Louisiana State University LSU Digital Commons

LSU Historical Dissertations and Theses Graduate School

1950 Vapor- Equilibrium of Non-Ideal Solutions. Virgil Orr Louisiana State University and Agricultural & Mechanical College

Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_disstheses Part of the Chemical Engineering Commons

Recommended Citation Orr, Virgil, "Vapor-Liquid Equilibrium of Non-Ideal Solutions." (1950). LSU Historical Dissertations and Theses. 7950. https://digitalcommons.lsu.edu/gradschool_disstheses/7950

This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion in LSU Historical Dissertations and Theses by an authorized administrator of LSU Digital Commons. For more information, please contact [email protected]. VAPOR-LIQUID EQUILIBRIUM OF MCM*ID^L SOLUTIONS

A Dissertation

Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College In partial fulfillment of the requirement® for the degree of Dootor of Philosophy

in

The Department of Chemical Engineering

by V irg il Orr B.S#, Louieiana Polytechnic Institute# 1

All rights reserved

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion.

UMI

Dissertation Publishing

UMI DP69328 Published by ProQuest LLC (2015). Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code

P r o Q u e s t

ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106 - 1346 MANUSCRIPT THESES Unpublished theses submitted for the master*s and doctor *s degrees and deposited in the Louisiana State University Library are available for inspection* Use of any thesis is limited by the rights of the author. Bibliographical references may be noted, but passages may not be copied unless the author has given permission.

Credit must be given in subsequent written or published work, r A library which borrows this thesis for use by its clientele is expected to make sure that the borrower is aware of the above restrictions.

LOUISIANA STATE UNIVERSITY LIBRARY ^ . S9419

ACKNOWLEDGMENTS

The author Is very deeply Indebted to Dr. Jesse Coates, director of this project, for hie advice and suggestions*

Grateful acknow1edgment le alec made*

To Dr* P. M* Horton for hie rn&ny helpful suggestions*

To Hr* S* E* Snyder for fabricating and maintaining the

equipment*

To Mrs* Grace Cameron, chemistry librarian, for her invaluable

assistance in preparing the literature survey*

To Mr* A* J* Gully for his assistance in performing the

statistical analyses on the analytical methods and in preparing

the graphs*

To hie fellow students for their advice and constructive

criticisms*

j

K '\ j •?« 11 iq 5 b

6 , ^ ■ 1 3 2 3 S : j TABLE OF CONTENTS

CHAPTER PACE

X. INTRODUCTION...... * ...... I

IX . THSORETIOAL CONSIDERATIONS...... 5

A. Derivation of th© Thermodynamic Relatione J

B. Development of th e In te g ra te d Forme o f th e

Gibbe-Duhem Equation »•**..#.«.. 8

0* Effect of on Activity

Ooefficienta • 22

Dm Other Correlations ...... 27

E. Conclusions . G ...... 28

I I I . AN EQUILIBRIUM FOR PARTIALLY MILCIBLE

LIQUIDS...... 55

A. Resume* of Previous Designs 55

B. Design of the New Type of Still ..... 59

0* Operating Procedure and Characteristics * 45

D. Proof of the Equilibrium Still ...... 50

E* Reagents ...... 64

F. Conclusions ^ 6 5

IV . EXPERIMENTAL DATA ...... 66

V. DISCUSSION CF RESULTS...... 100

A. Development of a More Versatile Equilibrium

S t i l l ...... 1 0 0

i i i CHAPTER FA OS

B* Proposed Modification® of the Existing

Design ..»•••...... a..** lO^ O* Experimental Results • •«••••.»•* 105

D* Correlation of Data ...... lOJj

VI. SUMMARY- * * . • 120

bibliographt ...... 12a

APPENDIX ...... 126

BIOORAPHT ...... I?!

lv LIST OF TABLES v TABLE PAGE

I* Experimental Vapor-Liquid Equilibrium Data for

Ethanol- at Atmospheric Preeeur© • » . * • 5 2

II* Experimental Vapor-Liquid Equilibrium Data for

n-Butanol-Water at Atmospheric Preeeur® • * • • 5 5

III* Vapor-Liquid Equilibrium Data of Stookhardt

and Bull on n-Butano1-Water at Atmoepherio

Preeeur® • •*•«••••*•••«** • . » 5 7

IV* Vapor-Liquid Equilibrium Data on Smith and

Bonner on n-Butanol-Water at Atmoepherio

Preeeur® .*.***••««• ...... ••• 60

V* Experimental Vapor-Liquid Equilibrium Data on

Ethyl Acetate-Water at Atmoepherio Preeeur® • • 62

VI* Phyeical Properties of Reagent® • ••••**• 64

VII* Experimental Data for n-Rexano-Ethanol at

250 mm* Hg* **«••*••••**•#.»•* 72

VIII* Experimental Data for n-IIexane-Ethanol at

595 mm* Hg* • •*••••*•••••**»*, 7 6

IX* Experimental Data for n-Hoxano-Ethanol at

760 mm* Hg* ...... 60

X* Experimental Data on n-Hoxan©-Ethanol at

1270 mm* Hg* • *•••*»*•«*•*<.«•• 64

v TABLB PASS

XI* Experimental Data on Hexane* Eft hanol at

X?45 i®b* Hg* ••**•«**•***•*#**« 8 8

HI# Experimental Data on n-Kex&n©* Ethanol at

2510 mm* Hg* * * ...... *•**.***••* 92

XIII* Experimental Data on n*Hexane Ethanol at

28^0 osa* Hg* **•**•*••••»«*••** 96

XXV* Oalibratlon Data for Goppor^Qonotanian

Thermocouples *«*•*»•*•**•**«** 129

XV* Solubility of m-Buianol in Water *•#**»•* 1^1

XVI. Solubility of Ethyl Acotato in Water *..*..

v i LIST OF FIGURES

FIGURE PAGE

1* Van Laar Constant® for Ethanol-Wat sr at 5°®0* « « » 15

St* Activity Coefficient Curves for Ethanol-Water

a t 50 °C. . * , ...... 14

5* Van Laar Constants for EthanoX-Water at Atmoepherio

Pressure » 1 * * ...... * ...... * » * • 15

4* Van Laar Constants for Sthanol-Water at Atmoepherio

P ressu re **.**««* ...... «*«****» 16

5* Aotlvity Coefficient Curves for Ethanol-Wator

at Atmoephsrls IJ

6 * Activity Coefficient Curves for Ethanol-

Chloroform a t 4j?°0# 16

7* Van Laar Constant® for Ethanol-Chlorofora at

^5 ° o ...... 19

6 . Othmer S t i l l...... 34

9* Jones* Schoonborn* and Colburn S till »««•»«* 55

1 0* Detail Drawing of New Type Equilibrium Still • * • 40

11* Oblique View of Original Design ...... 4 l

12* Front View of Original Design • • • • ...... 42

15* Front View of Modified Still ...... • • • 46

14* Oblique View of Modified S t i l l ...... 4

15* x - y Diagram for Ethanol-Wat or at Atmoepherio

P ressu re »•*»•«**••#« * * ...... 51

v i l FIGURE PAGE

%6 * x - y Diagram for n-Butanol-Water at Atmoepherio

Pressure * * * »*» .««**.*.*>»»»* * 55

17* Bolling Point Diagram for n-Butanol*Water at

Atmoepherio Preeeure *•*•«»*•••»•••* ^4

Id* Experimental Activity Coefficient Curves for

n-Butanbl-Water * * * *•*•••**••••«* 5 6

19* Aetivity Coefficient Curves for n-Butanol-Water

from the Data of Stockhardt and Hull »••»•»* $ 8

20* Aetivity Coefficient Curves for n-Buianol-Water

from th e D&ta o f Smith and Bonner »#*««••* 59

21* x - y Diagram for Ethyl Acetat©-Water at

Atmospheric Pressure • •••••*••*•«•** 6 $

22* Laboratory D is tillin g Column • •**••••*»* 6 j

2 5 * Jones* Sohoenborn* and Colburn S till1 * ****** 71

24* x - y Diagram for n-Hexane-Ethanol at 250 mm* Hg* 75

25* Point Diagram for n-Hexahe-Ethanol at

250 am* Hg* *#*••*»••*»•*•»»#* * 7 4

26* Aetivity Coefficient Curves for n-Hexane-Ethanol

a t 250 mm* Hg* ...... • 7 5

27* x - y Diagram fo r n-Hexane-Ethanol a t 595 art* % • 77

28* Diagram for n-Hexane-Ethanol at

595 mm* H g* .*«••*«•* 7 8

29* Activity Coefficient Curves for n-Hoxana-

Ethanol at 595 mm* Hg* * * 79

v t l i FIGURE

50* x - y Diagram for mHexane-Ethanol at 760 mm* Hg* • ©1

51* Bolling Point Diagram for n-Hexane-Ethanol at

760 Ban. Hg* .*.*•*«•*..«.«».**** ©3

5S* Activity Coefficient Curves for n-Hoxane-$thanol

a t 7 6 0 mm* Hg* • «••**.. ••*.**«..• ©5

55* x - y Diagram for n-Hexane-Ethanol at 1270 mm* Hg* ©5

J4* Boiling Point Diagram for n-Hex&ne-Ethanol at

1B70 mm* Hg* •**•*. *••**..»•••«. 66

55* Activity Ooofficiont Curves for n-Kexan«-Ethanol

a t 1270 am* Hg* * * ...... • **.*»...** ©7

J6* Boiling Point Diagram for n-Hexane-Ethanol at

15^5 *■» *•••*•**•«*.. •»....* ©9

57* x * y Diagram for n-Hexane-Ethanol at 1^4^ mm* Hg* • 90

5©. Activity Coefficient Curves for n-Hexane-Ethanol

at 1545 mm* Hg...... 91

59* x * y Diagram for n-Haxane-Ethanol at 2J10 mm* Hg* • 95

40* Boiling Point Diagram for n-Hexane-Ethanol at

2510 mm* Hg...... 94

4l* Aetivity Coefficient Ourves for n-Hexane-Ethanol

a t 25IO mm. Hg. 95

42* x - y Diagram for n-Hexane-Ethanol at 28^0 mm. Hg* • 9 7

4j* Boiling Point Diagram for n-Hexane-Ethanol at

285 O mm. Hg. 9 ©

44* Activity Coefficient Curves for n-Hexane-Ethanol

a t 2 Q 2 0 mm. Hg* 99

Ix FIGURE PAGE

42* Corrected Activity Coefficient Curves for

n-Hexana-Ethanol at Z%0 ram* Hg* • ••»»»•<.•• IQS

46* Corrected x - y Diagram for n-Hexane-Ethanol at

2%0 mm* Hg ...... **»*•*****•**•• 10?

4?* Correlation of the Data on n-Hexan©-Ethanol at

1545 ram* Hg* by Proposed Equation * »»«**•*• 109

4$* x - y Diagram for n-Hexan©-Ethanol at 15^5 ram* Hg*

with Limiting Slopes *«•**•*•*••*•»*« 111

4p* White Correlation of Activity Coefficient Data * « • 112

20* Copper—Constantan Thermocouple Calibration * * * • • 12$

21* Phyeieal properties of the n-But and-Water System * 122

22* Physical Properties of the Kthanel-Waier System * * lj4

22* Physical Properties of the n-Hexane-Ethanol System • 122

24* Oox Plot ****•*•*«•*•* • X$$

x ABSTRACT

Previous investigators developed apparatus for obtaining vapor-liquid equilibrium data and derived methods of oorrelatlng this data based on thermodynamic considerations* The methods used to obtain vapor-liquid equilibrium data have limitations that are usually imposed by the physical > of the systems being studied* These limitations have prevented determination of a sufficient amount of reliable data on the various types of systems encountered to permit development of suitable correlations for the extension of limited data or prediction of values from a minimum amount of related data* Some of the relations previously presented have found wide application as approximations and in some cases have been applied under conditions for which they are not valid*

A more versatile recirculation type equilibrium still has been developed for the determination of vapor-liquid equilibrium data on mieeible or partially miscible systems at ranging from below atmoepherio to considerably above atmoepherio* The still is of metal construction utilizing a centrifugal pump with a large by-pass line as a condensed vapor receiver* The pump will agitate the two liquid phases

00 completely that they behave as a homogeneous, single fluid and can be removed from this receiver in the ratio that is present* This still was found to satisfactorily duplicate literature values for the mieeible system, Ethanol-Water, and the partially miscible system, n-Butanol-Water, at atmospheric pressure*

xi Vapor-liquid equilibrium data for Ethanol-Water* n-Buianol-Wator#

Ethyl Aeeiate-Water# and n-Hexane-Ethanol at atmospheric pr0 0 sura and

£or n-Hoxana-Sthanol a t 2f50 mnu, 59$ mm.* 1270 mm*# I jb j mm** 2J10 mm.# and 2 8 5 0 nan* total pressure aro presented*

Tho following aquation la proposed for the correlation of activity coefficients derived from vapor-liquid equilibrium data for binary aye t erne rather than the commonly used empirical power series 8

^ J'l ~ /_ . / _ /f> / I /. / tT- P„ 1 /Z_ /77 __ iP ) where Y' — activity coefficient of component 1

Y 2 activity coefficient of component 2

y^ — mole fraction of component 1 in vapor

— mol© fraction of component 1 in liq u id

y^ — mole fraction of component 2 in vapor

x — mole fraction of component 2 in liquid 2 t •“ fugaeity of component 1 when under ite own vapor pressure * 1 at the temperature of the solution

f **• fugaeity of component 2 when under its own vapor pressure 2 at the temperature of the solution

fugaeity coefficient of component 1 evaluated at a reduoed

temperature corresponding to the critical temperature of

component 1 and a reduoed pressure corresponding to the

total pressure and critical pressure of component 1

fugaeity coefficient of component 2 evaluated at a reduced

x ii temperature corresponding to th© critical temperature of

component 2 and a reduced preeeur© cor responding to the

total pressure and orltleal pressure of component 2

— molal volume of component 1 In the liquid W1 —* molal volume of component 2 in the liquid Ma — vapor pressure of component 1

— vapor pressure of component 2

R — Constant

T — absolute temperature

^T T — total pressure

An improved method for prediction of activity coefficients and vapor-liquid equilibrium over moderate pressure ranges from accurate isebarle boiling point measurements over the entire composition range is presented* This method Is based on the proposed correlation equa­ tion and seems to permit prediction of this data as accurately as It can be determined experimentally* The equipment necessary to obtain the basic data is much less involved than that for determination of the vapor-liquid equilibrium data*

x i i i CHAPTER I

INTRODUCTION

The development of azeotropic and extractive aa

Important lnduatrial processes has stimulated m&ny investigations of the vapor-liquid equilibria of non-ideal systems* While a large amount of data have been obtained, correlation of this data in a generalized and conveniently usable form has proved to be an exceed­ ingly difficult problem#

The usual approach to correlation ie by ©valuation of th® deviations from the laws of ideal solutions employing equations based on thermodynamic ooneidorations# Use of thermodynamic analysis has proved to b© very valuable in correlation of vapor-liquid equilibrium*

For example, the Gibbs-Duhem equation, which in differential form expresses the necessary conditions for any phase equilibrium under conditions of constant temperature and pressure, ie commonly employed as a test for thermodynamic consistency of vapor-liquid equilibrium data, tfhile thermodynamic consistency does not in itself constitute a proof of equilibrium, it doee give strong evidence of such* Incorrect data may be consistent, but thermodynamically inconsistent data cannot be oorreot. Thermodynamic consistency should always be used as a guide when drawing a line through experimental vapor-liquid equilibrium data.

