physics. For example, this action revision of the nature of entropy and free
energy and the interaction between internal and translational action states has the
potential to advance reaction rate theory and many other processes occurring in the liquid
state, including those of life systems. In this area, the translational action will play a
special role, since it is closely related to changes in Gibbs energy for molecular
trajectories from the chemical potential of free reactants through reversibly activated
transition states to the chemical potential of products (see Kennedy, 2001, chapters 4, 6).

Boltzmann’s realistic collision model of entropy

Strongly relevant to the equations for entropy based on logarithmic functions of action
given in this paper is the approach used by Ludwig Boltzmann; he derived an equation
for entropy using his H-theorem by considering the mathematical behavior of a collision

24
integral using a realistic model. This theorem was based on integrating the average effect
on a single molecule of collisions with all the other molecules of the gas, spontaneously
increasing its entropy whenever commencing with a more ordered state. This led him to
essentially the same equations for entropy as those of Gibbs, while claiming “that the
mechanical basis is necessary to illustrate the abstract equations”, despite the current of
opinion at the time from Mach and others directed against the existence of molecules. For
example, Boltzmann (1896) gives an expression for the integral of the sum of the
entropies of the masses in the volume elements as [Rln(ρ-1T3/2) + const.], where ρ is the
number density of gas molecules. We can observe that this result only lacks a suitable
divisor in the logarithmic term required to remove the physical dimensions of action per
unit mass.

In his collision theory, Boltzmann establishes that the quantity H, which can be identified
with the negative of the entropy, must always decrease with time or remain constant – if
one assumes that the velocity distributions of colliding molecules are uncorrelated. H
remains constant (i.e. dH/dt is zero) only when the gas attains a most probable velocity
distribution, corresponding to that derived by Maxwell, a process requiring collisions.
This approach related entropy with probability and Planck’s derivation of the quantum of
action for radiation (1913) acknowledges Boltzmann’s statistical theory of isothermal
entropy (S = klnW). Relating this to the formula for entropy given in this paper, we
conclude that the position of equilibrium where H is constant, must correspond to an
isothermal state where the action is also stationary, although they will both be maximal.

In action theory (Kennedy, 2001) we have extended this realistic model (Fig. 1), with
molecular symmetry reducing the need for field energy because of the shorter free path
for energy between identical molecular structures. Conversely, asymmetry maximizes
the free path length between molecules of field energy and its magnitude in sustaining a
system’s structures. The heat capacity of such diverse systems and their ability or need to
store field energy at a given temperature is therefore greater. For all practical purposes,
the only variable affecting the entropy is the action; more diverse systems will
automatically display greater relative action because of the greater separation of identical

25
molecular structures, but these will also have a richer energy content for a given
temperature.

In general, only differences in action and entropy between alternate states are of
thermodynamic interest since the absolute entropy is not required. We have shown that
for translation the entropy change per mole given a change of state (1=>2) can be
expressed as a function,

ΔSt2-t1 = Rln[(e5/2@t2/ħ)3/zt] - Rln[e5/2(@t1/ħ)3/zt] = 3Rln[(@t2/(@t1)] (10)

assuming that the translational symmetry factor zt remains the same between state 1 and
state 2. Note that the kinetic or enthalpic aspect of entropy drops out. So equation (10)
for the change in entropy gives statistical effects only. Therefore, the precise choices of
the symmetry factor or of the translational radius are only of significance for estimating
the absolute entropy. Because of the statistical nature of momentum and position in
phase or action space, assigning an exact value to the most probable radius or symmetry
factor to each molecule is impossible since they will fluctuate around statistical mean
values. However, the amplitude of the fluctuations is of interest, since these will control
rates of transition.

Under isothermal conditions Gibbs energy varies with translational action and entropy
Given that the third law of thermodynamics states that the entropy at the temperature of
absolute zero is zero, this would require that the action ratio @t/ħ for a gas at this
minimum temperature must be slightly less than one, if it could exist as such, since the
translational entropy can then be considered as equal to 5/2R + Rln[(@t/ħ)3/zt], equivalent
to the Gibbs expression for entropy of ST = H – G where G is the free energy or work
potential of a monatomic gas at constant pressure. This suggests that the magnitude of the
function RTln[(@t/ħ)3/zt] has the same value as the free energy, although opposite in sign,
so that G = -RTln[(@t/ħ)3/zt] or RTln[zt(ħ/@t)3] and ΔG = 3RTln[(@tr/(@tp)] for changes
in action state at constant temperature.

