International Advance Journal of Engineering Research (IAJER)

Volume 3, Issue 1 (January- 2020), PP 01-06

ISSN: 2360-819X
www.iajer.com

Research Paper Open Access

A New Constant from Classical Nucleation Theory

John H. Jennings
Jennings Research & Editing 2530 Hillegass Ave. #307 Berkeley, CA 94704 USA
*Corresponding Author: John H. Jennings

ABSTRACT:- Blander and Katz give a formula in classical nucleation theory, J = A exp K, for homogeneous
nucleation (liquid-->gas). Jennings proved that dlnA/dK = 1/6K for all pure liquids by combining two theories,
taking the limit as polymer concentration-->0. This gives lnA = (1/12)ln(K2) + C, where C is the integration
constant. The conjecture is that C is a constant for fluids of low molecular weight. We used data for 10 sample
solvents, and solved for C. The surface tension drops out in C, which makes C more accurate, as the surface
tension is difficult to get at 0.89Tc, the limit of superheat. Tc = critical point in Kelvin. All quantities are
evaluated at the limit of superheat, which is approximately 0.89Tc for solvents. C = 75.379 ± 1.073 for the 10
solvents (a range of polar to non-polar). This eliminates the prefactor A, streamlining J: ln J = (1/12)ln(K2) +
75.379 + K is the exact new equation. Using information from Blander and Katz, it is possible to get an exact
value for K. K = - 64.5605.

KEY WORDS:- “homogeneous nucleation” “Flory-Huggins theory” “limit of superheat” “differential

equation” “polymer solutions”

I. INTRODUCTION
In the late 1800s, Josiah Willard Gibbs had the idea that there's a trade-off between lowering energy
and maximizing entropy where clusters of a new phase appear driven by the increase in temperature to
overcome the barrier to forming a new phase. In 1942, Flory and Huggins made a theory for the mixing of
polymer and solvent based on a lattice model. This paper shows how they join in a new way.

Over the years since Gibbs, homogeneous nucleation has been studied and developed into an exact
theory that gives a well-defined nucleation rate with a prefactor multiplied by an exponential term. The
prefactor slowly varies with rising temperature and in this paper, the author presents an exact derivation
supported by data that allows for elimination of the prefactor.

THEORY
Later on, these two trains of thought developed into 1) modern classical nucleation theory (CNT) as put
forth by Blander and Katz 1975 and 2) the model for surface tension of polymer solutions (STPS) refined by
Siow and

Patterson 1973. In 2012, Jennings combined the CNT/STPS equations and that later led to equation (19) in
Jennings 2014, here as (1), a general formula for bubble nucleation in polymer solutions. The data for (1) was
first presented in graphical form in Jennings and Middleman 1985.

ΔT = 3kT2wMW1/σaMW2 (1)

Blander and Katz’s (15) is the abbreviated formula (2) here for the nucleation rate,

J, discussed in Appendix 1.

J = A exp K bubbles/cc-sec (2)

Jennings 2012, proved in (11) there, essentially that

dlnA/dK = 1/6K (3)

Engineering Journal www.iajer.com Page | 1

 A New Constant from Classical Nucleation Theory

for all pure liquids by combining the CNT/SPTS theories and then taking the limit as polymer concentration--
>0. See Appendix 2 for an outline of the 2012 proof for Eq.

(3), which is the precursor to Eq. (1). The solution to Eq. (3) is

lnA = (1/12)ln(K2)+ C (4)

where C is the integration constant. Notice K is squared because K is a negative number. C is a pure number,
the same for 10 fluids, as equation (3) is general.

II. METHODOLOGY
In the early 1980s, Jennings and Middleman collected data on liquid---->gas nucleation as affected by presence
of polymer. In 2012, Jennings made a foray into a theoretical treatment that predicts the early 1980s data quite
well. In Jennings’ 2012 paper is contained the starting equation for this work, more clearly laid out in

Appendix 2. This is Eq. (3), a truly exact formula, which has a well-defined integration constant, putting the
whole theory on solid ground.

The pre-exponential factor, PEF, has been studied and was derived years ago. There is a paper by Shiau 2018,
where the temperature dependence of the PEF was investigated. Eq. (3) is the precursor to Eq. (1). Bovey and
Winslow 1979 give an exact equation for boiling point elevation due to addition of polymer, where the heat of
vaporization appears in denominator instead of the surface tension, but otherwise that equation is similar to Eq.
(18) in Jennings 2012.

