Unformatted text preview: Motivation
The surface tension plays a major role in interfacial phenomena. It is the
fundamental quantity that determines the pressure change across a surface due to
curvature. This in turn is the basis for the stability analysis of both thin films and liquid
jets. While experimental data exists for many pure fluids, it is the goal of this project to
develop an approximation that describes most pure fluids. This would allow the
determination of the surface tension even when experimental data is not readily available. DensityGradient Theory
The surface tension of an interface can be thought of as an excess energy due to
the presence of the interface. In the density gradient theory, the interface is not thought
of as a perfect dividing plane between the two phases. Instead there is a region in which
the density changes from that of the highdensity phase to that of the lowdensity phase.
The energy in the interface region is a function of the density at each point in the
interface. So, the surface tension represents the excess energy present in the interface
region due to the “gradient” in density.
So to derive a relationship between the energy in the interface and the surface
tension, we start with the energy balance for the interface region:
σ∗S = A + PV – Nµ Where A is the Helmholtz energy of the system, S is the interfacial area, P is the pressure,
V is the volume, N is the number of molecules and µ is the chemical potential.
Dividing the energy equation by the area gives (derived in Appendix A, part 1):
σ= (A +P –ρ µ) dx Where the limits of integration are from infinity to minus infinity.
This must be integrated over the interface. The density gradient model assumes a
planar interface in the sense that the density is only changing in the direction
perpendicular to the interface. For this derivation, we will call this the xdirection. Using the expansion for the Helmholtz energy derived in Appendix A, part 2, the expression for
the surface tension becomes:
σ= (∆a + ½ k(ρ) ρx2) dx Where ∆a is given by ao(ρ) + P –ρ µ
The quantity ∆a represents the excess Helmholtz energy per volume in the
interface. The ρµ term in this expression is evaluated at the bulk compositions, as is the
pressure. The ao(ρ) term represents the Helmholtz energy in the interface, which is a
function of density, which changes in the interface. Therefore, this expression represents
the excess energy, the energy in the interface minus the energy in the bulk.
This integral represents the total excess energy in the system due to the presence
of the interface. At equilibrium, thermodynamics requires that the system energy will be
at a minimum. So, the integral must be minimized in order to minimize the excess
energy. Minimization of this integral using the calculus of variation is shown in
Appendix B.
The result is:
σ= (2 k(ρ) ∆a)1/2 dρ Where the limits of integration are from ρl to ρg. Note that the variable of integration has
been changed from x to ρ. This allows the calculation of the surface tension without ever
having to determine the actual dimensions of the interface region. This is the result of the
density gradient theory that will be used to calculate the surface tension. Lattice Fluid Model.
An expression for σ can be found by using an expression for ∆a that is explicit in
ρ. The lattice fluid model provides the necessary equations. However, the integration of
these equations usually requires a numerical integration technique. In order to obtain an
analytical solution, a Taylor series expansion of ∆a will be used. First, the Sanchez Lacombe EOS expressions for a and µ are made dimensionless, as are k and σ. The
expressions are: µ = −ρ + P/ρ + T((1/ρ – 1) ln(1ρ) + 1/r ln ρ)
and
∆a = ρ µ(ρ, P) P  ae
Here ae represents the Helmholtz that the system would have if it were homogeneous.
The italics represent dimensionless numbers. The dimensionless relationships are:
T=T/T* T*=ε*/k P=P/P* P*=ε*/v* ρ=ρ/ρ*
σ=σ/σ* σ*=(P*)2/3(kT*)1/3 µ=µ/Nrε* a=a/P* Here, the quantities P*, T*, ρ* and r which describes the molecular size, are tabulated for
a large number of pure compounds.
In order to apply the dimensionless form of the Helmholtz energy, the equation
for the surface tension must be determined in terms of the dimensionless Helmholtz
energy. It is given by: σ=2 (k(ρ) ∆a)1/2 dρ Now ∆a is expanded in a double Taylor series, first around ρl and then around ρg.
