You are on page 1of 43

Copyright © 2008, 1997, 1984, 1973, 1963, 1950, 1941, 1934 by The McGraw-Hill Companies, Inc. All rights reserved. Manufactured in the United
States of America. Except as permitted under the United States Copyright Act of 1976, no part of this publication may be reproduced or distributed
in any form or by any means, or stored in a database or retrieval system, without the prior written permission of the publisher.
0-07-154211-6
The material in this eBook also appears in the print version of this title: 0-07-151127-X.
All trademarks are trademarks of their respective owners. Rather than put a trademark symbol after every occurrence of a trademarked name, we use
names in an editorial fashion only, and to the benefit of the trademark owner, with no intention of infringement of the trademark. Where such designations appear in this book, they have been printed with initial caps.
McGraw-Hill eBooks are available at special quantity discounts to use as premiums and sales promotions, or for use in corporate training programs.
For more information, please contact George Hoare, Special Sales, at george_hoare@mcgraw-hill.com or (212) 904-4069.
TERMS OF USE
This is a copyrighted work and The McGraw-Hill Companies, Inc. (“McGraw-Hill”) and its licensors reserve all rights in and to the work. Use of this
work is subject to these terms. Except as permitted under the Copyright Act of 1976 and the right to store and retrieve one copy of the work, you may
not decompile, disassemble, reverse engineer, reproduce, modify, create derivative works based upon, transmit, distribute, disseminate, sell, publish
or sublicense the work or any part of it without McGraw-Hill’s prior consent. You may use the work for your own noncommercial and personal use;
any other use of the work is strictly prohibited. Your right to use the work may be terminated if you fail to comply with these terms.
THE WORK IS PROVIDED “AS IS.” McGRAW-HILL AND ITS LICENSORS MAKE NO GUARANTEES OR WARRANTIES AS TO THE
ACCURACY, ADEQUACY OR COMPLETENESS OF OR RESULTS TO BE OBTAINED FROM USING THE WORK, INCLUDING ANY
INFORMATION THAT CAN BE ACCESSED THROUGH THE WORK VIA HYPERLINK OR OTHERWISE, AND EXPRESSLY DISCLAIM
ANY WARRANTY, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO IMPLIED WARRANTIES OF MERCHANTABILITY OR
FITNESS FOR A PARTICULAR PURPOSE. McGraw-Hill and its licensors do not warrant or guarantee that the functions contained in the work will
meet your requirements or that its operation will be uninterrupted or error free. Neither McGraw-Hill nor its licensors shall be liable to you or
anyone else for any inaccuracy, error or omission, regardless of cause, in the work or for any damages resulting therefrom. McGraw-Hill has no
responsibility for the content of any information accessed through the work. Under no circumstances shall McGraw-Hill and/or its licensors be liable
for any indirect, incidental, special, punitive, consequential or similar damages that result from the use of or inability to use the work, even if any of
them has been advised of the possibility of such damages. This limitation of liability shall apply to any claim or cause whatsoever whether such claim
or cause arises in contract, tort or otherwise.
DOI: 10.1036/007151127X

This page intentionally left blank

Section 4

Thermodynamics

Hendrick C. Van Ness, D.Eng. Howard P. Isermann Department of Chemical and Biological Engineering, Rensselaer Polytechnic Institute; Fellow, American Institute of Chemical
Engineers; Member, American Chemical Society (Section Coeditor)
Michael M. Abbott, Ph.D. Deceased; Professor Emeritus, Howard P. Isermann Department of Chemical and Biological Engineering, Rensselaer Polytechnic Institute (Section Coeditor)*

INTRODUCTION
Postulate 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Postulate 2 (First Law of Thermodynamics) . . . . . . . . . . . . . . . . . . . . . .
Postulate 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Postulate 4 (Second Law of Thermodynamics) . . . . . . . . . . . . . . . . . . . .
Postulate 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4-4
4-4
4-5
4-5
4-5

VARIABLES, DEFINITIONS, AND RELATIONSHIPS
Constant-Composition Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
U, H, and S as Functions of T and P or T and V . . . . . . . . . . . . . . . . .
The Ideal Gas Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Residual Properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4-6
4-6
4-7
4-7

Pipe Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Nozzles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Throttling Process. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Turbines (Expanders) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Compression Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Example 1: LNG Vaporization and Compression . . . . . . . . . . . . . . . .