Instead of using the differential form of th© Gibbs-Duhem equation, certain investigators*^' ^ have developed integrated forms which

1 2

will be discussed more fully in later sections* These equations are commonly applied under conditions of oonetant pressure and varying temperature for which they are not valid, having been derived at oonetant temperature* Of these integrated forms the Van £#aar^* ^ equatione are probably the most widely used* They contain two empirical constants which are supposedly independent of composition at constant temperature for any particular substance, but which are known to vary with temperature* The assumption that these constants are independent of temperature has lead to poor results in many oases and points up a very definite need for inveetigation of the effects of temperature variation on these constants* Since liquid "activity coefficients" are a qualitative measure of the deviations from ideal solution behavior, and since the constants of the various integrated forms of the Gibbs-

Duhem equation are e&elly related to these activity coefficients, there is a need for a study of the variation of these coefficients with tempera­ ture* CHAPTER I I

theoretical considerations

A. DERIVATION OF THE THERMOWTNAMIO RELATIONS*

7 The Gibbs-Duhem equation la & rigorous thermodynamic expression

of the necessary conditions for phase equilibrium of a closed eyetom

at constant temperature and pressure* It was first deduced by J* 1 0 o Willard Gibbs in 1079 and later* independently by P# Duhsm in 1866*

It has been shown that the change in free energy* F* In a homogeneous

phase as a function of the temperature* T» the pressure* P* and the

number of moles of each component present* n^* n^*• »** ie

, / 3f)^r + /f^) <*p +(l£ )d n , + L2£ \d n z + •••

Since one of the eondltions of a closed system in equilibrium at

constant temperature and pressure ie (SfJ^p * 0* Equation (1) reduces

** (if) oln, + (§ £ ) = O (2) C an'/77pt ne;-° ' p, V * *

where

\a( ^ n , ) = r * / which ie known as the partial molal free energy* Since Equation (2)

ie a partial molal equation* the Independent and dependent variables

can be switched yielding the following expression for a binary system s

». ( % k 5 4

In order to define seme terms* a digression Is In order* The partial molal free energy* V* is related to the fugaeity* f* at oonetant temperature fey the following equation which ie a partial definition of ftogaeity*

< * >

Qualitatively* fugaeity is a measure of the esoaping tendency of any particular substance in any particular phase and th© concept was 24 introduced by G* N* Lewis to account for the behavior of real as compared to perfect gases* It should be remembered that when two phases are in equilibrium* the fugaeity of any particular component is the same in both phase*# s© if the fugaeity of a component in the vapor phase can be calculated* the fugaeity in the liquid phase Is obtained simultaneously*

Fugaeity of pure gases Is related to pressure through the basio 17 thermodynamic fu n ctio n s as described by Hougen and Watson w ith th e following results*

(5) where

f s the fugaeity of the pure gas at temperature, T* P * the pressure* z * the compressibility factor* z * for one mole*

This equation is easily integrated graphically or analytically if a suitable Is evaluated* This integration has been carried out in conjunction with the theorem of corresponding states and the ratio of fugaeity to pressure tjp has been plotted against 5

reduced temperature and pressure* This ratio (|>) is coraraonly termed the Hfugaeity coefficient* and these plots provide a very convenient method for obtaining the fugaeity of a pure eubetanoe at any condi­ tion* of temperature and pressure*

For treatment of problems involving solutions it is convenient to define another thermodynamic property which is directly related to fugaeity and hence also to free energy* This property* called activity* a* ie defined as the ratio of the fugaeity of a component In a given state to its fugaeity in an arbitrarily defined standard state at the same temperature* Thus*

(

a = activity f * fugaeity in the given state f°* fugaeity in the standard state at the same temperature and pressure which can be evaluated by Equation (5) under its own vapor pressure at the temperature* then corrected for pressure*

This correction for the effect of pressure on the fugaeity of the

liquid is derived from the same basic coneid©rations as Equation ( 5 )

for gases* The expression is

(7)

where

fugaeity of the pure liquid under its own vapor pressure at bll* vviltyw abur*f X • vm - molal volume of the liquid at th© temperature* T* and pressure T t •jf • total pressure of the system* P s vapor pressure of the pure liquid at the temperature* T* R s Gas Constant* 6

Combination of Equations ( 6 ) and (4) give# /er^u a = T -~f 0

T * partial molal free energy In the given state at temperature T* F° * partial molal free energy in the standard state at the e&me temperature T*

Xn the case of liquid eolutione it ie eonvenlent to define the standard state as the pure component at the temperature and pressure of the * With this oholee of etandard state the activities become equal to the mole fractions* x# in which form ideal solutions* since f in Equation ( 6 ) can be replaced by xf°. It ehould be noted that this is a variable etandard state since in the ease of oonetant pressure vapor-liquid equilibrium* the temperature of the solution varies with composition* or when isothermal the pressure varies with composition*

The effect of this variable standard state on activity should be studied*

The statement* that whoa ideal liquid solutions are formed* the aetivity equals the mole fraction* does not hold for non-ideal liquid 24 solutions* To account for this 3* fl* Lewis introduced the activity coefficient* Y , which when multiplied by the mole fraction* x* gives the activity* Thus

YJC - CZ = ( 9 a) or

K- -5- -- — o 9 “ JC p f (9b) The standard state has already been defined and may be summed up in the following manner# A standard state of unit activity is often very convenient* and it is known that the behavior of a solution approaches 7

the ideal ( ^solvent approaches unity) as the pure solvent Is approached

(x approach©© unity)* Hence the product* V^xf or the aetivity alee approaches unity* and the pure component at the temperature and pressure of the mixture yields a convenient etandard state of unit activity*

For binary systems under conditions where the vapor phase may be treated as a perfect gas (low pressures) without much error and where the vapor phase may be assumed to represent an ideal solution* Equation

(9 b) reduces to the following for component It

r f - r - H * 71l Z ' Z , p * z , < 1 0 > where

— mole fraction of component 1 in the liquid phase* y, - mole fraction of component 1 in the vapor phase*

- of component 1 at the temperature of the solution* “ vapor pressure of component 1 at the temperature of the solution*

-jj* a total pressure of the system*

The asterisk represents a pressure sufficiently low that the above assumptions will hold*

From Equations (5) and ( 8 ) one form of the Gibbs-Duhem equation for binary mixtures can be written n, d JU a, &£. - ° (ix) It is desirable for graphical manipulations to express this in terms of activity coefficients since they vary less with change in composition than do the activities* For a binary mixture the following development can be written a

Xf -h Xa - I c J X f + d Z & = O

( 12)

Equation ( 1 1 ) ©an be written in the following equivalent form

Xi d Q i ~t~ d -A t Q^-O

How subtraction of Equation (13) from Equation (11) give©

X . = o

d J U r , = — ^ c J / ^ ^ £ * t x . Xf (1?) This form ie more convenient for graphical integrations, but the following form ie more convenient for cheeking thermodynamic consistency of vapor-liquid equilibrium data by use of the elopes of the usual activity coefficient versus mole fraction plots*

(!*)

B. DEVELOPMENT of THE INTEGRATED FORMS OF THE GIBBS-DUHEM

EQUATION*

These integrated forms, which are the expressions most commonly used to show the variation of activity coefficients with composition at conditions of constant temperature, are most conveniently derived by consideration of} the free energy changes accompanying the formation of a solution from its components in their standard states* According 1 7 to Hougen end Watson the free energy change of mixing, F^, is given by the following general equation where the standard state ie taken as 9

the pure components at the temperature and pressure of the systems

^ ^ > 7 ~ * 7 * E c' < 5 T /V E l where rr- ■S'/I,' fu ' - ^ ^ + rtg. F k + ' *' o ^ n,' r?,Ft° + n t F2 + - ' ~ F Combining Equations (15)# ( 6 )# and (9b)

where * m J*. x * = ^ A ^ ' “

/ 7 ^ - X . Y t = n ,^* i + n*'£~ rh + m" For an ideal solution all aetivity coefficients are equal to unity and the term reduces to aero* This term was designated by

Seateh&rd and Hamer^ as the excess free energy# F®# of the solution*

Thus# the free energy# F# of any solution may be written as

F = J E /& ' F i° + ^ (17) where

F esl Cc z th© excess free energy*

The partial molal free energy of any component is obtained by dif­ ferentiating Squation (1?)# thus for component 1#

— — _ ^ j. 3 F e F, = F , ° + ------(18) Expansion of the ssoond terra combined with Equation (12) gives

a ^ / + rt, ~ ^ /7/ fc T \^ £ u x , >*-/ (19) Then#

( 2 0 ) / T = F , ° + F T ^ X , + ^ 10

By comparison of Equation* (20) and ( 6 ) It. may be seen that.

tf = s n d e f z r t£ - ( 22)

Wohl-' pointed out that the empirical equations which have been

commonly used for eerrel&tion of activity coefficients represent special

cases of the following equation for the excess free energy* F®*

— %, ^ t 'Z'. ^ +/E f ' 2i fZ'l Xf / ** ft **.■&£& where

- mole fraction of component 1* qi » effective molal volume of component 1 * zi * effective volume fraction of component 1* aih etc.* empirical constants correeponding to the indicated * " groups of components in the summations* Subscripts if h, j9 and 1 each may correspond to any component of

the mixture in the terms of the indicated summations*

Equation (2J) is designated a four suffix equation as characterized by the last summation term Indicating interactions of groups of four * If the last term is dropped there remains a three-suffix

equation. Each additional summation introduces added constants to permit improved representation of increasingly complex relationships*

The relations which have been used in the past to correlate vapor- liquid equilibrium data have assumed that Equation (2J) or some simpler form of Equation (2^) is a satisfactory expression for the exoess free

e n e r g y $ which is questionable* Therefore the value of such equations

is doubtful except as an approximation# 11

The effective molal volume* q» in Equation ( 2^) serves to relate the effective volume fractions* z# to the mole fractions, x, thue

^ ______7 *,+ 9% tz + t t y * * * ' " (24) and

^ = %+ % U +% <25> Since Equations (21) and (22) are valid* thermodynamically con­ sistent expressions for the logarithm of the activity coefficients at constant temperature can be obtained by differentiation of a suitable

BS/| fora of Equation (2^)* tfohl^ did this for a binary system of com­ ponents 1 and 2 using a three-suffix form of Equation ( 25) w ith th e following equations be'lng developed*

‘ ( 2 6 )

^ r * * z ? C e + g (* % ~ 8) (27)

These equations contain three constants, A, B, and q^/q^, which must be empirically determined for each system under consideration* By use of various assumptions regarding the ratio of the effective molal volumes, q^/qg, the number of empirical constants may be reduced to two*

Margules2^ in effect assumed this ratio q^/q^ to be unity* With this assumption Equations (26) and (27) reduce to the Margules equations £ as modified by Carlson and Oolbum •

^ r , - *11*+ £ = (£re~*} (26)

y-£ = xffjs+ zQ i-B ) **7 = (29)

Beatehard^ and his coworker ^ took the effective molal volumes as equal to the aotual molal volumes of the pure components so 12

that tj/tjj equaled

Van Laar^°f assumed the ratio q^/q^ equal to A/B and roduced equations (26) and (2?) to the following forms as rearranged by Carlson wad Colburn?» _ £ . „ Z _ /?Z* r , = " Z * - (50)

2 8 X , a -&fr tg.- - (x, + %Ze.)Z’ ^ which may be rearranged to a form suitable for solving for A and B d ir e c tly . . >. - j 2

T fJ (52 )

4 t 4 t D <»> The work of White^ suggested the rearrangement of these equations In the straight line form as follows*

^ y ,/£= ^ ^ (34)

■ £ ( % ) + - & Differentiation proves that equations ( 5 0 ) through (53) satisfy the general relation of Equation (14) provided A and B are considered as true constants* The result of this differentiation is as follows*

( 4 ^ 1 , = : ~ ¥ * > ' 0

0 1 ( H * £ ) i , . o ~ - However these "constants" are not constant even under the conditions for which the equation was derived as can be seen from Figures 1, 2* 9* 4,

5* 6, and 7* Figure 1 is a plot of the Van Laar constants A and Bf which were calculated by use of Equations (J2) and (53) bh© isothermal data MSTANTS VERSUS GOMPO : 0 R ETHAN OL1— WA ewT 50

LEGEND to qf es, Sohounborn, id ..Gotbur

0 10 2 b 30 4 3 50 do 70 80 90 MOLE PERCEN T ETHANlOL in l iq u id FIGURE j rf' jQQEF CURVES THANQL1 At bo6 c

LEGEND Data! of Jonos, Schoenbbrn, arid dolt T h e rm o d y namically a is le n t 15

[van laar constants .versus ... COMPOSITION for : ETHANod—WATER At I 76Q: mm PRESSURE - ______

j : ) ; ' ILEGEND 0 Data of Risder and Thompson

0.6

0.2

MOLE PERCENT ETHANOL! IN LIQUID

i L 1 ! L FIGURE 3j _.i____ 16

VAiN U V A R G "ANTS: VERSUS: temperature "HANOti— WATER A I 26.Q tnm. , PhESSL)RE ...... _

L ^ G E

0:4

eo 1 0 0 |TEMFmRA|-Uft& (°C.)

[. L.iu mBE— I ' ! * 1 ' 17

LEGEND eder and hompson

fr-3

100 v iq le. p e rg e m t; ETHAMOli IN LIQUID f i g u r e . 5 I : .... ACTIN') TY . COEI: ETHA NQLl-r AT . . .45

LEG EN p D ata of S catchdrd and Raym Sond C o n s i s t e n t With Log /,

Ip 2 D : 30 4 D TO 8 MQLB PERCEN ETHANC L1 IN L VKRSUS R ETHANOL1— ;

! " ' jLEGENQ : i . 1 Q Data of Scdjt chard jand feaymfefid j L._ ,_4 _ l~4.

jMQLE PERCENT ETHANO IN: LIQUID FIGURE 7 20

32 0 for Ethanol- Water of Jones# Schoenborn# and Go1burn at JS>0 0## ag&inet

mole percent Ethanol in the liquid* Figure 2 is a plot of logarithm of

the activity coefficients against mole percent Ethanol in the liquid# and

wae inserted for purposes of comparison since Equations (JO) and (Jl) show that log iTj equals A at equal aero# and likewise log ^ equals B at

equal unity* The circles show what log should be if it were thermo*

dynamically consistent with log Y^ according to the Gibbe-Duhem Equation*

Figures J and 5 are similar plots for the constant pressure (7^0 m illi­

meter) data of Bieder and Thompson^ on the same system* The constante

would be expected to vary here since the temperature varies from a low

value ©f 7^*15° 100,00° 0, In Figure 4 the constants are plotted

against the temperature corresponding to the various liquid concentra­

tions of Figure 5* Choice of any of the many possible values would lead

to a poor data fit over some range of the ooncentratlone which is un­

desirable# therefore these are approximations which should only be used

in the absence of any but a minimum of experimental data* The highly

precise data of Scat chard and Raymond^ on Ethano1- Oh 1 or o form at 4^° C*

is shown in Figures 6 and 7* The circles show the data to he more thermo­

dynamically consistent than the Ethanol-Water data# but the variations of

the Van Laar constants is still evident*

If the validity of the Van Laar type equations is accepted# vapor-

liquid equilibrium data can be calculated over the entire concentration

range from a knowledge of one point. It is apparent from examination of

Equation (l) that at the azeotroplc point where x - y# the activity co­

efficients can be evaluated from a knowledge of the vapor pressure of the

particular components at the azsotropic temperature together with the total 21

pressure* fleing Equation* (28) and (29) ar (5&) (25) th© constants

A and B can be evaluated fro® a minimum of data* This method la quoted by Hougen and Watson*^ a* giving fair re suite when the aseoiropic point lies between 0*25 and 0*75 &eie fraction.