A preliminary study to this effort was published by Jennings 2019 with seven solvents and this paper has more
solvents and brings classical nucleation theory to its logical conclusion and includes the clearer derivation of the
starting Eq. (3) in

APPENDIX 2. The author feels that the additional solvents and results justify a new paper and this is of
interest in the field of thermodynamics and chemical engineering.

III. RESULTS AND DISCUSSION

The reason for undertaking this work was when we noticed that the integration constant varied little
among different solvents. This is interesting because: 1) equation (3) is mathematically correct and exact for
pure liquids and 2) the integration constant C should be a universal constant as long as BLANDER/KATZ’s
equation (2) holds. This is true as long as the Poynting correction is valid, that is, the vapor pressure of the
solvent is appreciably greater than the ambient pressure. Remember, equation (2) is evaluated only at the limit
of superheat for all quantities because Blander/Katz’s formula is to give the limit of superheat for liquids The
novelty of this study is that a strange formula, equation (4), gives a novel new constant in physical chemistry,
based on (15) in BLANDER/KATZ’s highly cited paper.

C was then evaluated with data at the limit of superheat for ten common solvents, inserting the data in (4) and
solving for C. The surface tension cancels out in (4) and this is fortunate, as the surface tension is hard to
estimate at 0.89 of the critical temperature, where the limit of superheat is.

Data
Here is a table of the calculations for C from July 27, 2019 until January 8, 2020.

Solvent C (integration δ(SI) solubility MW gm/mol

(low MW) constant) Hildebrand molecular weight
heptane 74.10 15.3 100.21
hexane 74.60 14.9 86.18
carbon
tetrachloride 74.79 18.0 153.82
cyclohexane 74.82 16.8 84.16
pentane 74.84 14.4 72.15
cyclopentane 75.11 16.2 70.14
benzene 75.14 18.7 78.12

Engineering Journal www.iajer.com Page | 2

 A New Constant from Classical Nucleation Theory

ethanol 75.91 26.2 46.07

methanol 76.54 29.7 32.04
water 77.94 48.0 18.02

The higher the solubility parameter is, there is a trend to higher C. The lower the MW is, there is a trend to
higher C.

C = 75.379 ± 1.073

Anyway, the results above show that C = constant through the range of sample solvents going from non-polar
pentane to polar water. Notice there is little scatter. Some of the solvents are alkanes along with a variety up to
polar solvents. It appears that C is a constant for low MW solvents. For Blander and Katz’s formula in
APPENDIX 1 and APPENDIX 2, this enables the prefactor A to be eliminated.

ln J becomes simply:

ln J = (1/12)ln(K2) + C + K (5)

See Appendix 1 for Blander and Katz's exact expression for J and details on the workup of the data. See
Appendix 2 for derivation of Eq. (3), the starting equation.

For Eq. (5), the nucleation rate J is solely a function of K. Using a CASIO fx-115ES scientific calculator
iterating with C = 75.379 and Blander and Katz’s value (for nucleation of J = 104 – 106) the author took the
geometric mean for J, and obtained equation (6). Getting the value for T (Kelvin) yields the temperature of
liquid--> gas nucleation, which obviously needs a sophisticated computer program.

1830.84 = T (Pe - PL)2 δ2 / σ3 (6)

Finally, Blander and Katz say when the value of J changes by a factor of 1,000 to 10,000 the prediction of the
limit of superheat changes only one degree Celsius.

So, the uncertainty in C is ± 1, or only one power in the natural logarithm of J, a factor of e, a minuscule
uncertainty. K = - 64.5605 for the ten solvents.

IV. CONCLUSIONS
These results bear out that C is a strange new universal constant, which should be true for solvents of
low molecular weight. Again, it needs to be borne in mind that the surface tension is hard to determine up at the
limit of superheat and the surface tension conveniently drops out! This study is not concerned with direct
calculation of the nucleation rate, but it is shown here that the pre factor in classical nucleation theory for liquid-
-->gas has been eliminated and K has been evaluated exactly for all solvents.