4-15
4-15
4-16
4-16
4-16
4-17

4-17
4-18
4-18
4-19
4-19
4-19
4-19
4-20
4-20
4-21
4-21
4-21
4-21
4-22
4-23
4-26

4-26
4-27
4-27
4-27
4-28
4-28
4-28
4-29

OTHER PROPERTY FORMULATIONS
Liquid Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Liquid/Vapor Phase Transition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4-13
4-13

SYSTEMS OF VARIABLE COMPOSITION
Partial Molar Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Gibbs-Duhem Equation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Partial Molar Equation-of-State Parameters . . . . . . . . . . . . . . . . . . . .
Partial Molar Gibbs Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Solution Thermodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Ideal Gas Mixture Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fugacity and Fugacity Coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Evaluation of Fugacity Coefficients. . . . . . . . . . . . . . . . . . . . . . . . . . .
Ideal Solution Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Excess Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Property Changes of Mixing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fundamental Property Relations Based on the Gibbs Energy. . . . . . . .
Fundamental Residual-Property Relation. . . . . . . . . . . . . . . . . . . . . .
Fundamental Excess-Property Relation . . . . . . . . . . . . . . . . . . . . . . .
Models for the Excess Gibbs Energy. . . . . . . . . . . . . . . . . . . . . . . . . . . .
Behavior of Binary Liquid Solutions . . . . . . . . . . . . . . . . . . . . . . . . . .

THERMODYNAMICS OF FLOW PROCESSES
Mass, Energy, and Entropy Balances for Open Systems . . . . . . . . . . . .
Mass Balance for Open Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
General Energy Balance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Energy Balances for Steady-State Flow Processes . . . . . . . . . . . . . . .
Entropy Balance for Open Systems . . . . . . . . . . . . . . . . . . . . . . . . . . .
Summary of Equations of Balance for Open Systems . . . . . . . . . . . .
Applications to Flow Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Duct Flow of Compressible Fluids . . . . . . . . . . . . . . . . . . . . . . . . . . .

4-14
4-14
4-14
4-14
4-14
4-15
4-15
4-15

EQUILIBRIUM
Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Phase Rule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Example 2: Application of the Phase Rule . . . . . . . . . . . . . . . . . . . . .
Duhem’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Vapor/Liquid Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Gamma/Phi Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Modified Raoult’s Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Example 3: Dew and Bubble Point Calculations . . . . . . . . . . . . . . . .

PROPERTY CALCULATIONS FOR GASES AND VAPORS
Evaluation of Enthalpy and Entropy in the Ideal Gas State . . . . . . . . .
4-8
Residual Enthalpy and Entropy from PVT Correlations . . . . . . . . . . . .
4-9
Virial Equations of State. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4-9
Cubic Equations of State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4-11
Pitzer’s Generalized Correlations. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4-12

*Dr. Abbott died on May 31, 2006. This, his final contribution to the literature of chemical engineering, is deeply appreciated, as are his earlier contributions to
the handbook.
4-1

Copyright © 2008, 1997, 1984, 1973, 1963, 1950, 1941, 1934 by The McGraw-Hill Companies, Inc. Click here for terms of use.

. . . . . . and Flash Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Liquid/Liquid and Vapor/Liquid/Liquid Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-38 4-39 4-39 4-40 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-36 4-37 4-38 THERMODYNAMIC ANALYSIS OF PROCESSES Calculation of Ideal Work. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis of Steady-State Steady-Flow Proceses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Solute/Solvent Systems. . . . . . . . . . . . . . . . . . . . Example 7: Lost-Work Analysis . Example 5: VLE at Several Temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-30 4-31 4-31 4-32 4-32 4-34 4-34 4-35 4-35 4-35 4-35 Equilibrium Constants . . . . . . . . . . . . . Lost Work . Equation-of-State Approach . . . . . . . . . . . . . . . . Complex Chemical Reaction Equilibria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VLE. . . . . . . . Chemical Reaction Stoichiometry . . . . Example 4: Flash Calculation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemical Reaction Equilibria . . . . . . . . .4-2 THERMODYNAMICS Data Reduction. . . . . . . . . . . . . . . . . . . . . Example 6: Single-Reaction Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Standard Property Changes of Reaction . . . . . . . . . . . . . . . . . . . . Extrapolation of Data with Temperature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . K Values. . . . . . . . . . . . . . . . . . . . . .