K urt Wahl"" ^ reviewed th e foregoing method© and equation© and has made recommendations as to whieh equation give© best result® for various types of eystems* His recommendations are as follows«

It is demonstrated that the usual Msrgules equation will be quantitatively useful mainly for system a of relatively small dissymetry for whieh it is about equivalent with the u su al Van Laar equation* The l a t t e r holds fo r many ©yeterne up to somewhat higher degrees of dissymetry# When it does not fit the data it 1© suggested to use the three-suffix Soatchard and Hamer equation or, with special advantage In certain cases the four-suffix Van Laar equation* In cases of still higher dissymetry in which the curve for the activity coefficient with the lower end value shows a maximum or & very flat course in the neighborhood of its end value, the three-suffix q-equation may be applied* Still more use­ fu l will be the four-suffix Mar gules equation* 3 Benedict and his coworkere^ used an equation of the same type as

Equation (25) together with curve fitting of experimental P-V-T data by the method of least squares to evaluate certain constants of the binary systems and then used those to evaluate ternary data composed of t :,e three binary mixtures* They claimed that this circumvented the inaccurate assumptions of the Van Laar type equations*

Their expression relating the activity coefficient of any component to the vapor-liquid equilibrium compositions for the ©am© component Is a modification of Equation (9) based on the theory that the second coef­ ficient of the virial equation of state is sufficient to express the deviations from th© perfect gas laws# Hie equation is

r = — <*> * i x , ^ e 22

where v^ is th© molal volume of the liquid component 1 at the temperatur© end pressure of the solution* ami Is the second virial coefficient of the virial equation of state

/ V = r t ( / + y ■ ; / (?7) where B and 0 are the viri&l coefficients and are functions of tempera- ture* the additional term of Equation (55) corrects for the non­ ideality of the gaseous phase and the effects of pressure on the liquid*

0 . EFFECT OF TEMPERATURE OH ACTIVITY COEFFICIENTS

The equations developed by White^2 from considerations of the Van der Waala equation of state for mixtures of gases included a temperature factor and are valid for low pressures only* if at all, since the Van der

Waale equation of state is not adequate to express the behavior of most systems* His equations are a * i -rJ o f r, =• (56)

& l£ Jn!'S75 f c = (59) where eonstants *att and ^b* are the constants of the Van der Waals equaticnt^^p2J^ ^ )r By inspection It can be seen that these equa­ tions are equivalent to the usual Van Laar equations* if the temperature is constant* His equations are probably superior to the usual Van Laar equations for constant pressure data at low pressures sine© the tempera­ ture factor is taken into consideration* Theoretically it should be 25

possible to caleulat© these constanta from the oritleal constants of the pure components* but according to White these methods are not reliable*

In addition* the term# "R|gM and a**« not sufficiently defined except that they are the Van der Waale constants for the mixture*

Partial differentiation of Equation (9b) with respect to temperature at constant composition yields

/9 A ji) =/s-A/>) _/EA _i°) V 9-T / X t ~ ' & T J X , 1 ^ 7" h% (* D

The change of fugacity with temperature and hence the right hand side of

Equation (4l) can be related to heat quantities by the method of Lewis 25 and Randall whieh yields an expression for the variation of activity

coefficients with temperature* a i r , ( u f — /fc , — f r 1 =■------* 7 - *

where

HP Is the molal enthalpy of component 1 at the temperature of tne solution but at a pressure sufficiently low that it behaves as an ideal gas*

TT is the partial molal enthalpy of component 1 in th© solution and at the temperature of the solution with respect to the same reference state as the molal enthalpy of the gas* H|*

H? Is the partial molal enthalpy of pure component 1 or at a state of infinite with respect to component 2 referred to the same reference state as the molal enthalpy of the gas* HJ*

0 ^ “ H |) 1b the partial molal enthalpy relative to th© pur© com­ ponents at the temperature of the solution or the differential heat of solution*

The differential heat of solution is described as that quantity of heat

involved when one mole of the component is added to such a large quantity 2b

of the solution that the overall concentration remains essentially constant* According to Hougen and Watson*^ the use of Equation (4g) for calculating the effects of temperature is complicated by the fact that values of (Sfj “ Hj) vary considerably with temperature and are rarely known over any extensive range* 22 Jones, Sehoenborn, and Colburn attempted to compare the theo­ retical effect of temperature ae defined by Equation (4g) to the ex­ perimentally determined deviations of activity coefficients or more

specifically to Van Laar constants with temperature for the system

Bthanol-Water* The results were inconclusive partly due to the narrow temperature range ( 50 ° 0* to do® 0*) studied and partly due to the lack of reliable heat of solution data* In reality it seems more logical to determine heat of solution data from activity coefficient data since the latter are more easily and reliably determined*

Among the more recent attempts at correlation of the effects of 4 temperature on activity coefficients is that of Berg and McKlnnie who developed a purely empirical relation in straight line form at constant

p**iU o» , T p ) 0 i * S = K - x 7 5 < * » where

£ is an empirical constant*

Tjj is the reduced temperature of the solution, T/Tc solution*

A plot of log Y at constant composition versus iiiM 0,43 should Tr be a straight line with a slope of K and an intercept of zero* This inter­ cept will be zero because at the critical point of a solution th© 25

properties of liquid and vaper are indi stinguishable giving an activity coefficient of unity whose logarithm is aero* At the same time the T ti function is sere because is unity at the critical point* However when this function was tested on experimental data only one of the systems presented by the authors gave a straight line in accordance with Equation

(41). It is believed that if the temperature range of this system was extended to the lower it would show curvature also. Obvious­ ly a correlation of this type is not very beneficial if extrapolation over a relatively wide range of temperatures is not possible without extensive experimental data. j

Meries and Go I bum3® sailed the term of the A. Benedict equation *2^% a correction factor* Seheibel ' presented an empirical method of calculating values of 2 as a function of the vapor pressure, erltioel pressure, and critical temperature of the pure com* penent and the total pressure on the system* Th© equation is

where is th© vapor pressure of component 1, is the total pressure*

PG is the critical pressure of component 1, and T^ Is th© reduced temperature T/Tq for component 1. It is somewhat involved so a nomograph was presented for rapid evaluation. This correction is Important at higher pressures, but Is also significant when differences between the boiling points of the components of the mixture are large so that one or more of the components will be present in the liquid at a temperature con­ siderably different from its own boiling point* 2$

An approximate method for estimating the effect of temperature on aseetropee ie given by Carl sen and Golburn^* It is assumed that the ratio of activity eeeffieiente V* ie independent of temperature at

& given concentration* Then from Equation (10) at th® az©©tropic con­ centration where x^ * y^#

p=>* (JL) ( 4 5 } V Ye. A p * Then by plotting against x from data at a known temperature and 02 1 on th e earns s e a ls a s P3*°t F j/P j a g a in st tem perature* th e azeo tro p io concentration can be found that will give the same value for th© ratio i ' / / any temperature desired*

The following empirical equation may then be used to estimate the total pressure corresponding to the aaeotropio concentration estimated

the preceding method! / (X,P, + ?*?*) where

PazfP - —vapor pressures of the * component 1* and com- 2 ponent 2 at temperature t*

~“vaPor pressures of the azeotrop©9 component l.t and oom- ^ ponent 2 at the temperature t 1•

x x —mol© fractions of components 1 and 2 in the aaeotrope 1 ^ at temperature t*

x1, x* —mole fractions of components 1 and 2 in the azeotrop© ® at temperature t 1*

From the azeotrope compositions and pressures estimated in thi® manner corresponding constants for the Van Laar equation may be gotten from Equations (^2) and (2?)* In this way activity coefficient and 27

vapor-liquid compositions may be estimated over limited ranges of temperatures and pressure©* This method ha© not been sufficiently explored to estimate the errors involved, but they would seem to be many and serious in view of the assumptions involved* Inspection of experimental data from the literature shows that '^ / ' f at constant

composition le not independent of temperature* However* if Equation

(4j>) were modified u sin g Equation (J6)* i t ©on b© seen th a t

An extension of this idea to th© method of Carlson and Colburn just presented might be of value in estimating more correctly th® effect of temperature on azeotropee*

D. OTHER CORRELATIONS

Gilmont and co-workers^* presented a method of correlating vapor- liquid equilibrium data by relating the relative *

to composition by use of an empirically determined power series* Their claim is that this method eliminates the necessity of knowledge of total pressure for isothermal data or boiling points for constant pressure data* Whatever virtues the method might possess are overshadowed by the feet that it is purely empirical, tedious, and it departs from th© funda­ mental equilibrium concepts* 35 Redlich and Kister In a review of th© thermodynamics of non- eleetrolyte solutions developed a method of algebraic representation of 28

activity coefficient data baaed on the excess free energy# PS# of

Seat ©hard as defined by Equation (17). For one mole of mixture# the following can be w ritten« E —------=. j£ . Y] 4~ j?<>q Z ,3 o>? 3 &fcT T ' f t

If the left aide of Equation (48) ie termed Q and differentiated with reepeet to x^* one obtain*

~nrd z f = ^*?ir 0 V k £4s>) Since according to Equation (48) Q equals zero at x equal aero X or u n ity

/ <*4 - I ' - h (5o ,

The authors then developed suitable power series of mole fractions of the more volatile component to represent the data and satisfy the foregoing requirements. An extension of this method to ternary eye toms was presented.

This method permits smoothing of experimental data when calculated in th© fo m log /'j'* sine© in accordance with E**ufttion ( 5 0 ) the posi­ tive and negative areas of a plot of log vereus must be equal.

The power series is again an empirical method and a digression from the fundamental thermodynamic relations involved*

S. CONCLUSIONS

In the foregoing discussion it has been shown that the methods used in the past to correlate vapor-liquid oquilibrium data have some serious lim itations brought about by unjustified assumptions in development of the methods# applications under conditions where they are not valid# or 29

complete obscuring of the subject by purely empirical methods which are useless without extensive experimental data. In n© ©as© has a broadly applicable method been presented to accurately relate the variations encountered in studies of non-Ideal liquid solutions with properties easily determined*

It appears to the author that it would be better to confine the considerations of the subject to the original fundamental relationehipe and use theee directly rather than use the numerous empirical methods presented which have overstressed the idea of referring everything to the ideal system by use of correction factors*

Th© following equation ie proposed for correlation of activity co­ efficients and vapor-liquid equilibrium* The component parts of th© equation are not new ideas* but the arrangement of th© equation and the method of application is different and more direct than the method© previously presented.

Equation (9b) can be written for each component of a binary mixture as followsi /7

Th© fugacitiee t ^ 9 f®* f # and can be evaluated from a knowledge

Of the total pressure of the system* temperature of the system* and 17 properties of the pure components as outlined by Hougen and Batson *

The evaluation of these fugacitiee Involves th© assumption that the fugacity coefficient as determined from reduced temperature and pressure 3©

generalised chart b ie correct*

The fugacity of each component f 9 or t * in solution ie given by 1 2

= n H i (51) fz =' ^ TT $£ where

- the fugacity coefficient of each component evaluated from

the generalized charts of Hougen and Watson at reduced

temperatures oorreepondlng to the temperature of th©

solution and reduced pressure corresponding to the total

pressure* 77".

7T - total pressure.

y - the mol© fraction of each component in th© vapor which is

in equilibrium with the liquid of interest*

By uee of Equation (7) the fugacity* fQ> for ©a eh component as a pure liquid at the temperature and pressure of the solution can be evaluated* Equation (7) ©an be rearranged to the form - V’mA T T-P ,) f,°- fp, e —*r-

Vhtz O r- . Si -- fprZ. e ^ <5 } where all terms were previously defined* The fugacity of the pur© liquid under its own vapor pressure at the temperature of th© system, fp* ie evaluated by uee of th© fugacity coefficient a© read from the g en eralised ©hart a t a reduced tem perature corresponding to th© tempera** ture of the solution and a reduced pressure corresponding to th© vapor pressure of the pur© component at that temperature* Then 51

f p , = ^ /? (55 ) Substitution of all these value® in Equation ( 9b) gives

^ - ^ r f ‘ ~ ~ z X , M

y _ a/ t Fz T T t y i . ____ ? lc 7 ~ F 2 ^ t W S ^ S l and dividing yj by Jr^ yields^- & ^ V n-C rr-P j TZt J L = e (55) T j. “ . y M ~ ~u*,,(jr^3 ) V *- / 'P f "V7T^ £ RT Taking th© logarithm of both ©ides and grouping like terms

P ^ _ £*-« HiL^-h + &t\ ~^V +Uw-JjT- %)_ Mi, (w-R) ^ ^ X, d f?t difri *** ** 2-™Rrm

The following observations o&n be made about Equation (^6)s

(1 ) Log becomes equal to Log Yt at s 0, eino© ^ is

unity at that point, and similarly Log Y k -- - L e g i s t

%x * 1*0.

(2) At some value of x^, ^ and Log = 0*

(5) At th© azeotropic point » y^ and s yg, hence

Log y x ~a JZ z 0 at that points* x y 1 2 x (4) It should be noted that * 2 is th© , C&. V a and tb© limit of Log y^xg as x^ approaches sero la of th©

*ly2 in d eterm inate form Log 0 . 75 ?2

According to L1Hospital1e Rule r . - * * % o 0 if it exists*

Hence

^ a f e £ j k t _ C J _ X y.-o y,-*o 7/ malting the substitution Xg s (l-x^) and yg ** (l~y^)*

By Henry's Law for dilute solutions at or very near x, a 0*

? / = * * /

/ o ^ \ / r ^<*2/ /£, = o This value* K* is the slope of the x-y diagram at x^ s 0. Substitu­ tio n gives

x£o A-~ If-**-* 9 g ,- + o

A line with a slope equal to this limit drawn through - 0 should be used as a guide to the approach of the equilibrium curve to the aero point on the x-y diagram*

(5) According to the development of Redlieh and Klster

i L = O O ^ cf v fig (6} All the terms on the right side of the equation except log J2 & can be oaleulated from a knowledge of the properties of the V a pure component coupled with knowledge of the temperature and total pressure of the system* GHAPTm 111

AH EQUILIBRIUM STILL FOR PARTIALLY MISOIBLS

A. RESUME1 OF PREVIOUS DESIGHS

The object of ell vapor-liquid equilibrium still a Is to obtain

for analysis sufficient samples of vapor and liquid which are in

equilibrium* Making such determinations ie not easy and to obtain

reliable values requires a highly developed laboratory technique* There

are several methods for experimentally determining vapor-liquid equilibria,

but the most widely used and generally most satisfactory method is by

circulating the vapor through a system and repeatedly contacting it with the liquid* Circulating type fall into two categories* (a) those

in which the vapor ie generated from th© boiling liquid and subsequently

condensed and returned as liquid to th© body of th© boiling liquid, or

(b) those in which the vapor is generated elsewhere in the Bystem and

subsequently contacted in the vapor state with the liquid, then recir­

cu lated •

Those stills that fall into category (a) are best illustrated by

the well-known Othmer still which is shown schematically in Figure 8*

This type of reoireulating still has two serious drawbacks in that the vapor leaving the surface of the liquid (A) may not b© in equilibrium

with the main body of the liquid in the still* Also since th® liquid being returned from the condensate trap (0) is different in composition

from the liquid in the boiler, It might flash due to its lower boiling point unless mixing is instantaneous*

53 " 8 ta HP A PD TAPPP

Til t u p to DPA/m TO Co n d p a /SPP .___ ! P~$UGHT SLANT POP DPA/N AGP

> ~ '^ F 'B p n d Cl o s p a s P o$$/b l p T.

/ i

CO

CvJ

/0A O P P N /N G ->MALL VPNT

B o d y - 8 0 0 m l . k j p l d a h l Fl a s h - W it h in -g in . o f \

P l u g o f Co c k \

■t OP£N!N6.

Standard cock 3mm.