26
Incidentally, the total entropy change during an isothermal chemical reaction at T
includes contributions from changes in translational action and entropy – δ(StT) is equal
to the change in Gibbs chemical potential – as well as changes in the internal action and
entropy representing changes in enthalpy as a result of revised bonding energies. It is
important to understand that the enthalpy term designated H refers to the sensible heat
that tends to change temperature. Thus, if a chemical reaction results in products where
atoms or electrons are more firmly bound with shorter radii, the reduced potential energy
will be compensated by increased internal kinetic energy and equal quantities of emitted
quanta, resulting in a release of heat as a reduction in Gibbs energy and an increase in
entropy and decrease in Gibbs energy of the surrounding system. This change in heat
content as a function of changes in chemical bonding will be reflected in the variation in
zero-point energies of all vibrations possible. In the absence of such chemical reactions,
the enthalpy change can be measured by the changes in kinetic energy and pressure-
volume work alone. This is obviously true with monatomic noble gases like argon that do
not undergo chemical reaction.

It is important to note that, within limits, the internal entropy for rotational and
vibrational states is a function of temperature only. It is unaffected by changes in
concentration, except at very high densities, in contrast to translational states. A low
concentration or pressure corresponds to a high action state of greater entropy. So at
constant temperature, changes in free energy are purely a function of changes in
translational action states, since internal entropy or enthalpy remains constant, or
fluctuate around a stable mean value though variations in internal states by absorption or
emission of radiation and re-equilibration with translational states. This is entirely
consistent with chemical work processes being directly mediated by translational inertia
and pressure, such as pressure-volume work.

Consequently, the free energy or action potential can be considered simply as an inverse
logarithmic function of the translational action and entropy at a given temperature T. The
lower the initial action, the greater the Gibbs energy and action potential, indicating a
spontaneous propensity for thermal energy to be absorbed as latent heat and the action to

27
increase – thus providing the dynamic functional basis for the second law of
thermodynamics. As a measure of the energy capacity of molecular systems, the
translational entropy must also include terms for kinetic energy and pressure-volume
work, indicated by the 2.5R or exponential (e5/2) term. But these, together with the
translational symmetry factor, cancel out for changes at constant temperature.

We can then write that

G = H – ST = H – (StT + SrT + SviT)

The enthalpy (H) is a term always referring to the sensible heat in a system that can be
sensed with a thermometer and related to the kinetic energy of its molecules. The
entropic energy ST differs in that it is only partly indicative of sensible heat, but includes
the potential energy stored in work such as thermodynamic work in molecular systems, or
gravitational work. According to the Carnot principle, in a reversible system this work
can reappear as sensible heat, raising the temperature. This potential source of extra
warming certainly applies to gases in the Earth’s atmosphere. Indeed, it is responsible for
much of the heat transfer to higher latitudes, released by frictional processes on the
Earth’s surface.

For monatomic gases we can rewrite this classic equation taught to all students using the
algorithms developed here as follows.

G = H - ST

-RTln[(@t/ħ)3/zt] = 1.5RT + RT – RTln[e5/2(@t/ħ)3/zt]

For the main diatomic gases in the atmosphere nitrogen and oxygen at ambient
temperatures by including rotational entropy but where we can neglect vibrational
entropy, we will have the following.

28
-RTln{[(@t/ħ)3Qe/zt][(@r/ħ)2/σr]} = 2.5RT + RT – RTln{[e7/2(@t/ħ)3Qe/zt] [(@r/ħ)2/σr]}

Disallowing a role for the enthalpy of chemical reactions at ambient temperatures in the
troposphere we have,

G = H – ST = 3.5RT - ST

Alternatively, we can write for Helmholtz energy in constant volume conditions a

modified equation, varying only slightly in the RT or PV term.

-RTln{[e(@t/ħ)3Qe/zt][(@r/ħ)2/σr]} = 2.5RT – RTln{[e5/2(@t/ħ)3Qe/zt] [(@r/ħ)2/σr]}

A = E – ST = 2.5RT – ST = G - RT

For polyatomic molecules or at temperatures where vibrational entropy and energy are
more relevant, it is simple to add the vibrational entropy terms to both sides of the
equation.