NOMENCLATURE
a surface area of solvent molecule
A prefactor
B coefficient
C constant of integration
d density of liquid
J nucleation rate
k Boltzmann constant
K exponent
M molecular weight of liquid
MWi molecular weight: solvent 1, polymer 2
Pe equilibrium vapor pressure
PL ambient pressure
PV vapor pressure

T temperature Kelvin
Tc critical temperature in Kelvin

Engineering Journal www.iajer.com Page | 3

 A New Constant from Classical Nucleation Theory

w weight fraction polymer

δ Poynting correction factor
ΔT rise in superheat in Centigrade
σ surface tension

ACKNOWLEDGMENTS
This article is adapted from Jennings, J. H. (2019). Eliminating the Pre-exponential

Factor in Classical Nucleation Theory. Chemical Science International Journal, 28(3),

1-6. https://doi.org/10.9734/CSJI/2019/v28i330140

APPENDIX 1
1) J = A exp K
J ≈ 3.73 x 1035 (d2 σ / M3 B)1/2 exp [ - 1.182 x105 σ3 / (T (PV - PL)2)]

A is the prefactor and K is the exponent, according

to this detailed equation (15) in Blander/Katz.
2) There is a Poynting correction, δ, where

δ ≈ ( PV - PL) / ( Pe - PL)

B is close to 2/3. δ and B are discussed in Blander/Katz.

3) In the detailed expression for J (bubbles/cc-sec) above, the other units are as follows: Pe (equilibrium vapor
pressure), PV (vapor pressure) and PL(ambient pressure) are all in atmospheres; T is the limit of superheat for
the solvent (Kelvin); σ is the surface tension (dynes/cm); in the pre factor, M is the molecular weight of solvent
(g/mole); and d is its density (g/cc). All quantities are at the limit of superheat for each solvent.

4) Sources for the data were CRC Handbook,

JASPER, ANTOINE EQUATION (Iran Website),

BLANDER/KATZ and various Internet websites for densities. However, the density of cyclopentane was
estimated and also the limit of superheat of carbon tetrachloride was estimated. Admittedly, this is a bit
imprecise, but the mathematics dictates that C is a pure number.

APPENDIX 2
Proof of dlnA/dK = 1/(6K) for all pure liquids J = A exp K from Blander and Katz classical nucleation
theory liquid→gas.

The starting equations are from

1) Blander and Katz (Z) and 2) Siow and Patterson (A) and (B).

See JH Jennings, International Journal of Thermodynamics article, Ref. (10). page 127-128.

35 2 3 1/2 5 3 2
J ≈ 3.73 x 10 (d σ / M B) exp [ - 1.182 x10 σ / (T (PV – PL) )] (Z)

(σ - σ1) a / kT = ln (φ1S / φ1) + ((r -1) / r ) (φ2S - φ2) (A)

ln[(φ2S / φ2)1/ r / ( φ1S / φ1)] = (σ1 - σ2) a / kT (B)

Now, near φ2 = 0, Eq. (B) becomes

φ2S = φ2 exp [ r (σ1 - σ2) a / kT ]

Putting in the numbers, ∂φ2S/∂φ2 ≈ 10-38 for MW2 = 2000, r = 13.4 and

Engineering Journal www.iajer.com Page | 4

 A New Constant from Classical Nucleation Theory

even less for higher MW.

First, it is necessary to prove Eq. (11) in Ref. (10).

lim w2-->0 (∂lnA/∂w2)/(∂K/∂w2) = 1/(6K)

35 2 3 1/2
A = 3.73 x 10 (d σ / M B)

Omitting a few steps, because they are obvious and constants, we have

∂lnA/∂w2 = (1/d) (∂d/∂w2) + (1/2σ) (∂σ/∂w2) - (1/2B ) (∂B/∂w2)

In calculating ∂K/∂w2, we note that ∂T/∂w2 = 0, as T and w2 are orthogonal.

PL is ambient pressure and therefore is constant; δ is the Poynting correction factor.

5 3 2 5 3 2
K = - 1.182 x10 σ / (T (PV – PL) ) = - 1.182 x10 σ / (T (Pe – PL) δ2 )

5 2 2
∂K/∂w2 = - 1.182 x10 (3)σ (∂σ/∂w2) / (T (Pe – PL) δ2 ) +

5 3 3
1.182 x10 (2) σ (∂Pe/∂w2) / (T (Pe – PL) δ2 ) +

5 3 2
1.182 x10 (2) σ (∂δ/∂w2) / (T (Pe – PL) δ3 )

We prove these four differentials are zero, which simplifies it, page 128 of Ref. (10).