OTHMER STILL

F IG U R E 8 J0NES' SC,0£B!« . C o l b u r n s t , l l

F IG U R E 9 The still of Jones, Schoenborn, and Oolburn was on© of the first of the type defined by Category (b)* This still is schematically shown

In fig u re 9 where a liquid representing the composition of the vapor floras from the condensed vapor receiver (A) to the flash vaporiser (B) where it is totally vaporised before bubbling through the liquid in chamber (C)* The vapors are then condensed in (D) and recir* culated* This type still seems to satisfactorily circumvent the undesirable characteristics of category (a)# but oar© meet be exercised to prevent of the vapor bubbling through the liquid* r©fluxing after it leaves the liquid* and entrainment of the liquid with the vapor passing through*

There ©r© two major difficulties encountered in studies of partially aiscible systems in addition to those usually inherent in equilibrium stills* The first is that the vapor from any but the most dilute samples will* on condensing, form an Immiscible mixture* Therefore the recircu­ lation type of apparatus has not been used in the past for immiscible mixtures* since the condensate on separating into two layers cannot be returned to the still with the two liquid layer© in proper proportion* tftille a stirrer might be utilised there may b© some question with the usual types of stirrers as to whether or not mixing was thorough enough to cause the heterogeneous mixture to act litre a homogeneous phase*

Tests conducted by this investigator under circumstances very similar to those encountered in equilibrium stills, showed that th© two phases could not be satisfactorily circulated after ©ven th® most thorough agitation with laboratory type stirrers* 57

4 ? Stockh&rdt and Hull eliminated recirculation merely by distilling off* small quantities from a large quantity ©f a mixture of known com­ position after first refluxing in a tilting condenser arrangement, but this method involves slight errors caused by differential condenser hold­ up* and in addition encounters the second major difficulty discussed below*

The second undesirable characteristic offered by these partially miecible systems lies in the great difference In composition between th© v&per and liquid in the miecible region and the small concentrations of the dilute component in the liquid* For example in the aiscibl© region

©f n-butanol in water* Figure 16* which extends to about two mole percent n—butanol* th© vapor is from fifteen to thirty times as rich in n-Butanol as the liquid* Therefore* If an equilibrium study is undertaken where a

liquid sample is distilled* its composition with respect to the dilute component will change very rapidly and arrival at the desired steady state

conditions In the still is virtually impossible* 44 Th© still proposed by Smith and Bonner with minor variations ouch as a magnetic agitator* and fixed sample take-off device seems essentially th e same in p rin c ip le as th a t of Stookhardt and Bull and w ith th e same inherent errors* They found the vapor evolved from the boiling liquid to be appreciably superheated* so the boiling points reported were taken on an ebulllomater independent of the vapor-liquid equilibrium d©tormina­ tio n s* 7 Oolburn* Schoenborn* and Shilling proposed a still where vapor was generated from sep arate b o ile rs o f pur© components and mixed in th e vapor state to form a mixture of th® desired composition* Th© amount© of vapor generated from each boiler was controlled by the heat input to each* 38

This Taper was babbled through the liquid which changed to th© equilibrium composition* The vapor was condensed on leaving and analysed* This still set-up is not advantageous regardless of the accuracy of the results obtained because of the large consumption of reagents which are veiy difficult and expensive to obtain In any quantity in a sufficiently pure state for vapor-liquid equilibrium measurements*

Koraan^ presented a modification of this design whereby two streams of solution were mixed in any predetermined ratio and totally vaporised giving a vapor of a certain concentration depending on the rati© in which the two solutions were charged* A® in the still of Colburn* Schoenborn* and Shilling the vapor was bubbled through the liquid and subsequently condensed* Thie has the disadvantage of being a batch operation where the liquid has only a limited time in which to come to equilibrium* How­ ever* it has the advantage of not requiring great quantities of pure reagents as the resulting condensed vapor even if it 1® partially miacibl© can be reused as feed solution*

The still of Colburn* Schoenborn, and Shilling ae well as that of

Norman had a very good thermal insulation feature* They utilised a jacket surrounding the liquid equilibrium chamber through which the vapor feed must pass before entering the liquid chamber* Thie gives automatic

Insulation at approximately th© correct temperature since any heat loss will merely condense some of the vapor which is easily removed or re­ vaporised* This also prevents appreciable superheating of the vapors entering the liquid chamber* B* DESIGN OF THE NEW TYPE OF STILL

Since all th© © tills previously proposed for determination of vapor*liquid equilibrium data of partially raiscibl® systems have certain

lim itations and disadvantages ae previously mentioned as well a© not

being of a suitable design to permit work at pressures greater than

slightly above atmospheric* it was thought to be of considerable

importance to improve on these design©* Th© nature of this invostiga—

tion was such that a still capable of operating at pressure* considerably

above atmospheric was desired so that a wide range of temperatures could

be included in the investigation*

Griswold* Andre©, and K loin^ presented a high pressure recircula­

tion type still for miseibl© mixtures whereby the pressure in th© still was controlled by the rate of heat input and the rat© of coolant circu­

lation in the condensere* Thie new still resemble© their design only

In that th© materials of construction are th© same* it is a recircula­ tion type still* and th© rate of coolant circulation to th® condenser

is Intended to control the pressure within th© still*

Figure 10 is a detailed drawing of the new design* and Figure® 11 and 12 show th© ex te rn a l appearance of th© o rig in a l declgn* The most

important differencea in this still and any still proposed thus far are*

(1) It is a recirculation still of the type described under category (b)* utilizing a centrifugal pump with a large by-pae® line a© an agitation

chamber and as a condensed vapor receiver which should give complete

mixing* (2) It combines the good features of th© Oolburn* Schoenborn* and Shilling type and Norman type stills with th© very Important advantages of continuous operation over any period of time desired with uso of a LOUISIANA STATE UNIVERSITY — BATON ROUGE, LA VAPOR-LIQUID E.OU;L':BR:'Gm STILL 7 FOR JM Ml SC IBLE LIQ UIDS

.'HAWN BY: ORR SCALE: 1/2"- l" FEBRUARY 10,1950 1/8” Std. Pipe Nipples

Drilled For 3/16" Stove Bolts

Upper Side Drilled a Tapped For 1/4* $td. pip« Nipple DETAIL "B

Under Side Drilled For 5 /8 “ Boiler G auqe G lo ss (See d etail “A") Vent Or Pressure Gauge

; j/4" Copper k- LEGEND N Tubing-^ 7 , VAPOR RECEIVER BY-PASS LINE NEEDLE VALVE VAPORIZER THERMOWELL VAPOR JACKET VAPOR INLET Rubber Washers EQUILIBRIUM CELL (For Com pression)^ THERMOWELL CONDENSER 3/8 Cap Screws SIGHT (8 Required) 1/4" Copper Tubing 3/16" Stove Bod (4 Required) DETAIL “A” 2" Std. Welding Cap (See Detail Drilled For 3 /8 “ Std. Pipe 5/8" Boiler Gauge Glass

1/4" Needle Valve

Ferruled Fitting Centrifugal Pump

J/4" Copper Tubing (Brazed To Vaporizer)

Outer Assembly—3/8* Std. Pipe ft Fittings Inner A ssem b ly — 3/8 * 2 0 G o u g e Tubing With Nichrome Spiral Inside

FRONT AND SECTIONAL VIEW

1/8" S td . S a m p le Coch FIGURE 10 OBLIQUE VIEW OP ORIGINAL DESIGN

PIGllRE 11

Litnwn

J 42

EBONT VIEW OB OBIGINAL DESIGN

PIGDKE 12 4 5

definite small quantity of reagent b»

The condensed vapor receiver (A) is a bronze centrifugal pump with explosion proof motor manufactured by the Eastern Centrifugal Pump

Company* This particular pump id equipped with a packing gland specially designed by the pump manufacturer to permit pumping of alcohols* The materials handled during this investigation are primarily alcohols end water, the only exception being n-hsxanc* A mechanical seal was used at

first, but failed due to lack of sufficient liquid head to keep th© seal

lubricated and cooled* A relatively large fey-paes line {B) is used on the pump to permit continuous circulation and agitation of the condensed heterogeneous vapor while th© desired quantity is withdrawn and fed to the total vaporizer (D) through the needle valve (0)« This needle v&iv©

constitutes a major control device during the operation of th© still*

The vaporizer Is constructed of three-eighths inch standard pip© and fittings with a section of twenty gauge three-eighths inch copper tubing forming & concentric cylinder inside thus forcing th© incoming

liq u id downward around th e hot w all, where i t Is vaporized* The vapor rises through the center of th© tubing where the nichrom© spiral picks up superheat from the vapor and transfers it back to the cold entering

liquid through the tubing wall* The heat is furnished by resistance ribbon externally wound through which a controlled current is passed*

The current is controlled with a MVariactt controller* As the vapor

leaves the vaporizer its temperature Is measured by thermocouple (ft) before passing through the adiabatic jacket feature (F) to enter th© liquid chamber (H) by bubbling in through tube (3). Ae the vapor bubble© through the liquid there is materiel transfer until suoh time ae the liquid and vapor compositions have come to the equilibrium values* The temperature of the equilibrium vapor leaving the liquid is measured by the thermocouple (I) before passing to the condenser section (J}* Both thermocouple (ifi) and (i) are specially calibrated copper-constantan couples connected by a double pole* double throw switch to common leads which sire connected to a Leeds and Norihrup,

Typo *K* potentiometer capable of determining voltages to 0*00001 volts*

Both cold junctions were immersed in a ice bath as a reference point*

The condensed vapor then passes back to the condensed vapor receiver*

Th© sight glass (K) is provided to observe the operation of the still*

Daring the atmospheric pressure runs the still was vented through a con­ denser connected to th e Y above th© sig h t g la ss as shown in Figure 10*

The vent condenser was used to prevent vapor losses from th© still*

During pressure rune a pressure gauge and vent replaced the condenser a© seen in Figure 15*

This particular model was conetruotod of standard pipe sis®, schedule

40 iron pipe and fittings wherever possible* Copper tubing and ferruled fitting® found application in the feed tub© from the needle valve to th© vaporiser, thermocouple wells, inside tub© of th® vaporiser, and inlet to the liquid chamber* Th© sight glaee is five-eights inch boiler gauge glass mounted in special flange© a© shown in detail© BAP and pBn in

Figure 10*

Corrosion presented some problem, but this factor could bo eliminated by proper choice of metals of construction depending on the typ© of system *5

to be studied* Stainless steel would probably be satisfactory for most systems* A cheaper construction with a silver plated interior might also be expected to be satisfactory*

Any nonvolatile material that might have been left in the vapor receiver had a tendency to be deposited in the vaporizer by the pump and was taken out of either of the samples to be analysed* This per­ mitted sufficiently good an&lyaea to get reliable data on the systems studied*

Assembly ©f the equilibrium chamber body is possible by a flange assembly welded to the outer jacket and machined so as to form a smooth pressure and leak tight joint with a suitable solvent resistant gasket*

The flanges are joined by eight three-eighths inch eap screws* The

inner equilibrium chamber is easily inserted by screwing into the cap at the top of the vapor jacket* Thermowell (I) is easily removed for

charging or sampling*

This construction is of a simple and rugged nature which is a 4 distinctly desirable characteristic for equilibrium stills since most are of some intricate glass design and must be prepared by an experienced glass blower* Since the material used is metal the usual problem of breakage is not a major one*

0* OPERATING PROCEDURE A I® CHARACTERISTICS

The operating procedure and characteristics are Included since the development of the still wee not considered comp lot© until it was shown that its mechanical operation was satisfactory and that it would give reliable data on both aiecible and immiscible liquid systems* 46

To sot the still in operation for the rune at atmospheric pressure a sample of the mixture to he studied was charged to the condensed vapor receiver in a quantity sufficient to bring the level into the eight glass with the needle valve closed and the pimp running* Similarly a suitable sample was charged to the liquid chamber by removing the thermo­ well (l)# Current was then applied to the vaporiser until it became suf­ ficiently hot to vaporise any feed from the condensed vapor receiver*

The coolant flow was started to the condenser section and the needle valve opened en© quarter to one half turn and subsequently adjusted to hold the level in the vapor receiver approximately constant* This level will change slightly until such time as equilibrium is attained* and there is not further material transfer as the vapor passes through the liquid sample*

The temperatures at points E and I were cheeked and the current to the heater regulated so that the temperature at point E was never higher than at point I* but so that all material entering was still completely vaporized as indicated by a constant level in the condensed vapor re­ ceiver* Once these simple adjustments were made the still was allowed to eperate until constancy of the temperature at point X Indicated there was no further ohange in composition of th© mixture in the equilibrium chamber*

An additional period of operation was allowed to be sure equilibrium was reached* At this time the temperatures were read and recorded and the samples were withdrawn for analysis* If the operation was on partially misolble liquids and in the range of concentration where th© liquid sepa­ rated into two phases it was necessary to analyze only the vapor sample since the vapor corresponding to any two phase liquid sample should be th© same* 4?

Fop the runt above atmospheric pressure it wae found necessary to modify the apparatus slightly as shown in Figures 15 and 14 by inserting a one-quarter inch gate valve (L) just after the liquid chamber in th® vapor line so that the two samples oould be isolated after a run. This wae necessitated when the liquid sample was found to flash into the con­ densed vapor receiver on reduction of the pressure previous to withdrawing the samples at the end of the run* This valve was added at the expense of the condenser section which was much larger than necessary In the original design. As a result of this change it was necessary to increase the length of the sight glass* The substitution of the pressure gauge and vent for the condenser at the vent point for the atmospheric work had been anticipated*

It was also found necessary to oool the contents of the pump a slight degree to eliminate vapor locking at the pump intake. This was done by the crude but effective method of wrapping a towel wet with cold water around the by-pass line and pump Intake line*

Operation wae begun exactly as for atmospheric operation with the vent open. When the still was operating satisfactorily under atmospheric pressure the condenser coolant was stopped. In this manner the s t i l l was filled with vapor forcing the air out through the vent. When vapor had escaped for a short while the vent was closed and the pressure allowed to build up to the desired value by proper adjustment of the current input to the vaporizer and rate of circulation from the condensed vapor re­ ceiver. When the pressure reached the desired value th© coolant was again etarted and the rate of circulation adjusted to maintain this value.

The rat© of of th© vapore filling the still controlled th© FROHT VIEW OF MODIFIED STILL

FIGURE 13 49

OBLIQUE VIEW OS' MODIFIED STILL

FIGURE 14 50

pressure within the still* This method of pressure control is advantageoue ever pressurising with an inert gas since the pressure-temperature rela­ tions ere more likely to be correct and th© solubility of inert gas in­ creases with increased pressure*

The coolant was water maintained at a temperature ©lightly lower then room temperature by a refrigerating type constant temperature bath* The water was pumped by a centrifugal pump giving oonetant flow and regula­ tion of the flow wae by a sorew clamp choke on the line to the condenser*

The criteria for equilibrium were the same as for th© atmospheric run* at which time the needle valve (0) and the gate valve (b) were simultaneously closed and the power to the vaporizer removed* This action isolated the two samples* When they had cooled sufficiently to reduce the pressure* the samples were removed as before for analysis*

D. PROOF OF THE EQUILIBRIUM STILL

Although the still was specifically designed to handle partially mieeible ©yet©me it should work equally as well on miseibl© systeme* To test the performance of the still* the miseibl© system Ethanol—Water was chosen as representative with results of many investigations being avail­ able in the literature* The most recent of these is that of Rieder and

Thompson whose data at atmospheric pressure check with that previously determined* They present smoothed values based on their own data* These values are represented by the line of Figure 15 for comparison with the experimental points obtained on the still being tested* By reference to Figure 15 and Table I it can be readily seen the agreement i© good throughout the concentration range* 51

H 60

M0MPS0N

O . E X P ERtMeNTAL

30 4 0 50 60 70 80 90 ! MOLE PEftCEtjlr ETHYL

FlGJJ■LE 15

POR-rJ ERIUN BAH. F03 ETHANOL$ M 4 52

TABLE I

EXPERIMENTAL DATA CM STHTL *WATER AT ATMOSPHERIC PRESSURE

Mole Percent Ethanol

Temperature { 0 0 * ) L iquid Vapor

97.7 0 .6 0 1 0 .5

94.5 2 .2 6 2 2 .1

87*0 9 .5 44.0

8 1 .8 2 8 .5 5 8 .5

8 1 .5 5 2 .0 59.5

8 1 .0 55*o 6 0 .7

76.4 79.0 8 1 .5

Th* n*Butyl Alcohol*Water system# which has been th© subject of 47 investigation by Stockhardt and Bill 1 and more recently by Smith and 44 Benner » was chosen as being representative of the partially misoibl® systems* Their data are shown in Figure 16 and Tables III and I? and compared with the data from the new still as shown in Table 11 and

Figures 16# 17* and 18. As oan be seen* the agreement is very good throughout the region for which the apparatus was designed—that region where the vapor and liquid separate into two phases*

The deviations noted in th© range of high concentration of n*

Butanol will be explained later by a statistical analysis# and it is firmly believed this deviation is not attributable to the equilibrium still but to th© method of analysis as will b© dieoussed later* 55