Physicists and chemists have preferred to consolidate the different kinds of entropy into
the single term S, perhaps for simplicity. They have then relied on differential calculus to
establish thermodynamic relationships. However, this has partly concealed its true nature,
obscuring the relationship between the partitioned kinds of action. Furthermore, action
thermodynamics proposes that the external action (translation) and the internal actions
(rotation, vibration, electronic orbital) defining the respective translational and internal
free energies or entropies can sometimes be considered as physically opposed [5] during
collisions and able to vary independently. Since these can now be calculated more easily
and integrated numerically if required, the need for solutions based on calculus or
differential equations is diminished.

In his engaging book on statistical mechanics, Schrödinger (1946, p. 52) derived the
relationship for the Gibbs potential as nklnζ = U + PV – TS, by calculating from the ‘sum

29
over states’ ΣNi/N = e-ε1/kT + e-ε2/kT + e-ε3/kT+ e- ε4/kT ……+ e-εn/kT……… ; nkTlnζ is the
thermodynamic potential (or free energy) for n molecules, a function of an inversed
action ratio ζ. He defined the factor 1/ζ as a function of the translational partition
function (2πmkT/h2)3/2V, divided by the number of particles (n) – that is, as a translational
action ratio as defined in this paper. By contrast its inverse ζ is an ‘inaction’ ratio
indicating the free energy. For a perfect monatomic gas, PV is equal to RT and so U + PV
is equal to the enthalpy H, which does not change for individual molecules of a chemical
species – unless the temperature changes.

In contrast to translation and rotation, vibrational action states higher than the ground
state are largely unoccupied at ambient temperatures; most greenhouse molecules in the
atmosphere are still in their coldest vibrational states, despite them radiating as required
by the Stefan-Boltzmann equation proportional to the fourth power of the temperature in
Kelvin. Thus vibrational action and entropy are minimal in Earth’s atmosphere. This
ordered state of low vibrational entropy is fortunate for life on Earth, otherwise stable
molecules and structures would be impossible. In kinetic theory it has usually been
assumed that molecular trajectories are linear, with no interaction between molecules.
But the linear model is clearly only an approximation for translation. Whether the
translational trajectory of the molecules is considered as curved or straight is irrelevant,
given that the speed of energy transfer vastly exceeds that of the molecules; relatively to
the thermal field energy bath referred to by Clausius in 1875, molecules are almost
stationary.

It should be understood that all of the thermodynamic changes considered in this paper
refer to changes possible in the physical state of small molecules in the atmosphere,
mainly under conditions at the Earth’s surface or using standard reference states. On
Earth, only in the case of water do enthalpic changes associated with melting and
vaporisation also need to be considered. Nor has attention been paid to changes in state
following chemical reactions. In the case of chemical reactions the enthalpic changes
related to internal action and entropy reflecting changes in bond energies either releasing

30
or absorbing heat must also be included in the enthalpy as discussed above. However,
chemical reactions will not be considered here.

Such reactions occurring at constant temperature require transfer of heat across the
system’s boundary, resulting in entropy changes external to the system. Action
thermodynamics applied to transition states in chemical reactions will be considered in a
separate paper. It is important to understand that chemistry at the Earth’s surface is
dependent on its physical environment. The phase changes discussed in this paper are
strongly influenced by the thermodynamic and the gravitational properties of the milieu
in which these changes occur.

Greenhouse gases and temperature equilibration in the gravitational field

The entropy calculations in this paper for gases have been considered for equilibrium
conditions. This assumes that temperatures in the atmosphere equilibrate with surface
insolation instantaneously. This is clearly not the case and the atmosphere will exhibit
non-equilibrium conditions, constantly responding to the changing intensity of radiation
from the Sun. This suggests that a key role of greenhouse gases may be their function in
facilitating heat transfer by radiation and convection from the surface, cooling it and also
in re-radiation from within the atmospheric profile, though consistent with its local
temperature.