β β εj As superscripts. pressure expansion 3d virial coefficient. SE. Gi. U.j ρ σ Φi φi φˆ i psi Rate of entropy generation.THERMODYNAMICS 4-3 Nomenclature and Units Correlation. pressure expansion Interaction 2d virial coefficient Interaction 3d virial coefficient Heat capacity at constant pressure Heat capacity at constant volume Fugacity of pure species i Fugacity of species i in solution Molar (or unit-mass) Gibbs energy Acceleration of gravity ≡ GE/RT Molar (or unit-mass) enthalpy SI units J/mol [J/kg] m Dimensionless Btu/lb mol [Btu/lbm] ft2 Dimensionless cm3/mol cm3/mol 2 3 cm /mol cm3/mol cm6/mol2 cm6/mol2 cm9/mol3 cm9/mol3 kPa−1 kPa−1 kPa−2 kPa−2 kPa−3 kPa−3 cm3/mol cm3/mol cm6/mol2 cm6/mol2 J/(mol·K) Btu/(lb·mol·R) J/(mol·K) Btu/(lb·mol·R) kPa kPa J/mol [J/kg] psi psi Btu/(lb·mol) [Btu/lbm] ft/s2 Dimensionless Btu/(lb·mol) [Btu/lbm] Dimensionless Dimensionless m/s2 Dimensionless J/mol [J/kg] Equilibrium K value. identify phases Volume expansivity Reaction coordinate for reaction j Defined by Eq. S. HR. Si. (4-196) Heat capacity ratio CP /CV Activity coefficient of species i in solution Isothermal compressibility Chemical potential of species i Stoichiometric number of species i in reaction j Molar density As subscript. ∆G. Hi. Customary System units psi Symbol Definition Pisat Saturation or vapor pressure of species i Heat Volumetric flow rate Rate of heat transfer Universal gas constant Molar (or unit-mass) entropy Q q ⋅ Q R S ⋅ SG T Tc U u V W Ws ⋅ Ws xi xi psi J m3/s J/s J/(mol·K) J/(mol·K) [J/(kg·K)] J/(K·s) Btu ft3/s Btu/s Btu/(lb·mol·R) Btu/(lb·mol·R) [Btu/(lbm·R)] Btu/(R·s) K K J/mol [J/kg] J J J/s R R Btu/(lb·mol) [Btu/lbm] ft/s ft3/(lb·mol) [ft3/lbm] Btu Btu Btu/s Dimensionless m Dimensionless ft m/s m3/mol [m3/kg] Z z E id ig l lv R t v ∞ Denotes excess thermodynamic property Denotes value for an ideal solution Denotes value for an ideal gas Denotes liquid phase Denotes phase transition. Gi. GE.and application-specific symbols are not shown. Symbol A A âi ⎯a i B ⎯ Bi Bˆ C Cˆ D B′ C′ D′ Bij Cijk CP CV fi fˆi G g g H Ki Kj k1 M M Mi ⎯ Mi MR ME ⎯E Mi ∆M ∆M°j m ⋅ m n n⋅ ni P Definition Molar (or unit-mass) Helmholtz energy Cross-sectional area in flow Activity of species i in solution Partial parameter. SR. density expansion Reduced third virial coefficient 4th virial coefficient. (4-304) Fugacity coefficient of pure species i Fugacity coefficient of species i in solution Acentric factor yi Superscripts Subscripts Greek Letters Γi(T) γ γi κ µi νi. Si. pressure expansion 4th virial coefficient. Vi) Partial property of species i in⎯ solution ⎯ ⎯ ⎯ ⎯ ⎯ (Ai. Vi) Residual thermodynamic property (AR. H. ∆CP°) Mass kg Mass flow rate kg/s Number of moles Molar flow rate Number of moles of species i Absolute pressure kPa j U. G. Ui. UE. ∆S. GR. Ui. ∆V) Standard property change of reaction j (∆Gj°. denotes a heat reservoir Defined by Eq. density expansion Partial molar second virial coefficient Reduced second virial coefficient 3d virial coefficient. ∆U. Hi. HE. liquid to vapor Denotes residual thermodynamic property Denotes total value of property Denotes vapor phase Denotes value at infinite dilution c cv fs n r rev Denotes value for the critical state Denotes the control volume Denotes flowing streams Denotes the normal boiling point Denotes a reduced value Denotes a reversible process α. Eq. VR) Excess thermodynamic property (AE. ∆H.S. UR. VE) Partial molar excess thermodynamic property Property change of mixing (∆ A. V) Mach number Dimensionless Molar or unit-mass pure-species property (Ai. yi /xi Dimensionless Equilibrium constant for Dimensionless chemical reaction j Henry’s constant for kPa solute species 1 Molar or unit-mass solution property (A. cubic equation of state 2d virial coefficient. (4-151) Absolute temperature Critical temperature Molar (or unit-mass) internal energy Fluid velocity Molar (or unit-mass) volume U. Customary System units kPa Work Shaft work for flow process Shaft power for flow process Mole fraction in general Mole fraction of species i in liquid phase Mole fraction of species i in vapor phase Compressibility factor Elevation above a datum level Dimensionless lbm lbm/s SI units ω K−1 mol °R−1 lb·mol J/mol Dimensionless Dimensionless Btu/(lb·mol) Dimensionless Dimensionless kPa−1 J/mol Dimensionless psi−1 Btu/(lb·mol) Dimensionless mol/m3 lb·mol/ft3 Dimensionless Dimensionless Dimensionless Dimensionless Dimensionless Dimensionless Dimensionless Dimensionless . density expansion 2d virial coefficient.S. ∆Hj°.