∂δ/∂φ2 = 0, ∂d/∂φ2 = 0, ∂Pe/∂φ2 = 0, and ∂B/∂φ2 = 0 for w2 near 0.

1. d = d1 + (d2 – d1) φ2S

∂d/∂φ2 = (d2 – d1) (∂φ2S/∂φ2), which vanishes for w2 near 0.

2. Pe = Pe(0) φ1S = Pe(0) (1 - φ2S)

One can see by inspection that here ∂Pe/∂φ2 also vanishes.

3. B ≈ 1 - 1/3 (1- PL/PV) Here the approximation Pe = Pv is used because

this is a very small correction and they are close. Hence,

∂B/∂φ2 = (-1/3) (-1) ∂/∂φ2 PL/Pe

= (1/3) PL ∂/∂φ2 1/Pe and this vanishes too.

4. δ = 1 - dG /d + 0.5 (dG / d)2 and dG = Pe MW1/RTl,

ideal gas and Tl and φ2 are orthogonal. With a little algebra and using the previous

results it is readily seen that ∂δ/∂φ2 also vanishes.

Next, examining Eqs. (16) and (18) on page 129 we realize that:

∂φ2 = (d1/d2) ∂w2

Engineering Journal www.iajer.com Page | 5

 A New Constant from Classical Nucleation Theory

The solvent and polymer are incompressible, so d1 and d2 are taken as constant.

Finally, using the fact that the four differential quantities are zero gives:

lim w2→0 ∂lnA/∂w2 = (1/2σ) (∂σ/∂w2) and simultaneously,

5 2 2
lim w2→0 ∂K/∂w2 = - 1.182 x10 (3)σ (∂σ/∂w2) / (T (Pe – PL) δ2 )

Therefore, their ratio becomes:

lim w2→0 (∂lnA/∂w2)/(∂K/∂w2) = 1/(6K)

This is all only as polymer concentration approaches zero, or pure liquid.

So, we have the starting equation with an integration constant that was

calculated for seven different low molecular weight solvents.

dlnA/dK = 1/(6K) [nomenclature omitted]

REFERENCES
[1]. CRC Handbook for Chemistry and Physics 66th Ed.
[2]. Iran website for ANTOINE EQUATION CONSTANTS
http://teachers.iauo.ac.ir/images/Uploaded_files/ANTOINE_COEFFICIENTS_FOR_VAPOR_PRESS
URE%5B1%5D%5B4619259%5D.PDF
[3]. Various Internet websites for solvent densities
[4]. Blander M., Katz J.L., Bubble Nucleation in Liquids, AIChE J., 21(5):833-848 (1975)
[5]. Bovey, F.A. and Winslow, F.H., MACROMOLECULES: An Introduction to Polymer Science,
Academic Press, (1979)
[6]. Jasper J.J., The Surface Tension of Pure Liquid Compounds, J Phys Chem Ref Data,
[7]. 1: 841-1009 (1972).
[8]. Jennings, J.H., Eliminating the Pre-exponential Factor in Classical Nucleation Theory, Chemical
Science International. Journal, 28(3): 1-6 (2019).
[9]. Jennings J.H., Homogeneous Nucleation from Polymer Solutions, Polymers Research J., 8(4):311-
319 (2014).
[10]. Jennings J.H., Limit of Superheat of Polystyrene-Cyclohexane Solutions: Theory, Int. J.
Thermodynamics, 15(3):127-132 (2012).
[11]. Jennings J.H., Middleman S. Homogeneous Nucleation of Vapor from Polymer Solutions, Macromol.,
18:2274-2276 (1985).
[12]. Shiau, L., The Temperature Dependence of the Pre-exponential Factor and Interfacial Energy for
Aqueous Glycine Solutions based on the Metastable Zone Width Data, J. Crystal Growth., 496-497:18-
23 (2018).
[13]. Siow K.S., Patterson D., Surface Thermodynamics of Polymer Solutions,
[14]. J. Phys. Chem., 77(3):356-365 (1973).

Engineering Journal www.iajer.com Page | 6