When atmospheric molecules are heated internally by absorption of infrared radiation,

increasing their vibrational or rotational action and entropy, translational modes of action
will immediately respond mediated through subsequent collisions. Relaxation times for
such excited states are of the order of 10-7 sec. The absorption of a quantum will decrease
their internal free energy whilst increasing their inertia and capacity to exert pressure,
potentially doing gravitational work while moving to higher altitude and thus lowering
the local temperature as kinetic energy declines. This may seem paradoxical but it is
consistent with the virial theorem whereby increases in gravitational potential energy are
matched by decreases in half its magnitude of kinetic energy (Kennedy, 2001, chapter 5

31
and Endnotes) and this idea will developed further in a companion paper on the
barometric formula in assessing the possible role of greenhouse gases in global warming.

Such dissipation processes for absorption and emission of radiant energy may give a
special role for greenhouse gases in the atmosphere, since the major gases nitrogen and
oxygen have little if any such absorptive activity. Their presence enhances the rate of
transfer of radiant energy from the Earth’s surface to higher altitudes. Indeed, this could
be the most important role of polyatomic gases like carbon dioxide. On the whole,
greenhouse gases are regarded negatively because of their proposed role as agents in
global warming; but it is important to also consider for experimental testing their possible
benefits, such as elevating the atmosphere and cooling the surface of the Earth. Gases
with higher heat capacities (including nitrogen and oxygen) also tend to cause the
atmosphere to be more elevated, because the temperature lapse rate with altitude is less
than for monatomic gases of similar mass. Thus an atmosphere of carbon dioxide of
mass 44 Daltons would be more elevated than one of argon of mass 40, despite its greater
weight.

We can examine the relative absorptivity of the greenhouse gases and the existence of
absorptive-emissive lines in the infrared (Tables 3, 4), recalling that the earth’s surface
has a maximum emission range of around 5-30 μm wavelength (10000 cm-1=1 μm, 1000
cm-1=10 μm; 100 cm-1 = 100 μm) whereas sunlight is confined to the 0.3-5 μm range.
The longer wavelength of terrestrial radiation compared to sunlight is a result of
absorption of sunlight by surface materials and re-equilibration of the quanta with the
much cooler surface of the Earth, compared to the boiling ocean of hydrogen atoms of the
Sun. Obviously, polyatomic molecules absorb in the 5-30 μm wavelength band of the
infrared and the more complex the molecules are, the greater the number of absorptions.

The quanta associated with changes in rotational and translational action must be of
longer wavelength, in the microwave and radiowave range of frequencies not resonant
with the earth’s major energy primary emissions from sunlight. But the infrared radiation
absorbed by greenhouse molecules will be converted to these lower frequencies as a

32
result of work done in subsequent molecular collisions during temperature equilibration
in the atmosphere – the process known as equipartition. It is of interest that the quanta
able to promote equality of kinetic temperature with equilibrated molecules range from
infrared for vibrational freedom to microwaves for rotation and radiowaves for
translation, thus broadening the spectrum of the energy involved.

The theory of quanta as enunciated by Planck (1913) pointed out that the intensity of
black body radiation within an evacuated space equilibrated with matter bounding this
space was entirely a function of temperature – despite this material particle being
restricted to just one speck of carbon at the centre. Therefore, even a space containing
mainly non-greenhouse gases such as nitrogen and oxygen should still radiate with the
same intensity of infrared radiation, an essential feature for the correct operation of
equipartition. Conversely, the quantum theory demands that very few (but not none) of
the nitrogen and oxygen molecules will be excited vibrationally by such radiation,
compared to greenhouse molecules. Furthermore, the intensity of the radiation passing
through the boundary of a parcel of air is not a function of the proportion of the
molecules able to be excited, or even their total number per unit volume, but only on the
absolute temperature of the equilibrated surface. In a space relatively free of greenhouse
gas molecules at a given temperature, the rays of infrared quanta will travel further before
absorption so the total heat capacity for such quanta will be less. However, equilibrium
assumes that the radiation density detected and radiated through the boundaries will
always satisfy Stefan-Boltzmann theory for surfaces emitting radiation.