internal energy finds its origin in the kinetic and potential energies of molecules and submolecular particles. J. which for systems at internal equilibrium is an intrinsic property of the system. Van Ness. causing it to retrace the initial path in the opposite direction. G. N. INTRODUCTION Thermodynamics is the branch of science that lends substance to the principles of energy transformation in macroscopic systems. and M. Thermodynamics and Its Applications. New York. and the system must eventually reach a final static condition of internal equilibrium. the existence of the primitive thermodynamic properties (see postulates 1 and 3 following) is recognized much more indirectly. J. McGrawHill. M. the system is closed (but not isolated). C. M. Abbott. Classical Thermodynamics of Nonelectrolyte Solutions: With Applications to Phase Equilibria. An isolated system exchanges neither matter nor energy with its surroundings. A system can be an object. Poling. One can equally well say that the properties of a system are fixed by its state. Work is again energy in transit between a system and its surroundings. Nevertheless. Prentice-Hall PTR. and J. Integration of Eq. Attributed to this boundary are special properties which may serve either to isolate the system from its surroundings or to provide for interaction in specific ways between the system and surroundings. Although the properties T. 1982. a final static condition may likewise be reached such that the system is not only internally at equilibrium but also in external equilibrium with its surroundings. W. New York. For such systems energy is exchanged between a system and its surroundings in two forms: heat and work. New York. from which are formulated a few postulates that form the foundation of the subject.J.. C. characterized by the macroscopic coordinates T and P. 2d ed... which is called the surroundings. Modell. P. M. Internal energy is quite distinct from such external forms as the kinetic and potential energies of macroscopic bodies.J. Note that dQ and dW. Smith. Van Ness. 3d ed. M. McGraw-Hill. On the other hand. Sandler.. analogous to that for heat. O’Connell.. such changes cannot continue indefinitely. For a closed (but not isolated) system. POSTULATE 1 There exists a form of energy. 2001. Prentice-Hall PTR. its boundaries may permit exchange of matter or energy or both with its surroundings. a quantity of matter. J. and H. The first law of thermodynamics states that energy is conserved. functionally related to the measurable coordinates that characterize the system. all forms of energy must be considered. 3d ed. M. including the internal energy.. M. H. The number of properties for which values must be specified in order to fix the state of a system depends on the nature of the system. differential quantities representing energy exchanges between the system and its surroundings. and its mass is constant. When a system is isolated. Molecular Thermodynamics of Fluid-Phase Equilibria. The sign convention. The concepts of state and property are again coupled. and is ultimately determined from experience. Prausnitz. Introduction to Chemical Engineering Thermodynamics. When a system is displaced from an equilibrium state. 5th ed. and M. it cannot be affected by its surroundings. imagined to enclose the system and to separate it from its surroundings. Like heat. R. the system is said to be open. McGraw-Hill. is called the boundary of the system. These laws are primitive. makes numerical values of W positive when work is done on the system by the surroundings and negative when work is done on the surroundings by the system. 3d ed. Chemical and Engineering Thermodynamics. New York. serve to account for the energy change of the surroundings. which continues until its properties attain new equilibrium values. the total quantity remains constant. 4-4 Thermodynamics finds its origin in experience and experiment. and V may be detected with measuring instruments. C. If the exchange of matter is allowed. E. the first law requires that energy changes of the system be exactly compensated by energy changes in the surroundings. A quantity of heat Q represents an amount of energy in transit between a system and its surroundings. Schaum’s Outline of Theory and Problems of Thermodynamics. Although it is a macroscopic property. 1997.. When applied to closed (constant-mass) systems in which only internal-energy changes occur. for associated with the condition of internal equilibrium is the concept of state. 1999. This coupling is characteristic of the primitive concepts of thermodynamics.. t . and is not a property of the system. they cannot be derived from anything more basic. Wiley. H. and E. changes may occur within the system that are detectable with measuring instruments such as thermometers and pressure gauges. 7th ed. but conversely energy is an essential thermodynamic function because it allows the first law to be formulated. and is not a property of the system. de Azevedo. I.. The Properties of Gases and Liquids. (4-1) gives for a finite process ∆Ut = Q + W (4-2) where ∆U is the finite change given by the difference between the final and initial values of Ut. Upper Saddle River. POSTULATE 2 (FIRST LAW OF THERMODYNAMICS) The total energy of any system and its surroundings is conserved. M. the system may be caused to interact with its surroundings so as to interchange energy in the forms of heat and work and so to produce in the system changes considered desirable for one reason or another. The concept of equilibrium is central in thermodynamics. 1989.GENERAL REFERENCES: Abbott. McGraw-Hill. Upper Saddle River. The heat Q and work W are finite quantities of heat and work.. N. P. selected for study and set apart (mentally) from everything else. J... pressure P. that although it can be altered in form and transferred from one place to another. B. if only energy and not matter may be exchanged. a change of state. The first two deal with energy. The convention with respect to sign makes numerical values of Q positive when heat is added to the system and negative when heat leaves the system. It is therefore clear that postulate 2 depends on postulate 1. because such a process can be reversed at any point by an infinitesimal change in external conditions. but resulting from the displacement of an external force acting on the system. a quantity of work W represents an amount of energy. Van Ness. they are not properties of the system or functions of the thermodynamic coordinates that characterize the system. the first law of thermodynamics is expressed mathematically as dUt = dQ + dW (4-1) t where U is the total internal energy of the system. and M. For an isolated system the first law requires that its energy be constant. reproducible state when all its properties. The general restrictions shown by experience to apply to all such transformations are known as the laws of thermodynamics. However. N. In applications of the first law of thermodynamics. A system has an identifiable. S. Heat is energy crossing the system boundary under the influence of a temperature difference or gradient.. The words system and surroundings are similarly coupled. Prausnitz. For a closed system which interacts with its surroundings. known as internal energy. such as temperature T. If a system is not isolated.. or a region of space. A process that proceeds so that the system is never displaced more than differentially from an equilibrium state is said to be reversible. dUt is directly the differential change in internal energy of the system.. New York. Lichtenthaler. During such a process. it undergoes a process. Thus the first law of thermodynamics depends on the concept of energy. 1999. Abbott. An envelope. 2005. Tester. are fixed. and molar volume V.