4-

i

HArajT A O SMIT i AND B » EXPERIMENTAL. ^0 30 40 50 60 70 80 90 100 MOLE PERCENT n-BUTANOL IN LIQUID

X_M

T _ i/A £ C fe - . EQU1UBEILUM . F.QR In^BU WATER 7 6 0 mm.. P ASSURE

: T o

<9

EXPERIMENTAL

9 0 MOLE PERCENT n-BUTANOL IN

POINT DIAGRAM FDR ! BUtanql —Water FIGURE 17 i I 55

TABLE I I 1 EXPERIMENTAL DATA OK n-BUTANOL -—WATER AT ATMOSPHERIC PRESSURE

1 Temp* itel« Psro*nt n-Butanol

Llauid Vapor Log ( • .9 .) JG- , r * L O g , C 2 98.0 0.59 15*2 45.70 1.6599 0 .9 2 5 ——

92.6 1.91 2 4 .7 5 54.55 1.5557 1 .0 0 7 0 .0 0 5 0 92.5 1 6 .0 5 24.4 5 .845 0.5 8 4 9 1.184 0.0754

92.7 25*5 25.1 2.540 0.4046 1.552 0.1245

92.6* 26 .5 6 24.92 2.5 8 5 0.5775 1 .5 5 6 0 .1 2 5 8

92 . 6 * 29 .1 0 24.7 2.148 0.5 5 2 0 1.595 0.1446

9 2 .6 48.80 24.75 1.544 0.1284 I .9 5 0 0 .2 8 5 6

95.2 55*7 2 6 .8 1.2 2 8 0.0 8 9 2 2 .0 2 0 0 .5 0 5 4

96.9 75.0 57.5 1 .1 0 0 O.o4l4 2 .5 8 2 0.4120

96.9 76.0 41.9 1.079 0.0550 2 .7 5 2 O.4597

1 0 1 .2 6 1 .2 45.7 1.089 O.057O 2.775 0.4455 105.1 84.0 47.2 1.019 0.0082 2 .921 0.4655

1 0 6 .5 8 6 .5 55.9 1 .065 0.0275 2 .6 9 5 0.4506

107.4 8 8 .5 6 1 .5 1.054 0.0141 2 .5 8 5 0.4121

1 07.8 8 9 .2 6 0 .9 1.011 0.0048 2 .6 9 5 0.4506

111.9* 95-7 8 O.5 1.097 0.0402 2.086 0.5195

II 5. 5* 95*1 85.4 1.091 0.0578 1.905 0.2799 114.2 9 6 .6 6 5 .5 1.009 0.0059 2.674 0.4272

116.6 9 6 .1 95*5 0.996 ------1.975 0.2951

1— Component 1

*—Theee values read from a plot of the other valuee* C0EFF1C1E

•XPfRIMEN o n S isT e

«

2 Pts. 2 Pts.

10 20 30 40 50 €|0 70 ®° 90 100 MOL^ PERCENT n-BUTAMOl! INI LIQUID; ' ' FIGURE IS I ! 57

TABLE I I I

BATA OF STOCKHARDT AND IIULL^ 7 ON n-BUTAKOL1— WATER AT

kfmmmRio pressure

Temp* Mole Percent n-~ButanoI a Log If^ Log f s t o -) L iquid Vapojr S.. flz.

96*4 0.20 4.90 51.0 1.7076 1.000 0

96*8 0.60 11.60 45.2 1.6555 1.000 0

95*7 1.20 19.20 40.5 1.6075 1.029 0.0124

95 .0 2.00 24.40 52.0 I .5052 0.995 *■■■“

92.8 42.50 25.0 1.564 0.1942 1.705 0.2517

92.9 44.8 25.0 1.476 0.1691 1.784 0.2514

95.5 50.4 26.4 1.545 0.1281 1.880 0.2742

9 6 .5 69.5 55*8 1.120 0.0492 2.475 0.3956

96.7 70.8 54.5 1.088 O.0566 2.545 0.4057

97.9 74.5 57.1 1.067 0.0282 2.650 0.4300

108.8 95*0 64.8 I .015 O.OO56 5.700 0.5682

109.6 94.5 67.7 I .005 0.0015 4.250 0.6284

111.5 96.1 75*5 1.000 0 4.660 0.6684

1—Component I* j

— LOG i- p 2 Ip 1M0LE

KfaCENff TANQ 50 60 . 70 80 90 : l

I SURE A G im T Y iCQEFPl QtEN c u FDR ALER 0: mm. iPP

SMITH AN D BONNER ONSISTEN T WITH £

2 R aimis

10 2)0 30 ; 40 50 6 0 70 80 90! MOLE) PEBGEtt n-BUTA -1QL IN LIQUID J

■I.''. 60

TABLE XV kh i DATA OF SMITH ABD BONN SIR CM n-BUTABOL — WATER

AT ATMOSPHERIC PRESSORS

Teanp. Hole Feroent a-Butanol

Liquid Vapor K Log • * i -X . f c

u o .9 5 95*0 7 4 .7 1*020 0 .0 0 8 6 5.494 0.5455

1 0 6 .8 5 9 0 .8 6 1 .2 1 .0 2 8 0 .0 1 2 0 5*552 0.5255

io 6 .4 o 9 0 .5 5 ?.S 1.028 0*0120 5.54S 0.5240

1 0 0 .8 5 81.9 4 4 .4 1 .0 5 2 0 .0 2 2 0 5*010 0.4786

9 6 .6 5 70.9 34.0 1.108 0.0449 2*5$5 0.4125

9 6 .& 69.7 33.4 1 .1 2 1 0.0496 2-555 o.4o4o

94.00 58*5 27*6 1 .2 2 2 0 .0 8 7 1 2.179 0 .5 5 8 5

95*02 4 5 .4 2 5 .O 1.462 0 .1 7 0 6 1.789 0 .2 5 2 6

95*00 45.0 24.7 1.476 0.1691 1.785 0*2512

92.70 24.8 24 .6 2 .7 0 8 0.4326 1 .5 2 1 0*1209

92*70 9 .9 24 .6 6 .7 8 3 0.6514 1 .1 0 5 0.0426

92.70 9*8 24.6 6 .8 5 2 O.8 3 9 8 1.102 0.0422

9 2 .8 0 2 .0 24.0 52.640 1.5157 1.019 0.0082

92.85 1.9 25*7 33.810 1.9291 1.020 0.0086

95*40 0 .9 16.1 4 3 .5 6 0 1.6391 1 .0 1 0 0*0045

95.80 0 .6 1 5 .0 44.820 1.6513 1 .0 0 7 0 .0 0 5 0

1—'Component X* 61

Th© results of a thermodynamle consistency test on the data of

Stookhardt and Hull (T able III) by graphical Integration of Equation

(15) o f Chapter I I a re shown in Figure 19 where the eirelee show what hog ^.should be in order to be thermodynamically c o n siste n t w ith Log

T h is t e s t shows t h e ir d a ta to be inoonsistent and therefore incorrect.

Similar tests were carried out on the data of Smith end Bonner

(Table IT) with results as shown in Figure 20. Their data are satis­ factory up to approximately 60 mole percent n-Butanol where deviation begins. Their equilibrium still is believed to be the source of their deviation since they analyzed for water content by Karl Fischer reagent* a v ery precise and accurate method*

Figure 16 and Table II show the results of similar tests on the data obtained on the new still* Again the data ere satisfactory up to approximately 60 mole percent n—Butanol, where deviation is noted* This deviation can be attributed to the analytical method which was by density measurement in the misclble range (46*3 to 100 mole percent)*

A statistical analysis was performed on this analytical method under optimum conditions. The tests were of such a nature as to give the maxi­ mum precision obtainable with the existing equipment# taking into con­

sideration the preparation of solutions# measuring their densities# reading the compositions from a plot of density versus composition# and any other hidden factors involved* This statistical analysis showed that w ith all other factors being equal any analysis could be duplicated within

0*52 mole percent 6 j percent of the time# within 0*655 mol© percent 95 percent of the time, and within 0*96 mole percent 99*1 percent of the tlmo.

On the x-y diagram where each individual point is the result of two such 6 2

analyses* each of th* above figures should be multiplied by th© square root of two to get the allowable limit for ©aoh case.

These praoiaion lim it a together with an assumed t 0*1® 0# tempera­ ture variation were transformed into the corresponding limits of the activity coefficients and th© resulting confidence limits for Log versus composition ere shown by the dashed lines in Figure 18* These confidence limits show that all the deviation of the activity coefficient curves from that expected could have been caused by the analytical method used* It Is concluded therefor© that the equilibrium still operates satisfactorily, and that no major portion of the variation observed is due to the failure of the still to operate properly* It should be noted that although the data give a smooth curve when plotted on the x»y diagram, th© activity coefficient plots may become very erratic over th e earns range*

The Ethyl Acetate—Water Syetem was studied briefly* but due to corrosion of the still only the inmiseible range was completed* This data is shown in Table 7 and Figure 21 and compared with the values 55 predicted by Winsauer from solubility data with the Van Laar equations*

The samples were analyzed by separating and weighing the phases at a constant temperature*

TABLE V

EXPERIMENTAL DATA ON ETHYL AOSTATE-WATER AT 760 am* PRESSURE

Mole P ercent Ethyl A cetate Temperature (° 0») Liquid Vapor 71.5 1*56 75»o 71.5 77.40 73.0 7 6 .0 I .0 3 7 1 .6 _L__ I

LBGEtflD wold ictdd By Griswold Solubility rimental ip 20 30 410 50 60 70 SO 90

MOLE PERCENT ETHYL iACETATE ; IN LIQUIC l

F IG U R E EH . --L

_ VAPOR—I tom ni EC XJQB- E T H Y L A DETATE t-IWAT b ;a PR ES Sl **F 6b

g* REAGENTS

The reagents ueed in all teeie were of analytical reagent grade redistilled In a four foot laboratory column packed with stainless steel helices. The middle fractions distilling at the correct tempera­ ture were taken as th© pure material in most cases* but in some cases were redistilled to give physical properties commensurate with litera­ ture values. The water came from the laboratory distilled water supply*

Table YI is a tabulation of the physical properties of the reagents u se d .

TABLE VI

PHYSICAL PROPERTIES OF REAGEKTS

Rgagyrt Present Literature P resent L ite ra tu re

Water 1*5525 1.55252 Standard 0.99707

Ethyl Alcohol l . » 9 4 1.55912 0 . 76k96 O.765IO

Ethyl Acetate 1.3700 1.5701 0.8947 0.8950

n-Butyl Alcohol 1.397* 1.5975 0.3053 0.80525

All plots of the physical properties vermis composition used for analysis were made up from data on the actual materials used rather than the literature values* 65

F . CONCLUSION

It has bean ©hown that thi©net# equilibrium ©till will give reliable value© for vapor-liquid equilibria and in addition possesses th© very

important character1stice of continuous operation on a d e fin ite ©mall

quantity of reagent© in the mieoiblo or immiscible range of composition

at pressures varying from below atmospheric to p re e euros considerably

above atmospheric*

In addition to being versatile the © till ie rugged* simple to

construct, easy to operate, and e&ftily cleaned*

These features make the still superior to any u n i t thus far pro-

poeed for the general study of vapor-liquid equilibrium* Although

this etill does not solve all the problem s involved* th© approach t©

somewhat different and w ill perhaps stimulate further development of

the ideas presented* CHAPTER IV

EXPERIMENTAL DATA

Study of the variation of activity coefficients with pressure and temperature would be greatly facilitated by th© availability of reliable data over a relatively wide range of pressures* Such data has been obtained on a binary system which has become industrially important with the advent of the synthetic fuels program* Ethanol and n-Hexane appear as products in thi© synthesis along with a host of other chemicals most of which form with each other bringing about difficulties In separation by distillation*

In a d d itio n d a ta were determ ined on th e binary ey©terns Ethanol"*

Water* Ethyl Acetate-Water* and n-Butane 1-Water as previously presented in Chapter 111*

The Ethanol used in the investigation was absolute alcohol ©old by

U* S* Industrial Chemicals* Inc** which was fractionated In a four foot laboratory column packed with stainless steel helices (Figure 2 2 ) to yield cuts having physical properties in line with published values* 25 The materiel used had a refractive index ) of 1*3594 a specific gravity (djp) of 0*78496*

Technical grade n-Hexane* containing approximately 4 mole percent impurities* chiefly methylcyclopentane* was purchased from The Phillips

Petroleum Co.* B a rtle sv ille * Oklahoma* and subsequently p u rifie d by batohwise extractive distillation with aniline In th© lab o rato ry column*

(F ig u re 22)* Aniline has been previously used ae a separating agent 66 LABORATORY DISTILLING COLUMN

PISURE 22 66

f©r these two compounds by a oountor current extractive distillation in

Columns but has not been recoramended as a laboratory method* Th© first attempts at separation were by extraction with aniline in a separatory ftmnel and subsequent distillation of the hydrocarbon layer* 19 This effected no separation* It has been reported that these compounds fo ra an aseotrop© b o ilin g a t a tem perature clo se to th a t o f pure jv-Hexane*

It is entirely possible that when the aniline is present in the ©till pot* i t e x e rts a s e le c tiv e so lv en t a c tio n fo r th© met by 1 cy c 1 op ©nt an© (B* F*

71*6° 0 *) sufficient to change ite volatility characteristics with respect © to t h a t o f n-*Hexana (B* F* 68*75 G*) to euoh an e x ten t th a t th e ace©* trope Is broken and a separation is effected* The n-Hexan© used had a refractive index (S^) of 1*5725 and a specific gravity (d^p) of 0 * 65501*

Analysis of the various samples was by measuring refractive index and by use of a refractive index versus composition plot obtained by making up solutions of known composition and reading their refractive indices* The refractometer was maintained aj 25*0° 0 * by com bination o f a heating and refrigerating type constant temperature bath and could be read t o ± 0 . 0001.

A statistical analysis was carried out to ascertain the precision of the analytical method used* This test was performed by reading the refractive indices of several solutions of known composition under the same conditions that the analyses on th© experimental work were performed*

For example, readings on any on® of these known solutions were made in the morning, afternoon, and at night and on different days* In this way it was assumed that the many random variations that might affect such a determination would be included* A minimum of six such readings were 69

made on each of five solutions whoso ooncentratione ranged from pur® n-Hexane to pur® Sthenol*

S t a tis tic a l a n a ly st# ©f th e se p re c isio n te s te showed an average standard deviation* ^ % o f 0 .8 weight percent and that the precision of the determinations did not vary significantly over the entire rang®.

However when thi® was converted to precision In term® of mol© percent* a uniform increase from 0*4 mole percent equivalent to the average standard deviation in weight percent at pure Ethanol to 1*49 mole percent equivalent to the average standard deviation in weight percent at pure

©•Hexane was found* This increase is due entirely t© the differences in molecular weight of the two components*

For the purpose of setting up confidence limit® about experimental data* each of these values should be multiplied by 2*0 to get limits in­ side which 99*1 percent of the data should fall to be entirely satisfac­ tory and multiplied by 1*96 to get limits inside which 95 percent of the data should fall to be entirely satisfactory* When the standard devia­ tion is used without a multiplier, 67 pereent of the data should fall within the limits. In the case of an x-y diagram where each point is the result of two such determinations, each of the above values should be multiplied by the square root of two to get the appropriate limits* This depends to some extent on the shape of the curve*

At first glanee those values seem rather high* but the tests were performed with th e same degree o f care as the experiment&l determinations and considerable importance is a tta c h e d to thorn. Such teeite could probably explain much of the poor data that hae been published* Such information might prevent investigator® from making claim® about their data that are 70

not wholly justified.

T hib information permits setting up confidence lim its on experi­ mental data or any data derived from this experimental data. To th e knowledge of the author only one investigator^ has published data with confidence lim its, and these were gotten by assuming a reasonable deviation* Statistical analyses are tedious and difficult to perform correctly* but the information derived from suoh tests is ample reward for the labor* No experimental data should be considered complete until such tests have been performed*

The experimental data at 2^0 mm** 595 amu* and 7^0 mm* was obtained on a glass still of the Jones* Sehoenborn* and Colburn type as shown in fig u re 2 5 * The data at the higher pressures was determined using the new type of still as described in detail in Ohapter 111* The develop­ ment of the equilibrium still made possible th© investigations over a range of pressures sufficient to show the effect of pressure on vapor- liquid equilibrium data on liquid activity coefficients* This data is shown in Tables VIT-XIII and Figures 24-44*

The confidence limits are illustrated in Figures 2? and 29* The limits plotted are the 2

FIGURE 23 72

TABLE VII 1 2 EXPERIMENTAL DATA FOR n- HEXAMS -ETHANOL AT 250 mm* Hg.