According to Clausius (1875) and the second law, to heat the Earth’s surface as a net
process the atmosphere would need to be hotter than the surface. Consistent with this
principle, most of the temperature increase at the surface of the Earth from energy fed
back from the atmosphere must be a result of the reversal of convective processes in high
pressure zones, when air is descending. The fall of atmospheric gases from higher
gravitational energy is a work process generating heat, all air molecules simultaneously
gaining kinetic energy and radiating equivalent heat quanta as required by the virial
theorem of Clausius. Clearly, this process can heat the surface, as occurs in high pressure

33
zones or anticyclones. However, this transfer of heat from the atmosphere must be
balanced by compensating transfers of radiant heat into the atmosphere in low pressure
zones as gravitational work is performed using heat. These reversible processes
demonstrate the Carnot principle that so impressed Clausius. Surprisingly, in climate
science little attention is paid to the reversible transfers between heat and work that are
implied in Lagrange’s earlier identity relating the second derivative of the inertia of a
system of particles (I=Σmr2) with respect to time and its kinetic (T) and potential energy
(V); these can be considered as surrogates for heat and work in a gravitational or central
force system.

½d2I/dt2 = 2T + V (11)

On the basis of reversibility so that the equality of equation (11) is zero, so 2T = -V,
inertial effects can be considered as sources of heating or cooling as seen in convection
and advection near the Earth’s surface (Kennedy, 2001, chapter 5). In fact, this equation
could be considered as a basis for the whole of climatology, a contention we will explore
elsewhere.

Table 4 also shows the specific frequencies of infrared radiation from the Earth that
different gases such as CO2, N2O and CH4 will absorb. But a CO2 molecule activated by
IR-absorption to vibrate more vigorously will transfer most of this energy to other air
molecules in the next collisions, thus increasing their action and entropy while dissipating
the activated internal state and increasing their Gibbs energy. The relaxation time for
such vibrational and rotational excitations is about 10-7 sec, given molecular collision
rates of 109 – 1010 sec-1, during which translational energies of all air are in increased by
this dissipation. As such, quanta from the Earth’s surface will not be exchanged between
air molecules. Nevertheless, new infra-red and microwaves quanta will be generated
appropriate for the reduced temperature with altitude.

Furthermore, the more dilute the gas (e.g. N2O and CH4), the greater its translational
entropy per molecule – although its vibrational and rotational entropies will be purely a

34
function of temperature. Thus, on absorbing a specific quantum of IR-radiation (exciting
molecular vibration) such a dilute gas will have a larger disequilibrium between its
vibrational action and its translational action. In a subsequent collision, the greater
inertia and amplitude of the vibrating atom should cause a more efficient transfer of
momentum to surrounding air molecules, irrespective of whether they are greenhouse
gases or not.

So this thermodynamic force and heating effect is transferred to N2 and O2 as a result of

collisions and the heated gases expand to higher altitude, exchanging their increased
kinetic energy for increased gravitational energy and cooling as a result. Perhaps it is
more apt to consider that the greenhouse gases such as water play an important role in
holding up the sky, enabling reversible gravitational work, thereby cooling the
atmosphere!

Adiabatic processes
The concept of reversible adiabatic changes in the atmosphere is often invoked in
connection with parcels of air moving by convection or advection – with no heat entering
or leaving the parcel. As a result, any changes in the thermodynamic state of the air
parcel, such as temperature or pressure changes, must be achieved with the constraint of
no transfer of heat into or out of the parcel. In effect, this requires that these changes be
isoentropic, according to Clausius’ definition of entropy (1875).

dS = ∫dQ/T

As a result, given the case for the relationship between action and entropy made
preceding, there must also be no change in the total relative action. Since translational
action is given by (3kTIt)1/2, this requires that any adiabatic change in temperature T must
be accompanied by a compensating change in the inertia It. This is effectively achieved as
a change in volume or density of the gas. For example, in a reversible Carnot cycle heat
engine with a source temperature Tsource and a sink temperature of Tsink , a monatomic
working fluid like helium or argon would obey the relationship

35
 (Tsink/Tsource)3/2 = V2/V3

Here V2 => V3 represents an isoentropic process in which external work is done at the
expense of the heat content of the working fluid, its temperature falling from Tsource to
Tsink and the volume increasing from V2 to V3. Some of the heat content of the parcel as
measured by molecular temperature has apparently disappeared as the work of reversibly
separating the molecules.

Adiabatic cooling during expansion of a working fluid while external work is being done
is a feature of the Carnot cycle, with no heat flowing into or out of the engine. If no work
is being done, such as expansion of a gas into a vacuum (previously prepared), no cooling
will occur. Thus, an adiabatic expansion process per se is not the cause of the lowered
temperature with altitude. The cooling effect called the adiabatic lapse rate for parcels of
air expanding into a lower pressure zone higher in the atmosphere is a result of
gravitational work being done. A subsequent paper will provide an alternative view of the
lapse rate as a function of the virial theorem that increased gravitational potential of the
atmosphere corresponds to a decrease in kinetic energy of half this magnitude.