The result is derived in appendix C. It should be noted that k is taken as a constant in this
expansion. Sanchez reports that for the SL EOS, the value of k is 1/6. This is the value
that will be used in this derivation. The result is : σ = 2 k1/2 ∆ρ3/2 [ρg+P/ρl +T [1/2(1 – 1/r)(1 + ρg/ρl)+1/ρl (ln(1ρl)+½((ρl–ρg)/(1ρl)))]]1/2
In this equation, the values for ρl, ρg, P, and T are such that they must satisfy the
EOS, which is ρ2 + P + T [ln (1ρ) + (1 – 1/r)ρ] = 0
and the chemical potential in each phase must be the same. This gives three equations for
three unknowns for a onecomponent twophase system with one degree of freedom.
As a side note, in order to get good agreement with experimental data, a factor of
1/3 was applied to the equation. So, the final equation is: σ = 2/3 k1/2 ∆ρ3/2 [ρg+P/ρl +T [1/2(1 – 1/r)(1 + ρg/ρl)+1/ρl (ln(1ρl)+½((ρl–ρg)/(1ρl)))]]1/2
this can be simplified to: σ = 2/3 k1/2 ∆ρ3/2 T 1/2 [ 1/2(1 – 1/r) +1/ρl (ln(1ρl)+½((ρl)/(1ρl)))]1/2
Here the reduced density of the gas and the reduced pressure have been assumed
to be negligible. The reduced gas density is of the order 103 and the reduced pressure is
of the order 104 to 108. Neglecting these quantities makes less than 0.1% difference in
the results. Calculations
The calculation procedure that was used is quite simple. Given the temperature of
the system, the reduced temperature can be calculated. Then a reduced pressure must be
guessed. With these two quantities, the equation of state can be solved for the two
density roots, the gas density and the liquid density. Then the chemical potential is set
equal in each phase. This allows for a new reduced pressure to be calculated. With this
new reduced pressure, the reduced densities are recalculated and then used to calculate a
new reduced pressure. This usually only requires two or three iterations and can be done
on a calculator with a root finding function. This was the method used to find all the
numerical results for this problem. Results Some of the results calculated using the above equation are presented below.
Substance Temperature σexperimental σpredicted %diff Benzene 20C 28.88 29.67 2.73 30C 27.49 27.76 0.98 80C 21.2 19.66 7.26 120C 19.5 19.55 0.26 110C 17.65 17.04 3.46 100C 15.71 14.82 5.66 propane 230K 15.94 15.91 0.19 nbutane 230K 20.64 20.93 1.40 isobutane 230K 18.30 18.50 1.09 diethylether 20C 16.96 16.22 4.36 oxygen 193C 15.7 15.02 4.33 26.8 27.70 3.36 ethylene Carbon Tetrachloride 20C It can be seen that the agreement is reasonable as an estimate for these
compounds. In order to determine the accuracy of this equation over a large temperature
range, the equation was used to predict surface tension values for propane, butane, and
benzene. The predicted values are plotted along with the experimental surface tensions
on the following page. It can be seen that good agreement with experimental data is
obtained over a large temperature range, almost 200K for butane. So, this equation gives
good estimates of the surface tension not only at ambient conditions, but also at points far
away from ambient. This equation surprisingly does a good job near the critical points of
the fluids even though the reduced pressure was neglected. So, overall this equation
fulfills the goal of giving a reasonable estimate for the surface tension of pure fluids. References:
Poser, C, I.; Sanchez, I. C. J. Colloid Interface Sci. 1979, 69, 539.
Vargaftik, N. B., “Tables in the Thermophysical Properties of Liquids and Gases,” 2nd
Ed. Wiley, New York, 1975.
Enders, S. ; Quitzsch, K. Langmuir 1998, 14, 46064614. Determination of Pure Fluid Surface Tensions Project Report
ChE 385M Surface Phenomena
Submitted by Andrew Metrailer
May 4, 2000 ...
View
Full
Document
This note was uploaded on 02/25/2010 for the course MEDICIN 101 taught by Professor Anders during the Spring '08 term at Aarhus Universitet.
 Spring '08
 ANDERS

Click to edit the document details