TOTAL PRESSURE

m i S R « t n-HSXANS1 T5HFE8ATURS WWP VAPOR LOG MSLfij •o.

41.6 4.5 58*5 . 85 OS .0 3 7 9 3 6 .1 6 .4 22.4 .8072 .0 1 9 6

34.5 1 5 .0 64.8 .6 8 5 2 .0 1 5 0

% »7 36.4 7 0 .2 • 3589 .1479

51.7 44.5 70.2 .2949 .1947

51.5 3 2 .1 71 *9 .2371 .2 3 6 6 50.5 5 5 .8 70.2 . 21X3 .2 6 6 1

3 1 .2 65*.6 71.9 •1325 .3 8 8 2

3 0 .? 78.4 72.7 .0 6 2 9 •5731

32.1 99.4 76*7 1 .1 6 2 7

3 2 .4 96.5 79-5 1.1686

1—Component 1

2— Component 2 75

:2

E

tjifl'T!!1?! *

m

H TI

w m C E ' i t r U H O

IJQMI WAG&AM

a 7h

3 ^ ?r& ±tB

10 2|0 3 0 A g o ; 6 0 7 0 8 0 9 0 1 0 0 MOLE PER CENT nrHEXANE IN

BOILING DIAGRAM

-HEXAN EliOrnm.l-g FIGURE .i'& 75

A

r

0 SO 6 90 IOO J i h L LIQUID

AtiTIVltY EC EFFICIENT C

FIGURI 26 n

TABLE V III

EXPERIMENTAL DATA FOB n-HEXANE^-ETHANOL^ AT mm. Hg*

TOTAL PRESSURE

MOLE PERCENT HEXANE

JMFERATURK LIQUID VAPOR L O O ^1 l o o 0 ; ° 9 , ......

58.1 1.0 7 . 7 . 7 4 4 4

50.5 5.8 3 4 . 5 .7888 .0809

47.9 7 * 2 5 1 . 4 . 8 7 4 0 . 0 0 8 1

4 8 . 0 9.0 5 5 - 0 . 8 1 8 8 . 0 4 0 2

4 8 . 0 1 0 . 1 52.2 . 7 8 0 0 . 0 5 5 7

4 4 . 5 1 4 . 6 54.5 .6 ^ 5 .0913

4 2 . 5 2 5 . 9 62.1 .5205 . 0 6 1 5

4 l . l 5 7 . 9 8 8 . 8 .2058 . 5 0 8 9

4 1 . 4 62.9 8 8 . 6 .1639 .3608

4 1 . 5 7 7 . 4 70.2 . 0 8 0 8 .5301

4 l . 8 7 8 . 4 8 8 . 6 .0852 . 5 9 1 8

4 2 . 2 90.2 7 1 . 1 .0095 . 8 8 4 2

4 2 . 4 9 8 . 2 7 8 . 4 .0022 1 . 4 8 1 1

1— Coaponeaat 1

2— -Component 2 E

IXPE FIDEN'

20 130 40 50 0 7 0 i. 8 MOLE PERCEN T n*rH£XA N E IN L

UliLifiRIUWL QJASiAM FOR f t - HEXAME'-r-ETHANOLT U ^ n r-.i

.

3 0 ; 4 7 0 8 0 9 0 XA.NE IW _ J 4 Q U J0

HE:XANf. • IMAM A T . .3.93 STENT GOhFIBEMO

0 ,j_ 10 .! 2 40 50 0 90 1C MOLE ERGE i l n-rHEX JQ U tD

Efif ICiEN T CilfiY

-i-E T 60

TABLE IX X 2 EXPERIMENTAL DATA FOR n-HEXANE -ETHANOL AT 7€ 0 mm* Hg,

TOTAL FRESSURS

M ag PERCENT n-HEXANS

TEMPERATURE VAPOR 1,03 « 1 L 0 3 I E 0 WSHE 0.

7 5 * 4 2 . 5 6 . 9 • 55 4 1 . 6 2 9 8

6 5 . 7 6 . 4 25.2 .6589 . 1 6 1 4

65 . 0 15.1 4 6 . 2 .6222 . 0 6 4 5

6 2 . ? 1 5 . 7 4 8 . 4 . 6 2 9 4 .0584

5 9 . 6 2 8 . 2 5 1 . 5 . 5 8 6 5 . 1 6 9 7

5 9 . 6 5 0 . 9 6 0 . 4 . 4 l 4 l .0954

5 8 . 2 52.1 62.9 . 2 2 4 8 .2545

5 8 . 2 62.2 65.9 . 1 5 5 0 .5454

? 6 . 2 6 6 . 6 6 5 . 7 . 1 5 6 7 . 5 7 4 7

5 8 . 5 7 4 . 2 6 7 . 5 .1014 . 4 6 0 2

5 8 . 6 75*1 6 6 . 6 . 0 8 6 4 . 6 0 1 0

5 8 . 8 8 2 . 6 6 8 . 8 . 0 5 4 6 •6085

6 0 . 0 9 6 . 5 7 6 . 7 .018? 1.1510

6 l . O 9 8 . 2 7 6 . 7 1 . 4 1 0 8

62.2 9 8 . 9 7 8 . 4 . 0 0 4 7 I .5065

62.1 9 9 . 0 8 1 . 8 . 0 0 4 ? 1 . 5 4 9 0

6 5 . 4 9 9 . 4 9 6 . 5 .0294 1 . 0 4 5 5

1— Component 1

2—— Component 2 81

8 trm

Si

U11

ERQEB 62 1

! r

&

vlCLf

EfiiML

AT T6C) 0 100

QQiLEE CURVES

0

FKBJflE m

TABLE X

EXPERIMENTAL DATA OK ir-HEXANEl -ETHANOL2 AT 1270 m . Hg. TOTAL PRESSURE

TEMPERATURE MOLE PERCENT n »KEXANE °C- LIUUID VAPOR LOO V, LOO ^Z~

$7.5* 1.9 5*5 .1717

8 5 .6 5 .1 2 2 .6 .8 7 5 6 * ------

79.7 9*5 55.9 ,6 5 8 0 .0461

76.2 16*7 46.0 .5775 .0457

74.0 27-5 54.4 *4505 .0 9 0 5

75-9 59.2 57.1 .5117 .1599

75*5 42*6 57.7 *2851 .1667

75.1 55.® 60.4 .2 0 9 2 .2 5 6 5

75*5 65.9 6 2 .0 .1599 .5191

75-5 71.1 65.9 .1069 .5945

75*5 64.4 6 8 .6 .0 4 5 7 .5711

76 .5 97.6 75*9 1.2227

1—Component 1

2—Component 2

•—Read from plot of other values* i'T liA G ii

ElfiUK E i 2 a

»

m

8 87

m m t

0 50

m tiu So

TABLE XI

EXPERIMENTAL DATA m n - HEXANS1- ETHANOL^ AT ram* Hg.

TOTAL PRESSURE

TU^SRATtmS MOLB PERCENT z^H S M g 1 *0. LIQUID VAPOR LOG «\ los (Te

89 . 0 * 3 .2 2 1 .2 •8774 .0 4 0 6

« 7 .3 4 .2 30.4 .9297 .0120

86.1 6 .4 36.1 .8397 .0 0 5 2

84.0 9 .6 44.5 .7708 ------

8 J .4 I 2 .5 46 .9 .7042 .0 0 5 0

82.1 14.9 5 2 .0 .6648

81.7 14.9 5 2 .2 .6911 .0 0 0 9

81.9 1 5 .0 5 2 .0 .6840 ------

80.6 I 8 .5 53-6 .6 2 1 2 * 0 2 5 *i

79.1 59*0 59.6 .5 9 2 6 •0626

79.7 35.9 57.9 *3738 *1004

78.8 58.6 60.4 .1886 .2762

79.1 5 8 .6 6 0 .9 .1 9 2 0 .2695

78.8 60.9 6 0 .9 .1764 .2969

78.8 70.2 60.4 .1106 •4206

79.1 91.9 6 8 .5 .0469 *9112

8 4 .3 98.7 75.9 ------1.4646

1—Component I

2-~ Comp onosit 2

♦—■Read from a plot of other values* H4 TO

&

F"

10 20 3 0 i 4;0 SO i sb 7D 8 9 0 1 0 0 MOLjE. PERCENT rj-HEKANE IN i ■ : I . ; 1 BOIUINO ! POINT . J DIAsRAM

--1 J l-r ri u FIG UfcEL fife 00 l a n i

IDWHX

¥ O C 0

:: 92

TABLE XII 1 p EXPERIMENTAL DATA CM n-HEXANE -ETHANOL AT 2510 ram* %<

TOTAL PRESSURE

’ERATURS HOLS PERCENT B-gBxm s. 0 0. «<®10 VAPOR MXS L Y'r

99.1 6.1 52.5 .8311 •0018

95*5 6 .0 34.6 .7698 *051?

9 5 .5 6.2 47.4 .8953

95»tf 10.6 48.9 >7962

92.2 24.5 57 .7 •5394

9 2 .2 57.7 61.7 .1962 .1974 92.2 70.6 61.8 .1087 •3546

92 .0 61.0 64.5 .0595 •5160

9 3 .5 . 97.4 68.0 .0056

93*4 97.9 66.0 .0104 94.1 96.4 69.4 .0026

93*6 98.5 68.6 .0026

95*5 98.6 71.4 —'— 1 .5 0 9 5

1—Component 1

2——Component* 2 a

e I

E: 9*

I "4

M 9060

NO ItT

b r i m e : E WITH

f~

0 3 . 50 . 0 . 1C T . ji-rHEXf

AO-TiVl (tIEN T

THANQLT A T 23L0 96

TABLE m i 1 2 EXPERIMENTAL DATA ON n-HEXANE -ETHANOL AT 28^0 mm. Hg. TOTAL PRESSURE

1 T©*?gRATURE HOLS PERCENT n-HBXANS

c» m v j s VAPOR W » J' 103

1 0 7 .0 8 .7 4o.o .6715 ——

101.1 12.6 47*5 .6446 ------

101.x 15*1 50.5 .6916 ------

99.0 14.9 54.0 .6499 ------

99.7 18.4 52*5 .9449 ------

96.9 26.0 95*0 .4194 .0139

9S.5 40.7 57*0 . 2 ^ 4 .0899

96.6 4 5 .2 57*0 .1902 .1241

9 8 . 7 57.0 58.4 .1002 .2146

1—'Component 1

2—Component 2 w n o i T i : <)

O i w i tH B iS 3 d 5 c 3 a r c m i

96 LEGE

TIM AT

:::±

r

lb 20 3.0 . 4 MOLE PEpdEt EXAIilE

ACTIVITY OQEEEI

. n ~ H FIGU RE CHAPTER V

DISCUSSION OF RESULTS

Th© discussion of the results of this investigation is broken into parts to facilitate proper orientation of the material in th® order of its occurrence*

A, ENVELOPMENT OF A *©RE VERSATILE EQUILIBRIUM STILL

The new type equilibrium still proposed in Chapter III Is th© result of the combination of desirable features of stills already proposed with new features necessary to properly cope with some physical limitations which had not been satisfactorily overcome previously* The most notable of these difficulties was the adaptation of a recirculating type still so that partially misoible solutions could be handled*

The use of a centrifugal pump with a relatively large by-pass line as a condensed vapor reoelvor has several advantages! (1) Complete agitation is achieved so that any vapor that separatee into two liquid phases on condensation behaves ae a homogeneous liquid phase at least insofar as return to the total vaporiser is concerned* (2) Th® pump supplies the driving force necessary for recirculation of the condensed vapor and eliminates the necessity of Buffiolent liquid head to supply this driving force* thus reducing th© overall size of the equipment* (^}

Use of the pump eliminates slugging and the intermittent operation that Is characteristic of the gravity return still of Jones* Schoenborn* and

Colburn^ until the operating conditions are properly ad justed* thus th©

100 1 0 1

•till I® more easily e©t In operation* (4) Centrifugal pumps com® in a variety of ft lees and materials of construction, bo that ohoioa of the pump ©an be suited to the type of system being handled and the volume of hold-up desired*

9ft# of a centrifugal pump as a oondeneed vapor receiver also possesses all the limitations characteristic of stuffing boxes when organic solvents are being handled* These difficulties can be overcome by proper choice of the packing material* Seme of the plastic packing materials that have recently become available are especially resistant to organic solvents*

Probably the most outstanding of these 1 b Teflon, whloh is apparently impervious to chemicals and is also very stable over a wide temperature range* Teflon was used during this investigation as gasket® for the sight glass with complete success*

The total vaporizer is necessary if a still of the type described as category (b) in Chapter III is desired* The new still is of this type# so special ear# must be exercised in order to completely vaporize all of the feed coming from the condensed vapor receiver, but not to appreciably superheat the vapor* If superheated vapor enters the equilibrium chamber, of the liquid sample will occur as this sensible heat is lost cm bubbling through the liquid* A desuperheater" was installed as a feature of this vaporizer in the form of a nichrorae spiral wound inside the innier tube through which the vapor must rise* This spiral will pick up superheat from the vapor and transfer it back to the cold incoming liq u id *

So trouble was experienced with superheating since it was found th e re was sufficient heat to totally vaporize th© feed at a temperature equal to 1 0 2

or lower then the boiling point of the liquid in the equilibrium chamber.

Thle wag true because the boiling point corresponding to the vapor ecu** position le lower than that of th© equilibrium liquid at all compositions other than the azeotroplc composition where th© two boiling temperatures ere the game* Th© installation o f a thermocouple at the entrance of the vapor to the equilibrium ie a feature not previously ueed to insure that

•uperheating of this vapor doe© not occur#

The liquid chamber was held at approximately th© correct temperature by a vapor jacket* Acy heat looses result in removal of superheat from the incoming vapor or condensation of a small portion of the vapor which

is revaporized when it returns to the vaporizer.

Controlling the pressure by the rate of condensation of the vapor had been previously employed* but not on a still where a total vaporizer wag used* For such a method to be successful* a balance of coolant rate* heat input* and rate of circulation must be attained* This balance is easily reached since at any set of conditions there ie an equilibrium pressure* This equilibrium pressure can be adjusted to the desired value by changing any one or any combination of the controlled variables*

The data obtained on the systems Ethanol**Water and n~Butano1** Wa t ar * representative of misoible and partially miscible systems* were found to satisfactorily conform to th© literature data within the limit® of the experimental errors* Therefore it was concluded the etill operated properly and data were obtained at several pressures on th© system n«

Hexane-Ethanol on which no data have been published*

In short the new still pocseeaee the desirable characteristics of

eontinuoue operation with a definite small amount of misoible or partial**

ly raiecibl© liquids at pressures ranging from below atmospheric to 10J

considerably above atmospheric* Th® unit Is of a simple* rugged de®ign* easily oonetrueted, and easily operated*

B* PROPOSED MOB IF I OAT ION OF EXISTING DESIGN

At the higher pressures some difficulty was encountered due to leeks*

Such leaks are hard to detect and are disastrous to equilibrium measure­ ments* The sight glass seemed to be th© principal source of leaks* so a modified design with a smaller sight glass of different design is sug­ gested* The volumetric capacity of the condensed vapor receiver is ap­ proximately sixty m illiliters* and when this amount i s charged it is not necessary to see the operating level* A sight glass of the type used in refrigerant lines would be sufficient to tell If there is circulation*

The operation could be easily controlled by other criteria* This type of sight glass could be directly inserted into the line by screwed fittings eliminating the necessity for the special glange construction.