We can write, for the isoentropic change per mole of monatomic gas, given the heat
capacity at constant volume Cv of a monatomic gas is 1.5R per oK

ΔS = Rln(@3/@2)3 = Rln(3kT3I3/3kT2I2)3/2
= Rln(T3r32/T2r22)3/2 = Cvln(Tsink/Tsource) + Rln(V3/V2) (12)

Given (Tsink/Tsource)3/2 = V2/V3, the change in entropy must be zero. This follows since the
effect of the decrease in temperature on action just balances the effect of the increase in
volume. So if the increase in volume (or decrease in pressure at constant temperature) is
10 times greater, (Tsink/Tsource)3/2 is equal to 0.1 and the ratio Tsink/Tsource is equal to
0.2154467.

36
In general, we can write that for Carnot cycles, Tsink/Tsource is equal to (V2/V3)R/Cv for
working fluids of greater heat capacity. This means that diatomic gases like N2 and O2
would have a smaller temperature range for a 10-fold increase in volume, with Tsink/Tsource
of 0.40. This also would translate to the atmosphere as a lower change in temperature
with altitude than with the monatomic argon.

These adiabatic changes considered above occur at constant gravitational potential, as

when a parcel of air moves laterally by advection, doing electrical work on a wind farm.
When a parcel of air moves reversibly by adiabatic convection to a higher gravitational
potential we have to consider the cooling effect of doing gravitational work in addition to
changes in the atmospheric pressure. The lower the pressure exerted by the weight of the
atmosphere above the parcel of air, the less pressure-volume work and heat is needed for
expansion. However, almost the same amount of heat is required to raise the gravitational
potential of air molecules no matter what the altitude. This variation in pressure-volume
work brings into question the use of the dry adiabatic lapse rate (9.8 C/km) at all
altitudes.

A descending parcel of air may be adiabatically compressed and spontaneously heats as

gravitational potential declines causing kinetic work and internal heat-work varying free
energy to be done on the air as it falls. We will show elsewhere that the increase in
kinetic heat shown by the temperature increase at the expense of gravitational potential
energy is matched by the decrease in Gibbs energy of the thermodynamic field, also
consistent with the virial theorem. Furthermore, the capacity to do work of the air parcel
declines as the atmospheric pressure increases and pressure-volume work becomes more
costly.

As appropriate for statistical thermodynamics, these calculations of entropy and free

energy relate only to the scale of randomized molecular motions of canonical ensembles.
Neither the kinetic energy nor the “work-heat” or potential energy involved in convective
and advective motions of parcels of air has been considered here. The thermal energy
required to initiate these higher order motions (i.e. neither vibrational, rotational nor

37
translational) is substantial, even though the kinetic energy generated is relatively minor
compared to that of the randomized molecular motions. But the “work-heat” required for
anticyclones and cyclones generated by thermal gradients in the gravitational field is
substantial. The observation here that the major part of the heat required per molecule
(sT) from absolute zero to 298 K relates to the “work-heat” compared to the sensible
kinetic heat is a striking observation rarely made. For example, for argon the total
entropy (18.6k per molecule or 154 J/C/mole), shown in Table 4 at 1 atm, is 12.4 times
the increase in entropy from 0 K for kinetic motion alone (1.5k). At 0.01 atm in the
atmosphere, the ratio is even greater. For all the molecules in a rotating parcel of air, the
potential energy or “work-heat” of motion in these coherent though gaseous “fly-wheels”
is orders of magnitude greater than the kinetic energy of their circulation. Even though
the dissipation of this “work-heat” as frictional heat at higher latitudes is a major
mechanism for the dispersal of solar energy from the equator towards the poles, this
source of warming is rarely properly considered in climate models.

Conclusion

Unfortunately, in recent years thermodynamics and statistical mechanics have largely

fallen into disuse, except by specialists. However, this should not continue, given the ease
of calculation of entropy and of free energy of gases from action displayed in this paper.
We will show elsewhere the utility of this approach. Action thermodynamics provides a
realistic modeling approach available to all, easily applied mathematically and
simplifying study of the links between heat and work and morphogenesis. We
recommend more widespread application of this more explanatory approach that so aptly
partitions the enthalpic and statistical aspects of entropy.