In the same way that the first law of thermodynamics cannot be formulated without the prior recognition of internal energy as a property. then by Eq. in the limit where the system has reached an equilibrium state. For reversible processes. For a closed (but not isolated) system it requires that any entropy decrease in either the system or its surroundings be more than compensated by an entropy increase in the other part. from which all other equations connecting properties of . Total-system properties (V t. pressure.). The result is Eq. The fundamental thermodynamic properties that arise in connection with the first and second laws of thermodynamics are internal energy and entropy.nj where the summation is over all species present in the system and subscript nj indicates that all mole numbers are held constant except the ith.n For an open single-phase system. pressure P. nV. U t. Equation (4-6) is the fundamental property relation for single-phase PVT systems. electric. and composition. pressure. Temperature. Equation (4-5) shows that for a single-phase.n ∂(nV) nS.n ∂(nU) d(nV) +  ∂ni i   dni nS. pressure. The second law requires that the entropy of an isolated system either increase or. such as internal energy U and entropy S. closed system. and composition. AND RELATIONSHIPS Consider a single-phase closed system in which there are no chemical reactions. The solution of any problem in applied thermodynamics is therefore found through these quantities. Such systems are usually made up of fluids (liquid or gas) and are called PVT systems. (4-5) shows that ∂(nU)   ∂(nS)  dUt = dQrev + dWrev Substitution for dQrev and dWrev by Eqs. This equation follows directly from the definition of mechanical work. are also functions of temperature. . functionally related to the measurable coordinates that characterize the system. This is quite direct for molar or specific volume V. (4-3) and (4-4) gives dUt = T dSt − P dVt Although derived for a reversible process. nV). U. DEFINITIONS. changes in this property may be calculated by the equation dQrev dSt =  (4-3) T t where S is the total entropy of the system and T is the absolute temperature of the system. different types of systems are characterized by different sets of measurable coordinates or variables. molar volume V.n     ∂(nU)   ∂(nV)  = T and nV. Applications of the thermodynamic postulates necessarily involve the abstract quantities of internal energy and entropy. These molar or unit-mass properties.. and S. represented by the plain symbols V. considered together. n2. are independent of system size and are called intensive. or gravitational) are negligible and that surface and viscous shear effects are unimportant. Comparison with Eq. Temperature. Postulate 5 affirms that the other molar or specific thermodynamic properties of PVT systems. ∂(nU) ∂(nU) Then d(nU) =  d(nS) +  d(nV) ∂(nS) nV. ∂(nU) d(nU) =  ∂(nS)   nV. approaching zero when the process approaches reversibility. The PVT system serves as a satisfactory model in an enormous number of practical applications. (4-1) where subscript n indicates that all mole numbers ni (and hence n) are held constant. we assume that nU = U (nS. These properties together with the two laws for which they are essential apply to all types of systems. are also intensive.n = −P nS. or that in the limit where the process is reversible. reversible process. pressure. so also the second law can have no complete and quantitative expression without a prior assertion of the existence of entropy as a property. Mt = nM. where important equations of this section are collected. nonreacting system. POSTULATE 4 (SECOND LAW OF THERMODYNAMICS) The entropy change of any system and its surroundings. Define ∂(nU) µi  ∂ni nS. pressure. It is equally well written as d(nU) = T d(nS) − P d(nV) (4-5) where n is the number of moles of fluid in the system and is constant for the special case of a closed. and the functional dependence of the thermodynamic properties on these conditions is determined by experiment. which can be measured. In accepting this model one assumes that the effects of fields (e. 4-5 For closed systems of this kind the work of a reversible process may always be calculated from dWrev = −PdV t (4-4) where P is the absolute pressure and Vt is the total volume of the system. n1. −P. When U. DEFINITIONS. not all of which are necessarily independent. This postulate imposes an idealization. However. and composition only. and V represent specific (unit-mass) properties. and leads immediately to the conclusion that there exists an equation of state relating molar volume to temperature. Note that n n1 + n2 + n3 + … = ni i where i is an index identifying the chemical species present. If such a system undergoes a differential. this equation relates properties only and is valid for any change between equilibrium states in a closed system. S. and is the basis for all subsequent property relations for PVT systems. Under these restrictions the composition is fixed.nV. remain constant.nV. and the composition variables. n3. the total entropy of the system plus its surroundings be constant. . For a system containing n mol of fluid. and composition are thermodynamic coordinates representing conditions imposed upon or exhibited by the system. n is replaced by m. (4-6) of Table 4-1. resulting from any real process is positive. such as mole fraction. which for systems at internal equilibrium is an intrinsic property of the system. magnetic. nonreacting. . St) do depend on system size and are extensive. where M is a molar property.VARIABLES. In consequence. The type of system most commonly encountered in chemical technology is one for which the primary characteristic variables are temperature T.nj   The expressions for T and −P of the preceding paragraph and the definition of µi allow replacement of the partial differential coefficients in the preceding equation by T. POSTULATE 5 The macroscopic properties of homogeneous PVT systems at internal equilibrium can be expressed as functions of temperature. and composition for any particular homogeneous PVT system. The equation of state is a primary tool in applications of thermodynamics.g.n ∂(nU) d(nS) +  ∂(nV)   nS. VARIABLES. nU = u(nS. and µi. AND RELATIONSHIPS POSTULATE 3 There exists a property called entropy.