Placing the pressure gauge directly into the Y connecting the con­ denser and condensed vapor receiver would minimise the possibility of leaks* Charging and removal of the vapor sample could be easily effected through the sample valve at the bottom of the pump* The apparatus could be vented by the removable thermowell above the liquid chamber*

0* EXPERIMENT/.i RESULTS

The experimental results on the system n-Hexane-Ethanol presented in

Chapter IV a re shewn to be thermodynamically consistent for the pressures

2 5 0 mm»t 595 mm«* 760 mm** 1270 ms** and 15^5 tm* data at 2^10 nun* and 2 6 5 0 mm* are not quite as reliable as that at the lower pressures* Since the data up to 1 5 * 9 «»• are consistent they are probably correct* 104

Analysis of this data shows that, the vapory-liquid equilibrium Hue is displaced by increasing the pressure from 2^0 mm* to 760 mm. as has been postulated*^ However, at a pressure of 1270 mm* the line denoting th e equilibrium compositions Is not displaced throughout the composition range and the experimental asseotroplc point is much lower than that pro- 52 dieted by the method of Gthmer and Ten Syck* The same is noted by oomr* p a ri eon o f th e d a ta a t 1545 ma* with that at 1270 mm* and the lower pressures* At first it might seem this is due to erroneous data, but as already mentioned the data are eons!stent and in all probability are cor­ r e c t within the lim its of the experimental error* However the vapor p re ssu re cu rves for these two components are seen to cross at a pressure o f 2500 am* and a temperature of 112° Q. (Figure 54—Appendix) A reversal of the volatility characteristics would be expected beyond this point* It seems this reversal begins when the conditions are such that th e vapor pressures of the two components begin to converge thus resulting in the crossing of the equilibrium curves at two different pressures* 52 The statements of Gthmer and Ten Sfcrck are thus Inoorreet as they stand and should be qualified* Grossing of vapor pressure curves Is not an unusual phenomena since several compounds that are not of the same homologous s e r ie s exhibit this property*

Be conclusions are drawn from the data at 2510 mm* and 28^0 mm* since thay have been shown to be inconsistent* An effort was made to determine the nature and amount of this inconsistency by adjustment of the activity coefficient curves for the 2^10 mm. data until they were thermodynamically consistent. This was done by first assuming the curve representing Log fr/ was correct and graphically integrating th© aibbsHPuhea equation to obtain th e curve re p re se n tin g Log ^2 that would be thermodynamically coneiBtent w ith Leg 4| (Represented by circles on Figure 4^)• This makes the con­ s is te n c y appear to be poorer than it really is einee all the inoon- eieteneies of both curves are thrown into one* To account for this the curve re p re se n tin g Log ^ wee assumed to b© correct and th© procedure rep eated to fin d Leg that would be consistent with Log (Represented by th e t r ia n g le s on Figure 4j?)* Then somewhere in between the experi­ mental curves and curves gotten by th© procedure outlined there should be cu rves consistent with eaoh other* The dashed lines on Figure 4^ represent such curves, but being consistent doer not make them correct»

However assuming they were correct* new values for the equilibrium curve were e&loul&ted and a re shown in Figure 4 6* The Inconsistency seems to have occurred between 10 and 40 mole percent n-Hexane*

The experimental vapor pressures of the as 0 0 tro p e o f n-Hoxan©~

Sfchaaol as a function of temperature were plotted on the Cox Vapor Pressure

Piet (Figure 54—“Appendix) where the behavior was the same as that for pure components even though the composition of the azeotrope changes with p ressu re* According to the method of Hutting and Horsley^ the aseotrope of the n-Hexane-Ethanol system would disappear at a pressure greater then

1 0 ,OCX) snu where the azeotrope vapor pressure curve and the Ethanol vapor pressure curve in te rs e c t*

D. CORRELATION OF DATA

Th© application of Equation ( 5 6 ) o f Chapter I I recommended pre­ viously for th© correlation of activity coefficient data, H LJ

D -i .4b . 50 70 8 ± i £ x i \ LQUID

ACTIVITY

51XAN CTHANQC j n o ^ J l c u - s 3>

N

c z ^ 3?* m ^ u S c

© ;a i 108

11 luetratad In 47 for tho ean*'t&nt preceuro data at I 5 4 5 ram*

In thie Figure all the component parto of th® equation are plotted

. . Vm ( tt ~ P) except the term -^ 0f ^ r r * 081011 of which were of the order of magni­ tude of 0.0010 but of opposite sign b o ao to cancel * All the term®

Inside the brackets wore calculated from a knowledge of the properties of the pure components, total pressure* and temperature* It should be noted that Log does not vary with composition for constant pressure data and Log _I^£k does not vary for isothermal data* Log f p t ; was determined experimentally and Log —— then calculated* Past In- ^<2 vestigetors12* ^ have developed empirical power series to express the data* but use of the above equation is simp la and convenient when broken into ite component parts and in addition adheres to the factors commonly used to express phase equilibrium*

According to Squation ( 5 0 ) of Chapter II

L o a ~ ° 3 lre which mesne the positive and negative areas under the Log 1 - versus Y cl, curve should be equal* Thie criteria ia satisfactorily satisfied since the positive area as drawn is 0*2^4 units and the negative area

Is 0*24? units*

Since it has been shown in Chapter II that

L c c j & ~ ■o and 'jli -'N h} "57. S moM o tm p m t n w QfFO‘r;Q£fiFIOH‘Nr fETcPf c0MP0SLT10N c AfcEOTfcOPlfc \ J21IA n C 0 LC 110

where is the Henry Law constant at x^ « 0 and kg ie th© Henry Law

con stan t a t Xg> "• Of and which are th© si ©pas of th© x-y diagram at

*i * 0 *a * o* This limiting slope can he used a© an aid in draw** ing the equilibrium ourve as shown in Example 1*

Example I t

Proa Figur. k j at ^ = 0, tog - tog kj * 1.10.

Thus kj, * 12*6 and a line having this slop© ie drawn

throu$» xi s 0 In Figure 48*

Similarly from Figure 47

th e n * 2 8 * 2 and a line having thie elope ie drawn

through Xg * 0 on Figure 48* The experimental data are

Been to conform to these requirements*

Accurate prediction of activity coefficients and vapor-liquid

equilibrium data from very limited data cannot be accompli shed until a

method is found to accurately predict th© terminal values of th© activity

coefficients. At present the Van Laar equations are the most commonly

used for thie purpose. The limitations of these equations have already

bead pointed out and need no elaboration*

The following method ie suggested for prediction of these end 17 values based on a method presented by Hougen and Watson but made more

accurate by substitution of fugaoity for pressure* The following equations

will hold for moderate pressures where the pressure correction terms*

9 can be neglected* fz ~ r e I ll

IQ 2 0 3 0 4 0 50 ! 60 70 80 90 PERCENT, n-HEXANI IN: LIQJID

...... v f KPFRIUFNTftl V/lPOR LIQUID RIUM DAT A f 6 r n-HEXA NE^rETHA NOtT! AT 1 5 4 5 .min. URE

E10U k E „ 4 8

T“ ~

:_4 - 112

j i - r r - ^ r , Y t =■ 2 / f y ,

_ *<}■£. T T 'O r r *Z.

Then Z£l'7T-'^7rj = *1 K

"Vx. ^ ^ Addition of these two equations gives

and sines y^ + yg s 1# and ^ J jr and are very near the same value the following can he written as an approximation*

/ ) 7 T - + - ^ X & ^ (. Z ' (1)

Rearrangement gives

f'frr, + ) ~ir — y& ?£.£/% h = is 5 ^ ------—------'‘v ^P , (2)

i I?”! As x^ approaches zero#

£ ^7 "; ^ L \ 77— u . Y, =. - _____ 5 _ < ------— X , - j> o ''z ^5)

^ r n ; +--^Jn-£ J j j - — _ ^ u » 5 ------— - f — W Y - t- p . Frcsa accurate isobari© boiling point measurements over the ©ntlro

composition range, apparent activity coefficients can b© calculated and extrapolated to x^ * 0 and * 0 at which point they are equal to the true activity coefficients* This should be more accurate than vapor**

liquid equilibrium measurement® because of the ease of making up a liquid of known composition thus eliminating an inaccurate analysis later in the process* Apparatus for accurately determining boiling points have been Ao described . If there is considerable curvature, extrapolation to the sere concentrations is facilitated by use of the White^ method wherein

(V bog Yj ) v e rsu s ^ and (T Log Yg )m°v e r s u s ^ is plotted* the postulated straight line is not always observed, but it Is sufficiently straight in the dilute ranges to permit accurate extrapolation*

Analysis of the Van Laar equations shows that the concentration at Vi which Log * 0 is given by

And since A * Log Yj at^ S O and B * Log at Xg Z 0f It follows that this concentration can be located by Equation (5)»

'3 t y y z - o (5)

This is believed to be an adequate expression for this condition* At this point the actual values of y, and l^oan be found by ue® of Equation {!)•

In many oases the boiling point curve will indicate the concentration at h which the aseotrope occurs* At this point Log ^1 *2 - q an<* kog xi &k 114

©an fed c a lc u la te d from th© equation proposed# This will give three or

four points on the Log versus x^ curve which if used in conjunction

with the specification /

©re sufficient to specify the activity coefficient curves# A check for

consistency fey integration of th© Gibbs-Duhem equation should be per­

formed and adjustm ents made#

This method was applied to the n-Hexane-Sthanol System at 1^4^ mm#

assuming a knowledge of the boiling point diagram only (Figure 56—-Chapter

I?) as shown in Example 2# End values of Log ^ * 1*11 and Log 1*60

were predieted as compared with doubtfully extrapolated experimental values

©f Log Y s 1*11'and Log^® 1#60 as seen in Figure of Chapter XT# The

predicted concentration at which Lc(j _^-^)was x^ * 0*546 as compared with

the experimental value of x^ * 0*552# The predicted value of Log * Log tg

at this point is 0 * 2 2 1 as compared with an experimental value of 0 * 2 5 0 #

To secure other points use is made of a characteristic brought out fey White's method of plotting activity coefficients* This is illustrated

In Figure 49 where it is seen that s *2 a 0*254 and T Log Yj s

T Log l£* This is indicated at x^ s % * s> . It has been found that these concentrations are correct for this characteristic but that the value© o f Log ^ obtained at this point do not conform to experimental data and will not be used# This is brought about by th© assumption of a straight line for this plot which is not oorreot# The values obtained indicate that the high values of Log Y are th© ones in question* since there are two possible values# 115

U

WHITE ik A T iP .N

A T . 15 4i IGUFIE 1X6

Example 2 1

From F igure 56 of Chapter XV the following data ie obtained at 1?45 BHtt* total pressures

*1 *2 B.P. °i

0.10 0.90 84.0

o.o4 0. 9<5 89.O

0.90 0.10 79.0

0.99 0.01 85.5

From theee data and a knowledge of the vapor preeeuree

and critical constants of the pure components the following

oan be calculated by methods presented in Chapter II*

X1

o .o 4 0.951 0.966 959 1124 0.10 O.951 0.966 IO65 1250

0.90 0.951 0.966 980 765

0.99 0.951 0.966 1150 940

Use of Equation (2) permits calculation of apparent

values of V| at the concentrations approaching x^ 3 0 ,

and an an&logoue equation for ^perm its calculation of

apparent values of !^at concentrations approaching x^ - 0 *

*1 frf ap p aren t Log apparent ^ a p p a re n t Log ^ ap p aren t 0 .0 4 9*25 #9661 0 .1 0 5 .5 2 *7^19 O.90 7*6? .8819 0 .9 9 36.80 1 ,565 When these values are plotted and extrapolated to 2 0 or xg 2 0 the end values of tog ^ and tog i^ar© obtained*

w , £ , „ -

To get the concentration at which Vj -

-jrX/ ~i - \f~fy i'C?o7 - - &■ ^ 9

■y . = £>.5^-< o ‘ t /.

Sow solving Equation ( 1 ) substituting for yields ^ c . t s / + t>.96(,) (rs+ s) - C . S-ti> O x> ) K + O. 4-5g. (7b, 5)

Y, - K - /•£ c 1

/L e $ t", ^ Le>J *~tL= C ‘

Using these values and th© end point values th® White correla­ tion can be plotted as shown In Figure 4 9 end the concentrations x *"0 5 found at which —i - —S and simultaneously (T Log ^1 } equals 0 .5 *2 *1 (T Log d^>)

. Z t ~ Q

, ' 4 ' / 1 ) - ° ^ o

( r ~ 7 L o d Yz_) ’ = /_ 3 < ^ ~ /.zo3 = £>-p3’° - £ ■ j A-z. ¥■' _ - , - < 7-^ .. j_ j"0"5! / 118

SVom Figure $£ at th® point of inter©eotion of the two line®

■S. - J j L - 0*2^ hence at ^ s 0,189 and ^ » 0,189 the logarithm *2 X1 Of the activity coefficient* are inversely proportional to their

absolute temperatures at th© two Indicated concentrations, that is9

oy - c .t* ? _ __

( L o t C r \ * , o . / r 9 How th© eurv© representing tog J[i_ can b© drawn satisfying th© fol- lowing condition© t

1* u >

2 . ^ t f , = - ^ K k - c - ' - i a

j t o a t £ . - 0 . 5 ^

h. > * = e-/ y ? 5. / Lc

•ln o e Log ^1 2 - o a t th a t point* V 2 TTr It has been observed that for several eystems a plot of Log versus Q_ x. is almost a straight line from the aero concentration having th© v X~t lower value of Log o to th© point where Log * 09 so such a line would A ~, 119

probably be a very good firat approximation* Also* th® curve connecting the point where Log 0 to the higher end value of Log ^shows quite T v pronounced curvature toward the concentration at whioh this high end value ocours* This can be seen by reference to Figure 47*

When the curve was drawn and properly adjusted according to the five points outlined* almost exact duplication of the experimental values was attained* hence the experimental curve of Figure 47 represents the predicted values and no second curve is drawn*

This method was also applied to the n-Butanol-Wat®r system presented in Chapter III with the seme good agreement being attained* Trials of this method on many systems should be carried out before definite con­ clusions as to its validity can be made* but it shows merit and is better than the methods previously used* OHAPTER VI

SUMMARY

The literature pertaining to the determination and correlation ef vapor-liquid equilibrium data for non-ideal liquid mixtures has been thoroughly surveyed• The proposed methods have been critio&lly examined and their lim itations pointed out* This examination showed a definite need for (l) better methods of determining vapor-liquid equilibrium, (2) better mothode of oorrelating and extending ex­ perimental data to pressures and temperatures other than that determined* and (5) a dependable method for predieting vapor-liquid equilibrium from data easily and accurately determined*

A new continuous, recirculation type equilibrium still has been developed that will operate satisfactorily on miscible or partially ml8oible liquid systems at pressures ranging from below atmospheric to considerably above atmospheric pressure* This still and one similar in principle were then used to determine vapor-liquid equi­ librium data on the binary systems, Ethanol-Water, n-Butanol-Water,

Ethyl A estate-W ater, and n-Hexane-Ethanol# Data was also determined at total pressures of 250 mm#, 595 » 1270 mm#, 15^5rQm** 2510 mm., and 2850 mm# on the system n-Hexane-Ethanol*

An equation for the correlation of activity coefficients derived from vapor-liquid equilibrium data is presented which is easily handled when broken into its component parts* At th© same time this equation

120 121

•hows the magnitude of the various factors influencing phase equilibrium.

An aid to drawing the approach of the equilibrium curve to the end points is demonstrated*

In addition a method for predicting activity coefficients and vapor* liquid equilibrium data for any system at moderate pressures based on the method of correlation presented is successfully applied* The only data required is a knowledge of the vapor pressures of th© components and acourat© isobarie boiling points covering the entire composition range*

Such data are much more easily and accurately obtained than reliable vapor*liquid equilibrium data*

It is sincerely hoped this work will stimulate more investigations along the same or related lines slnoe much experimental data is necessary to adequately test and improve the methods presented* SELECTED BIBLIOGRAPHY

I* Beatty, H. A*, and Oaling&ert, G*, T ests fo r th e Acour&cy o f Vapor-Liquid Equilibrium Data, Ind. Eng. Chom.. ^ 904 (1994).

