An Analytical Method o f Predicting Lee-Kesler-Plikker

Binary Interaction Coefficients Part I : For Non-Polar Hydrocarbon Mixtures

Tlm-modynamic Center

Solomon D.Labinov

Kiev?Ukraine

James R Sand Oak Ridge National Laboratory Oak Ridge, TN

Themphysical Properties June 1924,1994 Boulder, Colorado

Resented at the T w e l f t h Symposhm on

Solomon D. Labinog, James R. Sand

Paper presented at the Tbelfth Symposium on Thennophysical Properties,June 19-24,1994, Boulder, Colorado.
7 .
3

Guest scientist at Oak Ridge National Laboratoly from the Thermodynamics Center, K i e v ,Ukraine. Oak Ridge National Laboratoly Energy Division P . 0 Box 2008, Building 3147, MS 6070 Oak Ridge, Tennessee 37831-6070

Author to whom correspondence should be addressed.

DISCLAIMER
This report was prepared as an a w u n t of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

DISCLAIMER
Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

An Analytical Method of Predicting Lee-Kder-Pkker

Binary Interaction Coefikknts Part I : For Non-Polar Hydrocarbon Mixtures
Solomon D. Labinov James R. Sand Oak Ridge National Laboratory

ABsrRAcr An analytical method is proposed for finding numerical values of b i n q

interaction coefficients for non-polar hydrocarbon mixtures when the Lee-Kesler

ox)

equation of state is applied. The method is based on solving simultaneous equations, which are Pliicker's mixing rules for pseudocritical parameters of a mixture, and the Lee-Kesler equation for the saturation line.

For a hydrocarbon mixture, the method allows prediction of xii interaction coefficients (ICs) which are close to values obtained by processing experimental p-v-t data on the saturation line and subsequent averaging. For mixtwes of hydrocarbon molecules containing from 2 to 9 carbon atoms, the divergence between calculated and experimentally based ICs is no more than f 0.4%. The possibility of extending application of this method'to other non-polar substances is discussed.

KEYWORD3 algorithm; binary interaction coefficient; calculation; equation of

state; mixture; non-polar substance; thermophysical properties.

1.

INTRODUCTIION
In 1975 Lee and Kesler proposed an equation of state for pure non-polar

substances and described mixing rules for calculating pressure-volume-temperature (p-v-t) and thermodynamic properties of mixtures, but binary interaction coefficients were not mentioned in this original work. In order to calculate the pseudocritical temperature of mixtures, the authors proposed the well known equation:

- is the pseudocritical temperature of a mixture, K; where: T

Vcij = (1/8) ( V r+
Tcij = Pd* TJ

V p
%
;

T , , Tciare critical temperatures of the components i and j correspondingly, K;

are molar fractions of the components i and j correspondingly; xii equals one.

V , , Vciare critical molar volumes of the components i and j correspondingly,cm3/gm01;

xi, X ,

In 1978 V. J. Pliicker, H. Knapp, and J. Prausnitz extended this work by

publishing work applying the Lee and Kesler equation of state to phase equilibrium analysis in mixtures of non-polar substances with a considerable difference in critical
1

parameters? For this purpose the authors estabbhed that equation (1) did not agree with the experimental data, and they offered another formula for the pseudocritical temperature of a mixture:

q is an empirical coefficientand K~

is a binary interaction coefficient (IC) that does

not depend on composition, pressure and temperature. The authors assigned a universal numerical value of 0.25 to q, which they obtained by processing experimental data. The
K~~~~

was considered to be a fitting parameter determined by processing

experimental data exclusively. Numerical values of %s were calculated by the authors for approximately one hundred binary combinations of non-polar components, and an empirical correlation of these q ' s for mixtures of hydrocarbons and other non-polar Henceforth, whenever the Lee and Kesler (LK) equation of state was used for mixture available experimental data were used for the thermodynamic property calculations. molecules was established as a function of the parameter (V, Td)/(Vd Td).

analysis, the mixing rule indicated by equation (2) and x i s determined by processing

In this manner the main advantage of the LK equation, which is the ability to
calculate properties of mixtures by making use of critical parameters of pure components, was compromised. The original LK equation provided the opportunity
2

parameters might be estimated by a group contriiution method f r o m only a structural formula. Several unsuccessful attempts were made to restore this advantage by predicting K
~ from ~ component . ~

parameters? This lack of success is not surprising,

because Kijav is a unique fitting parameter that depends not only on properties of components but also on the extent to which the LK equation (and the models of

Benedict-Webb-Rubin (BWR)and Pitzer relation that are built-in) correspond to the physical nature of substances and their mixtures.'$ It is necessary to remember that
the basic BWR equation w a s developed for natural gas property analysis, and the

Pitzer relation is good only for non-polar substances like hydrocarbons. Meanwhile, the LK equation has come to be broadly used for various substances and their mixtures including polar substances.

Additionally, the optimal values of properties, and various authors optimize

K
K~~

differ . ~ for different thermophysical

by means of different minimization

functions working with different properties!