Acknowledgements

We wish to give credit to colleagues, including Barry Noller, Rodney Roughley, John
Knight and many others. As thought alone, this work required no funding.

38
References

Aylward G. and Findlay T. (1974), SI Chemical Data. 2nd Edition, University of

New South Wales Press, Kensington NSW, Australia

Brown, Arthur (1968) Statistical Physics. The University Press, Edinburgh.

Boltzmann, L. (1896) in J.A. Barth, Lectures in Gas Theory. Translated by Stephen G.

Brush, p. 147, 1964, Dover Publications, New York.

Clausius, Rudolf (1875) The Mechanical Theory of Heat. Translated by Walter R.

Browne, 1879 MacMillan and Co., London.

Glasstone, S. (1951) Thermodynamics for Chemists. 2nd Edition, Macmillan and

Company, London.

Gibbs, Josiah Willard (1902) Elementary Principles in Statistical Mechanics. Charles

Scribner’s Sons, New York.
Herzberg, G. (1945) Molecular Spectra and Structure Vol. II Infrared and Raman
spectra. p. 5, Van Nostrand, Reinhold and Company, New York.

Herzberg, G. (1950) Molecular Spectra and Structure Vol. I Spectra of diatomic

molecules, 2nd Edition, Van Nostrand, New York.

Hill, T.L. An Introduction to Statistical Thermodynamics. Dover Publications, 1960, p.

157, New York.

Kennedy, I.R. (2000) Action as a dynamic property of the genotype x environment

interaction: Implications for biotechnology. Acta Biotechnologica 20, 351-368.

Kennedy, I.R. Action in Ecosystems: Biothermodynamics for Sustainability. Research

Studies Press, 251 pages, Baldock, UK, 2001.

39
Moore, W.J. (1963) Physical Chemistry. 4th edition, Longmans London.

Planck, M. (1913) The Theory of Heat Radiation. Dover Publications 1959 New
York.

Rose, M.T., Crossan, A.N. and Kennedy, I.R. (2008) Sustaining action and
optimising entropy: coupling efficiency for energy and the sustainability of
global ecosystems. Bulletin of Science, Technology and Society, 28, 260-272.

Roderick, M.L., Sun, F., Lim, W.H. and Farquhar, G.D. (2013) A general
framework for understanding the response of the water cycle to global warming
over land and ocean. Hydrological Earth Science Stem Science 10, 15263-15294.

Shannon, Claude E. and Weaver, Warren (1949) The Mathematical Theory of

Communication. (Reprinted in 1972) University of Illinois Press, Urbana.

Schrödinger, Erwin (1946) Statistical Thermodynamics, Reproduced by Dover

Publications Inc. 1989, New York.

Wilson, E.B. Jnr (1940) Chemical Reviews 27, 17.

40
Figure 1: Calculation of translational and rotational action (@). Mean translational
action @t (A) is estimated as explained in the text from average separation of a = 2r by
allocating each molecule space of a3 = V/N where V is total volume and N is total
number of diatomic molecules like dinitrogen (N2). Relative angular motion dӨ/dt = ω is
estimated for molecules exhibiting the root-mean-square velocity, taking 3kT = mv2 =
mr2ω2. Then translational action is equal to (3kTIt)1/2. Rotational action @r (B) for linear
molecules such as N2, O2 and CO2 is estimated similarly, equated to (2kTIr)1/2.
.

41
 Compute molecular
Compute entropy andentropy
free
and
energy
free energy
Inputs
Pressure
Temperature
Mass of each atom
Bond length (s) for linear molecules
Moments of inertia for non-linear
Vibrational frequencies & degeneracy

Find translational entropy

If monatomic Goto Outputs

If linear
find rotational
moment of
inertia
If non-linear,
use Ia, Ib, Ic
Find rotational entropy

Find vibrational entropy

Outputs
Total entropy ΣS
Entropic energy ST
Gibbs energy G

Figure 2: Flow diagram for computing absolute entropy and Gibbs energy. A fully annotated
description of the relevant algorithms and subroutines to compute entropy and free energy is
available on request to the corresponding author.

42