2. Beebe, A. H. Jr., Coulter, K* E., Lindsay, R. A** and Baker, E# M«, Equilibria in Ethanol—Water System at Pressures Less Than Atmospheric, Ind. Jfcg> Qhaau. jj4, 1^01 (1942).

9* Benedict, M*, Johnson, 0. A ., Solomon, £., and Rubin, L. 0., Aseotropio and Extractive D istillation—Theoretical Aspects, Trang# Alt last# Chem. Sngrs.. 4^, 571 (1945).

4. Berg, C., and KcKinnie, A. C«, The Effect of Temperature on Liquid Phaee Activity Coefficients, Ind. Eng. Oheau, 40, 1909 (1946).

9* Oar Ison, H. 0., Extractive D istillation, Trane. Am. Inst. Qhem. is sp *» fa» 767 (1945).

6. Carlson, H. 0., and Colburn, A. p., Vapor-Liquid Equilibria, lad. Sam. Ohagu. ^4, 561 (1942).

7* Oolbum, A. P., Schoenborn, &• M», and Shilling, D«, Equilibrium S till for Partially Miooible Liquids, Ind. Eng. Chem., jjg, 1250 (1949)*

6. Dodge, B. F., Chemical Engineering Thermodynamics. McGraw-Hill Book Company, I n c ., Hew York, 1944'*

9 . Duhem, P ., O ospt. Rend. . 102, 1449 (1665).

10. Gibbs, J. W., The Scientific Papers of Willard Gibbet Thermo- dyrgftwtcp, Longmans, Green, and Co., 1900.

11*. Gillespie, D. T. C., Vapor-Liquid Equilibrium Still for Miecible Liquids, Ind. Eng. Chem.. Anal. Ed., 16, 979 (1946).

12. Gilm ont, R*, Weinman, E. A ., Kramer, F., M iller, E., Hashmall, F*, and Othmer, D. F., Thermodynamic Correlation of Vapor-Liquid Equilibria, Ind. Eng# Qhem., jg, 120 (19$0)*

15* 0 1 asstone. S., Textbook o f Physical Chemistry, Second E dition, I;. Van No stran d Company, I n c ., Hew York, 1946.

14. Griswold, J., Haney, J. D., and Klein, V. A ., Ethanol-Water System— Vapor-Liquid Properties at High Pressures, Ind. Eng. Qhem., Jjg, 701 (1949).

1 22 12$

15* Griswold* J.* Andres* Q«* and Klein* V* A** Determination of High £ r^ rd 9 ^ Llquid Equiubria# ^ s * ** iMt- am* asm -

16. Hong on, 0* A** and Wateon* K. M** Chemical Process P rin c ip le s * P a rt i£* Thermodynamics. John Wiley and Son©* Inc.* Hew York* l^T T ”

17* Hougen# 0* A.# and Watson* K* M.# Chemical Propose P rin c ip le s Charts* John Wiley and Sons* Inc** Hew York* Figure© 109 and rfa* i 946.

Id* Hougen# G* A.* Phase Equilibria Review* Ind. Eng. Ctoa». 40* 1565 (1 9 4 8 ). “

19* Horsley* L. H.* table of Aaeotropee and Non-Aseotropee* Ind. Eng* Chem., Anal. S|.* 1£, 5O8 (194?}*

20. Horeley* L* H»* Graphical Method for Predicting Azeotropiem and the Sffeet of Pressure on Aaootropio Constant©* Ind. toe* Chem*. A nal. jjd .g 1£* 60 5 (19^7)*

21* International Gritioal tables. McGraw-Hill Book Go** Inc** Haw York and London* 1955*

22. Jonee* C. A** Schoenborn* £. M*# and Colburn* A* P.* Equilibrium S till for Mieolble Liquid©* Ind* Eng* Qhem*. jg* 666 (1945).

25* Lange* A* #•* Handbook of Chemietry. Handbook Publishers# Inc.* Sandusky* Ohio* 194ZT

24. Lewie* G. H** Proo* Am* Aead. fl 57* 49 (1901).

25* Lewie* G* N.* and Randall* M** Thermodynamic© and the Free Energy of Chemical Subetanooe. McGraw-Hill Book Co.# Inc.* New York and London* 1925*

26* Licht# Wa»* Jr*# and Denzler# 0. G.# Aaootropio Mixtures— Variation of Boiling Point and Latent Keat with pressure# Cbemleal Engineering Progress. 44* 627 (1946).

27* Mar gules* H.» Ana* Akad. Wise. Wien*, Hath. Naturw* Klaeee* 104. II, 1245 (1895).

26* Hertee* T. S.# and Colburn* A* P.* Binary Mixtures of n-Butane* Ieobutane* and 1-But on© with Furfural# Ind. Eng. Chom.* 59* 767 (1947). 29* Norman* tf* S*f and Hand©* 0# H* G*# Th© D ehydration o f A lly l Alcohol By Aeeotropie D istillation* Trane. Inet, of Qhem. Engre. (B ritish)* 25, 76 (1945). 1 2 4

Nutting, H* S. | and Horsley , L# H# , Graphical Method fo r Predicting Jgffeot of Pressure on Aaeotropic Systems, Ind. J&g. Qhem** Anal, Sg#, 12, 602 (lp47). ~ *

51# Othmar, D* F. , Correlating and Vapor Pressure Data, t <£&• J§2£* Ohem., jg |, 8 4 l (1940).

52* Othaer, D* F*, and Ten ISyck, £* H*, Jr., Correlating Aaeotropic B*i&* Lad, Jtog. Chem*a 4j., 2697 (19% ).

32* Othaer, D. F., and Gilmont, R., Correlating Vapor Compositions and Related Properties of Solutions, Ind. Eng. Chem.» 36, 858 (1944).

34. Ferry, J* H«, Chemical Engineers * Handbook. McGraw-Hill Book Company, I n c ., New York, Page l'jo l, 1941*

32* R ed lich , 0 ., and Kieter, A, T., Thermodynamics of Noneleetrolyte Solutions, Ind. Sag. Chem., ftp. 341 (1948).

J6 • Rieder, R. M., and Thompson, A. R., Vapor-Liquid Equilibria Measured by » Gillespie S till, Ind. 3 ^. Qhem., 4l, 2905 (1949).

37* Scatehard, G., Equilibria in Nonelectrolyte Solutions in Relation to the Vapor Pressures and Densities of the Components, Chem# R«w., 6, J21 (1951). /

36. Scatehard, G., and Hamer, W. J., Application of Equations for the Chemical Potentials to Partially Misclble Solutions, £ • Am. Chem. Sop., 37» 1805 (1935)*

39* Scatchard, G., and Prentiss, S# S., Foints of Aqueous Solutions, 2* Am# Chem# Soc*, 26, i486 (1934).

4 0 . S catehard, G*, and Raymond, C. L*, Vapor-Liquid Equilibrium, J* Am# Chem# Soc#, 60. 1278 (1958).

41. Scheibel, £• G#, Activity Coefficient Correction Factor Nomograph, In d . Eng# Qhem#. 4 l, IO76 (1949).

42. Soott, T. A#, Jr., Refractive Index of Ethanol-Water Mixtures, Journal of Physical chemistry# 5°* 406 (1946).

43. Selected Valuesof Properties of Hydrocarbons, American Petroleum Institute Research Project"44, National Bureau of Standards, 1947#

44. Smith, T. £#, and Bonner, R. F., Vapor-Liquid Equilibrium Still for Partially Misclble Liquids, Ind. Chem*, 4l, 2867 (1949)#

43# Soudere, M., Selheimer, C* W., Smith, R* L#, and Brown, G# G ., pressure—Volume-Temperature R elatione of Paraffin Hydrocarbons, 2nd. En&. Chem*, 24, 51 >22 (1932). 125

46* Steinhauser, H* H*, and White, R, R*, Vapor-Liquid Equilibria Data for Tsrnfiry i‘>ixturfts^ iInd* Enga Ghssi^, 4l» 8912 (1949). h j . Stookhardt, J. S., and Hull, 0. M., Vapor-Llquld Equilibria and Roiling Point-Composition Relation© for the Systems n—Butanol— Water and Isobutanol—Water, Ind. Eng;, Ghem*fl 2£, l4^S (1951).

48. Stull, D* R*, Vapor Pressure of Pur© Substances, Ind. Eng* Chem*, 39, 517 ( 1 9 4 7 )* * '

49* Swietoslaweki, W*, Ebulliomatric Measurements., Reinhold Publishing C o rp o ratio n , New York, 1945*

5°« Van Laar, J. J., z. Phyalk. Chemie,JS* UW >)*

51. Tan Laar, J . J*, Z.Phveik. Qhemie. 8 I 5 , 55 (1929).

52* White, R* R*, Vapor-Liquid Equilibria of Non-Ideal Solutions, Trans. Am* Inst. Qhem* Engrs*. 4 l, 539 (1945).

55* Winsauer, W. 0*, Ohu, P* L», and Griswold, J*, Phase Equilibria in Ethyl Alcohol—Ethyl Acetate—Water System, Ind* Eng* Chem*, 41. 2352 (1949).

54* Wehl, K., Thermodynamic Evaluation of Binary and Ternary Liquid Systems, Trene* Am* Inst* Qhem* Engre*. 42. 215 (1945).

55* Wohl, K*, Uber den Gaszustand Niedrigsiedender Staff©, Z* Fhyeik* Chemie* B2* 77 (1929)* APPENDIX NCSMOUTUKS

a — activity

a Tan der Waal1 e con8ta u t a — empirical constants in Equation ( 2 5 )

*12* a21 ®mP*r*,ea* constants In Equations ( 5 6 ) and (59) A — empirical constant, Van Laar or Mar gules b — Van der Waal*s constant b ^ f bg — e m p iric a l c o n sta n ts in Equations ( 56 ) and (5 9 )

B — empirical constant# Van Laar or Margulee

B — second virial coefficient

0 — third virial coefficient d^— specific gravity referred to water at 4°C*

e — Mathematical quantity# e * 2*716

f — fugaeity

f°— fugaeity in the standard state

fp—. fugaeity cf a pure substance under its own vapor pressure

F — free energy

T partial molal free energy

T°— partial molal free energy in the standard state

F°—• free energy in the standard state

F — free energy of mixing pS— excess free energy

126 — melal ©nihalpy at a pressure sufficiently low that the ideal

law s hold 0 a — partial molal enthalpy at infinite dilution

IT — partial melal enthalpy k —- Henry Lae Oonetant

K — empirical constant in Equation (4j) n •"* number of melee of any component

Hjj— refractive index p — partial pressure

P — total pressure

P — vapor p re ssu re

Pq—~ critical pressure

Pp— reduced pressure

Pee* vapor pressure of the azeotrope at temperature» t 1 . Pas- vapor pressure of the azeotrope at temperature* V q — effective molal volume

Q — arbitrarily defined quantity of Equation (49)

R — universal gas constant t temperature

T — absolute temperature

T^— critical temperature

Tp— reduced temperature v#VfV - m olal volume o f pure oomponont in th© liq u id ® x — mole fraction of any component in a liquid mixture y — mole fraction of any oomponont in a vapor mixture z — effective volume fraction 128

8 compressibility factor

Z — correction faotor defined by Squat lone ( 5 6 ) and (44)

// •— total proeeure

7T^-* reduced pressure

— activity coefficient

A?t7-— fugaeity coefficient corresponding to the total pressure, 7T~

fVtgaelty coefficient corresponding to the vapor pres sure# P tfC — relative volatility

S u b sc rip ts

1, 2, 5 Component 1, 2# or J

If hf j, 1 • Component i, h, j, 1 129

TABLE XIV

CALIBRATION DATA FOR COPF ER-CGNSTANTAN THERMOCOUPLES

(Reference Point—lee Bath at 0°C.)

THERMO- FIXED STANDARD CORRECTED BAROMETER MILU- TEMP. COUPLE POINT USED (INCHES OF MERCURY) VOLTS ° G HUMBER ______. *

B.P. Ethanol 5 0 .2 8 5 .2 8 5 76.67

B.P* n-Hexane 5 0 .5 6 2 . 6 6 ? 69.11

B.P. Water 5 0 .1 5 4.274 1 0 0 ,2 0

f •?* Sodium Sulfate 1 .2 9 0 5 2 .5 8

B.P * Ethanol 5 0 .2 6 5 .2 8 5 78.67

B.P. n-Hexane 5 0 .5 6 2 .8 5 2 6 9 . l l

B.P. Water 5 0 .1 5 4.274 100.20

B.P. Ethanol 5 0 .2 8 5.285 78.67

B.P. n-Hexane 5 0 .5 8 2 .8 6 2 69.11

B.P. Water 5 0 .1 5 4.274 100.20 m

100

BO

GO 2.00 00 JA lLLLY OUT­

JRVEj F ( QFPCRrCmSlAHI^ 151

TABLE XV

SOLUBILITY**? OF n-BUTANOL IN WATER

TEMP* Weight Percent n-Butanol in Water

°0* Water Rich Fhaae Aiaohqi tM m .

2k 7 ^ 5 7 9 .0

25 7*4o 7 9 .7 1

26 7.52 7 9 .6 6

27 7.29 7 9 .6 0

26 7.21 79.55

29 7.17 7 9 .5 0

50 7 .1 0 79.40 O E lL fiT lH S

TANO.L Q .8 5 »

L 3J 8 0

Q JL4U Q

r ■ ± U V - 1.39 4 0

Q A 2 9 0

:T

! £ L iN

F jG U R E 155

TAELS W I

SOLUBILITY2 1 OF ETHYL ACETATE IN WATER

TEMP. Weight Percent Ethyl Acetate In Watw °n. Water Rich Ph&ga Ethyl Acetate Rich Phaae

0 10.06 97.71

5 9.40

10 8.95 97*39

1? 8 .2 9

20 7.86 97.02

25 7.47 96.80

30 7.16 96.60

40 6 .6 2 9 6 .0 6

50 95-55

80 94.99 3ilri9ld

3 ISOdlMOb s a u a a a i '

a s i m i j . 3 9 o

I I E D 9 S c u o jy 0 ir ~ |9 T W 3|0J^8Q 0 01 i ^ a i s f tN 'iiS a i

f“ 155

£200

x-uL-L

2 5 0 0

0 2 4 0 0 a m : 3;S70 ixJ 0 j7 2 0 C L36QC& H

Q170QC

0.68Q C

0 6 7 0 0

0 6 6 0 0

0 6 5 0 0 3 0 4 0 0 9 0 I T p e r c i : NT n-HEX ANE 'HANOL

CQM PQS1 ^ L E R iiiS . PHYSICAL PROPERTIESi

FOR . J . . . HEXANE •lE T H A N Q ..AT.;.. SL5* G

. ! ' j ; ...... F IG U R E b 3 i VAPOR i_m. Hg.)

ro o <»

t

8

I

x it) BIOGRAPKf

V irgil Orr wa© bom in Glenmora# Louisiana on February 2, 1923 .

Ho atten d ed Glonmora grammar and high schools and was graduated from

Glenmora High School in 1940* In September of 1

Louisiana Polytechnic Institute at Rust on, Louisiana and was graduated with the Degree of Bachelor of Science in Chemical Engineering in 1944#

He was employed by Cities Service Refining Corporation of Lake

Charles# Louisiana as a shift analyst from July* 1944 until his induc­ tio n In to th e Armed Services August 22# 1944#

After a brief training period he was assigned to the 29th

Engineering Photomapping Batalllon at Portland# Oregon# and later served in Manila# Philippine Islands#

He enrolled in the graduate school of Louisiana State University

In September* 1946# and was appointed graduate assistant in the

Department of Chemical Engineering in February# 1947# He was awarded the Degree of Master of Science in May of 1946# and at the present is a candidate for the Degree of Doctor of Philosophy#

1^7 EXAMINATION AND THESIS REPORT

Candidate: V irgil Orr

Major Field: Chemical Engineering

Title of Thesis: Vapor-Liquid Equilibrium of Non-Ideal Solutions*

Approved:

Major Professor and Chairman

Dean of the rraduate School

EXAMINING COMMITTEE:

Date of Examination:

May 15 * 1950