As a result, the set of xiF

values

reported in the literature is badly generalized and shows a large range of variability. It is understandable that
K~~~~

values obtained by different authors for the same

So, for non-polar and polar mixtures, the
K~

mixture differ considerably (Table 1 ) :

development of an analytical method intended for predicting

values only on the

basis of mixture component parameters is still a very compelling problem. In Part I of this article, the problem is solved for mixtures of non-polar substances.

INTERAcIlON COEFFICIENT CALCULATION

In this work an attempt is made to predict the values analytically, to value prediction, and to

develop a fundamentally grounded method for the

estimate the importance of factors that impact the accuracy of the prediction. The most simple assumption that makes it possible to obtain the K~~ value is:

T -

Th,
K~~~~

(3)

From this expression we have to find different values of

for each value of molar

concentration, x , and then to average them. Table I1 presents the xijreference 2 and the averaging is:
0.9 ( '(ir+ 2y11 ; x4.1
K ~ ~ ~ . % ~ ~

values from

values found f r o m equation (3); the formula used for

Kijo25ak =

(4)

K~~ values

were determined over a range of molar compositions from 0 . 1 to 0.9 in 0.1

increments; the sum of values obtained was increased by two because xiiaZr = 1when

= 0, and x = 1.0. Table I1 shows that the calculated values track the experimental

values well, but, naturally, differ from them, because formula (1) is not sufficiently accurate when mixtures of substances with large differences in critical parameters is in equation (1)differs encountered. It is evident that the multiplier, K!, needed for Tcii more from 1.0 with greater differences in critical parameters of mixture components. The equations given in reference 1 were used to find the value of the multiplier.
4

M i x i n g rules of the pseudocritical parameters for a binary mixture assume the form:

2 a =

= 0.2905

- 0.085

1 5 . 2 5 1 8- 1 5 . 6 8 7 5 / T k-1 3 . 4 7 2 1 1 n T k+0 . 4 3 5 7 " ;

- 5.92714 + 6.09648/Tk + 1.28862hTk - 0.16934nL

(9)

where:

Pk = l a w e
T b r
Tb

p,
w

= critical pressure, atm

=Tdrc

= the saturated t e m p e r a t u r e under 1 atm, K .

Z -

= the acentric. factor = the pseudocritical COmpressiWity factor of the mjxture = 'L? i f q = 1.0

Then, knowing that I n PIbr= In(1) - InP, = -InPo and using the following functions:
A = 15.2518 - 15.6875fI'bf - 13.4 721h(Tbf) + 0.43577Th6,

B = -5.92714 + 6.O9648flbr + 1.288621n(Th) - 0.169347T;

and using equations (7) and (9) equation (10) can be derived.

where:

the pseudocritical pressure of a mixture, atm, = f,[(TJI'JJ for a mixture k i K = fA('I'flJJ for a mixture B D i x A, = f,[(TflJ,] for a component 1 4 = f,[(TJI'J.J for a component 2 B, = f&"flJ,] for a component 1 & = f&Tfl&J for a component 2

PM=

based on: that a mixture is a pseudosubstance. U s i n g equation (10) the following may

shown for A and B above. Equation (10) f i t s the hypothesis that the LK equation is

(TDJmk is the pseudoparameter of a mkture, and the functions of fi and fi are those

be determined:

where:

R = 8204, (gas l a w constant; [atma3~mole-oIc]).

If equation (1 1) is set equal to equation (9,an equation w i t h two unknowns,

of two parameters: , T

(TJTJmi. and rcGt0, results. The dependence of (TdT,),, on x is defined by the change
and Tad both are pseudoparameters. The dependence of

TCmk on molar composition is given by equation (5); an equation similar to equation

(5) Cart be written for the pseudoparameter Tam&:

where: + = l/%d T , , , , Tbzare normal boiling temperatures of the components

1 and 2. Setting equation (11) equal to equation ( 5 ) and implementing equation (12),
will result in an equation with only one unknom wijl.,.

With these uijla0 values, the aWz

values may be calculated by setting equation (11) equal to equation (2) for each value

of x and subsequent averaging in accordance with equation (4).

3.

RESULTS

In Figure 1, the xiiO.% values are presented, which have been obbined by
processing experimental data for mixtures of hydrocarbons containing components with
2-9 carbon atoms in their structure, and the xiio=
described above? When the
K ~

values obtained by the m e t h o d

values ~ were ~ obtained, . equation ~ ~ (8) was used to

determine ZcmirFor some mixtures, the experimental data show values that slightly differ f r o m calculated values. This can be explained with the help of Figure 2. This figure shows that the actual 2,values of components (and, consequently, Z-) sometimes different from the values obtained f r o m equation (8). The maximum divergence between the calculated and experimentd 5 values (that is % K
~

are

was ~ approximately ~ ~ ~ ) & 0.4% with an average divergence f 0 . 1 % , which

may be considered quite satisfactory (see Figure 1).