These are Eqs. Additional thermodynamic properties are related to these and arise by arbitrary definition. whereas the mole numbers in Eq. (4-6) applies to a system of n mol where n may vary. It is therefore subject to the constraints i xi = 1 and i dxi = 0. (4-12) and (4-13). (4-6) are equivalent forms of the fundamental property relation. pressure. The quantity µ i is called the chemical potential of species i. (4-6) reduces this result to Eq. (4-14) through (4-17) of Table 4-1. nH. and composition. (4-22) through (4-25). The obvious next step is substitution for the partial differential coefficients in favor of measurable quantities. and S that result are given in Table 4-1 by Eqs. Eqs. (4-6). The resulting equations are of course less general than their parents. These equations and Eq. G Gibbs energy. This purpose is served by definition of two heat capacities. and S as functions of T and P or T and V ∂V S dA = −S dT − P dV (4-18) V ∂T i G H − TS S ∂P = ∂S (4-15) i A U − TS ∂P ∂T  ∂V (4-31)    − P dV V CV ∂P dS =  dT +  dV T ∂T V  (4-35) U Internal energy. Alternatively. Additional property relations follow directly from Eq. and appear under that heading in Table 4-1. 3. This provides two useful equations: dU = T dS − P dV + µi dxi i U = TS − PV + xiµi i The first is similar to Eq.4-6 THERMODYNAMICS TABLE 4-1 Mathematical Structure of Thermodynamic Property Relations Primary thermodynamic functions U = TS − PV + xiµi (4-7) H U + PV (4-8) For homogeneous systems of constant composition Fundamental property relations d(nU) = T d(nS) − P d(nV) + µi dni i dU = T dS − P dV (4-6) Maxwell equations  = − ∂S (4-14) i d(nH) = T d(nS) + nV dP + µi dni (4-11) d(nA) = − nS dT − P d(nV) + µi dni (4-12) d(nG) = − nS dT + nV dP + µi dni (4-13) dH = T dS + V dP (4-9) (4-10) ∂P ∂H dH =  ∂T ∂H dT +  ∂P P  ∂S dS =  ∂T  ∂S ∂U dU =  ∂T  ∂S dS =  ∂T V ∂U dT +  ∂V  ∂S dT +  ∂V dV (4-24) T  dV ∂P T ∂S = T  ∂P T ∂T ∂U V ∂S = T  ∂T V ∂U ∂S = T  ∂V T ∂H (4-23) T V  dP P  (4-22) T  dT + ∂P P dP ∂S = T  ∂T P ∂V ∂V (4-25) T T (4-28) ∂V + V = V − T  ∂T  ∂P − P = T  ∂T  V −P ∂V dH = CP dT + V − T  ∂T    dP (4-32) P (4-29) C ∂V dS = P dT −  dP T ∂T P (4-33) (4-30) ∂P dU = CV dT + T  ∂T (4-34) P = CV  (4-21) T Total derivatives P  ∂S P  =C  (4-20) T ∂T = − ∂P (4-17) Partial derivatives ∂H  ∂T ∂S V dG = −S dT + V dP (4-19) P ∂T = ∂V (4-16) i U. P. the Gibbs energy G is special. the terms in brackets must separately be zero. and it plays a vital role in the thermodynamics of phase and chemical equilibria. Those in common use are shown in Table 4-1 as Eqs. . however. H. mathematical operations requiring their independence are invalid. and S as Functions of T and P or T and V At constant composition. However. one at constant pressure and the other at constant volume: ∂H C P  ∂T P ∂U CV  ∂T V   (4-26) (4-27) Both are properties of the material and functions of temperature. and nG—as a function of a particular set of independent variables. Each expresses a total property—nU. The choice of which equation to use in a particular application is dictated by convenience. such systems are derived. (4-8) of Table 4-1 by n and differentiation yield the general expression d(nH) = d(nU) + P d(nV) + nV dP Substitution for d(nU) by Eq. because of its relation to the canonical variables T. this equation may be rewritten as d(nU) − T d(nS) + P d(nV) − µi d(xin) = 0 i Expansion of the differentials and collection of like terms yield dU − T dS + P dV − µ dx n + U − TS + PV − x µ dn = 0 i i i i i i Because n and dn are independent and arbitrary. Because ni = xin. nA. A Helmoholtz energy. CONSTANT-COMPOSITION SYSTEMS For 1 mol of a homogeneous fluid of constant composition. Here. (4-6) and (4-11) through (4-13) simplify to Eqs. because the mole fractions are not independent. However. and {ni}. application of the reciprocity relation for such expressions produces the common Maxwell relations as described in the subsection “Multivariable Calculus Applied to Thermodynamics” in Sec. n is unity and invariant. H. Multiplication of Eq. V can serve rather than P as the second independent variable. Another set of equations results from the substitutions n = 1 and ni = xi. The second of the preceding equations dictates the possible combinations of terms that may be defined as additional primary functions. U. H. (4-6). in which the partial derivatives are taken with composition held constant. (4-6) are. (4-7) through (4-10). called the canonical variables for the property. molar thermodynamic properties can be considered functions of T and P (postulate 5). Eq. H enthalpy. (4-11). where xi is the mole fraction of species i. because V is related to T and P through an equation of state. The total differentials of nA and nG are obtained similarly and are expressed by Eqs. Mole fractions are not independent of one another. (4-18) through (4-21) of Table 4-1. the variables of primary interest in chemical processing. The useful equations for the total differentials of U. Moreover. Because these equations are exact differential expressions.

01 0. and they show that Uig. (4-30). The coefficients of dT. (4-29). and dV are all composed of measurable quantities.02. CPig. (4-24). (4-27). Gig = Hig − TSig. For those ig ig properties that are independent of P—Uig. Van Ness. and the ideal-gas equation is a reasonable approximation. (4-28). (4-28). equations for (∂U/∂T)V and (∂U/∂V)T derive from Eq. The entropy. the ideal gas volume is given by Vig = RT/P. Eqs. independent of P and V. and Abbott. are functions of temperature only. For the entropy. (4-14).VARIABLES. and (4-31) to yield Eq. The following partial derivatives. AND RELATIONSHIPS Equation (4-15) of Table 4-1 may be divided by dT and restricted to constant P. CVig. The Ideal Gas Model An ideal gas is a model gas comprising imaginary molecules of zero volume that do not interact. Eqs. all taken at constant composition. (4-23). 4-1. and CPig. is a function of both T and P or of both T and V. and Eqs.1 0. 104. Equations (4-22). and (4-20) to yield Eq. values of which are given in Table 1-9. The most commonly used residual properties are as follows: Residual volume VR V − Vig Residual enthalpy HR H − Hig R ig Residual entropy S S − S Residual Gibbs energy GR G − Gig i ig where M can represent any of the properties listed. (4-33). Its PVT behavior is represented by the simplest of equations of state PVig = RT. (4-26). (4-21). p. (4-30) and (4-31) of Table 4-1. (4-32) through (4-35) become dHig = CPig dT CPig R dSig =  dT −  dP T P dU ig = CigV dT CVig R dSig =  dT +  dV T Vig (4-36) The ideal gas state properties of mixtures are directly related to the ideal gas state properties of the constituent pure species. where R is a universal constant. Similarly. Division of Eq. Equations (4-32) and (4-33) are general expressions for the enthalpy and entropy of homogeneous fluids at constant composition as functions of T and P. (4-25). 7th ed. and these with Eqs. (4-27) and (4-20) yield Eqs. Equations (4-34) and (4-35) are general expressions for the internal energy and entropy of homogeneous fluids at constant composition as functions of temperature and molar volume. Hig. Residual properties depend on interactions between molecules and not on characteristics of individual molecules. [Smith. Residual Properties The differences between true and ideal gas state properties are defined as residual properties MR: MR M − Mig (4-40) where M is the molar value of an extensive thermodynamic property of a fluid in its actual state and Mig is its corresponding ideal gas state value at the same T.. Because the ideal gas state presumes the absence of molecular interactions. New York (2005).] FIG. McGraw-Hill. P.98 and 1.02 1 Pr Z = 0. Eqs. yielding (∂H/∂T)P as given by the first equality of Eq. By definition. Equation (4-28) is completed by Eq. 4-1 For the Gibbs energy.98 0. (4-29) and (4-31) of Table 4-1 lead to very simple expressions for ideal gases: =   ∂V  ∂P 4-7 (4-37) Gig = yiGigi + RT yi ln yi i (4-39) i The ideal gas model may serve as a reasonable approximation to reality under conditions indicated by Fig. are obtained from this equation: ∂P ∂Vig  ∂T Vig =  =   ∂T P T R P =  =  Vig T V R P ∂P  ∂V T ∂Uig ∂Hig T ∂Sig  ∂P =0 T R = − P T ∂Sig  ∂V T R =  Vig Moreover. and (4-21) to yield Eq.001 0 1 2 Tr 3 4 Region where Z lies between 0. Hig. dP. CV . and Eq. and it provides the basis for comparison with true molar volumes through the compressibility factor Z. residual properties reflect deviations from ideality. Regardless of composition. (4-35). (4-32). whence by Eqs. Hig. (4-34). DEFINITIONS. (4-29) is completed by Eq. (4-17). and (4-29) combine to yield Eq. each weighted by its mole fraction: M = yiM ig ig i Z = 1. CVig. however. (4-37) and (4-38): In these equations Vig. Introduction to Chemical Engineering Thermodynamics. and CP —the mixture property is the sum of the properties of the pure constituent species. V V PV Z  =  =  Vig RTP RT 10 P = − Vig The first two of these relations when substituted appropriately into Eqs. Uig. and Sig are ideal gas state values—the values that a PVT system would have were the ideal gas equation the true equation of state. an alternative form of which follows from the mathematical identity:  RT G 1 dG − G dT d   2 RT RT . an additional term is required to account for the difference in partial pressure of a species between its pure state and its state in a mixture: Sig = yiSigi − R yi ln yi i i (4-38) Useful relations connecting these residual properties derive from Eq. (4-15) by dP and restriction to constant T yield (∂H/∂P)T as given by the first equality of Eq. and composition. They apply equally to pure species and to constant-composition mixtures. which is a function of both T and P. (4-26).