mixtures. The change of the (TOc) d u e against Vcfor pure hydrocarbons C, hypothesis descriiing a mixture as a pseudosubstance.
4.

pseudocritical volume of the mixture, Vd

In Figure 3 . the change of the calculated (IJIJli. values i s plotted against the
for ethane-propane and ethane-nonane

- C,

is also shown. The functions almost coincide which confirms the validity of the

CONCLUSIONS
The method outlined above makes it possible to obtain binary interaction

coefficients values, K ~ for , the LK equations of state w i t h an average deviation of 0.1% using the critical parameters of mixture components, and their normal boiling

, , . The method does not have any restrictions imposed by the nature temperatures, T
of the components and may be recommended as a general method for calculating the
K~~values when

the LK equation of state is applied to mixtures of non-polar substances.

5.

ACKNOWLEDGEMJ3UT3

Research sponsored by the Office of Building Technologies, U . S .Department with Oak Ridge National of Energy under contract No. DE-AC05-840R21400 Laboratory, managed by Martin Marietta Energy Systems, Inc.
REFERENCES

1.

B.I. Lee and M.G.Kesler, AIChE J., 21: 1 5 0 (1975). U. Pliicker, H. Knapp, and J.M. Prausnitz, Int. Chem. Proc., 17: 324 (1978).

2.
3.

S.W. WaIas, Phase Equilibria in ChemicalE n g k e h g , Butterworth Publishers,

Boston (1985).
8

REFERENCES
1 .

B.I. Lee and M.G.Kesler, AIChE J., 2 1 : 150 (1975).

U. Pliicker, 3. Knapp, and J.M. Prausnitz, Int. Chem. Proc., 17: 324 (1978). Boston (1985).

2.
3.
4.

S . W . Walas, Phase EquiZibria i n Chemical Engineering, Buttenvorth Publishers,

5.

ICs. Pitzer and G.O. Hultgren, J. Am. Chem. Soc., @: 4793 (1958).

M. Benedict, G.B. Webb, and LC. Rubin, Chem. Eng. Phys., 10: 747 (1942).

6.

K. Striim, State and 7kansprt Prop& of H i g h Temperature WorkingF W and Nonazeotropic Mixture: h . A, B, & C, Final Report, E A Annex XIII, Chalmers University of Technology, Giiteborg, Sweden, 1992.

Table L VariatMlls in LReKesler-Plikker Interaction CoefIicient from Experimental Data

Mixture
R-22/R-114

Author Hackstein Kruse Str6m Valtz Radermacher

~

q (%Y)

0 . 9 6 3 0 . 9 7 5 0 . 9 7 9 0 . 9 7 3

0 . 9 7 5 1 . 0 4 2 0 . 9 7 3
P

----

---I-

0.97

~-

Lame Strum Valtz Kruse Radermacher

0 . 9 9 7 1 . 0 9 6 1 . 0 4 6 1 . 0 3 8

----

R-22/R-152a

Lavue Stram Kruse Radermacher

1 . 0 1 3 1 . 0 1 4

----e--

Table IL Comparison of 1IF;rCperimental to Calculated Ledbler-PBcker heraction Coefficients U s i n g Equation (3) Mixture ethane & propane ethane & n-butane ethane & n-pentane ethane & n-hexane ethane & n-heptane ethane & n-octane ethane & n-nonane

%c,
Kijarp.
Kijuic

1 . 0 1 1 3

1 . 0 1

c,c 4
1 . 0 2 9

%c,

% c6

1.0333

1 . 0 6 0 5

1 . 0 6 4

1.OB73

1 . 1 0 6

c,c,
1 . 1 1 4 7 1 . 1 4 3

c,C8
1 . 1 6 5

c,c,
1 . 1 6 7 6 1 . 2 1 4

1 . 1 4 1 7

1.30

1.25

= 5 s

p
1

...

.oo

A -0.4
a

0.6

0.5
0.4

e 2 0.3
I L

0 f5 0.2
Z

. ....

0.1
0 0.24 0.25
I

0.26

0.27

0.28

0.29

0.30

COMPRESSIBIUP/ FACTOR (Zc )

FIGURE CAPTIONS

cpart I)

Figure 1 . Comparison of experimental to calculated interaction coefficients for mixtures of ethane with longer, straight-chain alkanes through n-nonane, and relative deviation ;. ~ + - K ~ , &; of experimental versus calculated K ~ S : 0 - I C ~ , ~ 0 - % K ~deviation. Figure 2 . Experimental and calculated compressibility factors, 2 , obtained f r o m equation (8), plotted against the acentric factor, a,for methane (1) through n-decane ( I O ) : 0 - experimental; - calculated. Figure 3. The reduced boiling temperature, Tdr, of pure substances C,-C, and the reduced pseudo-boiling temperature, (TJIJra for mixtures of ethanehonane and ethane/propane plotted against critical volumes (VeVcmJ: + - (TJIJ- for ethanehonane; 0 - (TbT,),, for ethane/propane; - Tbr, for C,-C,