and the values assigned to Hig0 and Sig0 are also arbitrary. The ideal gas state heat capacity CPig is a function of T but not of P. r PROPERTY CALCULATIONS FOR GASES AND VAPORS The most satisfactory calculation procedure for the thermodynamic properties of gases and vapors is based on ideal gas state heat capacities and residual properties. GR  = RT ∂Z   ∂T HR  = −T RT R G V H d  =  dP − 2 dT RT RT RT As a consequence. 210–211. (4-17) and for G by Eq. these are given by rearrangement of the residual property definitions: H = Hig + HR and S = Sig + SR These are simple sums of the ideal gas and residual properties. 7th ed.4-8 THERMODYNAMICS Substitution for dG by Eq.  RTG V RT H RT d  =  dP − 2 dT (4-41) This equation may be written for the special case of an ideal gas and subtracted from Eq. after algebraic reduction. Note also that the integrand (Z − 1)/P remains finite as P → 0. evaluated separately. (4-41) itself. Data for ideal gas state heat capacities are given for many substances in Table 2-155. EVALUATION OF ENTHALPY AND ENTROPY IN THE IDEAL GAS STATE For the ideal gas state at constant composition: dHig = CigP dT dT dP and dSig = CigP  − R  T P Integration from an initial ideal gas reference state at conditions T0 and P0 to the ideal gas state at T and P gives Hig = Hig0 + Sig = Sig0 + C T T0 ig P C T T0 ig P dT H=H + C T T0 ig P dT + H R T T0 ig P i Cig (4-52) P = A + BT + CT 2 + DT −2 R where A. and Abbott [Introduction to Chemical Engineering Thermodynamics. C. then by difference. Thus. Evaluation of the integrals ∫ CPig dT and ∫ (CPig /T) dT is accomplished by substitution for CPig. (4-52) are  CR dT = AT (τ − 1) + B2 T (τ ig P T dT P  − R ln  T P0 2 C D τ−1 − 1) +  T 03 (τ 3 − 1) +   3 T0 τ (4-53)  τ+1 D  RCT dT = A ln τ + BT + CT +   (τ − 1) τ T  2  T (4-50) 2 0 0 T0 Substitution into the equations for H and S yields ig 0 C dT P (4-51)  − R ln  + SR T P0 The reference state at T0 and P0 is arbitrarily selected. thus the units of CPig are those of R. with V = ZRT/P. a common form is S = Sig0 + T0 ig P 0 2 0 2 2 0 (4-54) . only changes in H and S are of interest. (4-43) yields GR VR dP d  =  dP = (Z − 1)  RT RT P P Because G = H − TS and Gig = Hig − TSig. They may be put into generalized form by substitution of the relationships T = TcTr dT = Tc dTr P = Pc Pr dP = Pc dPr The resulting equations are This equation in combination with a rearrangement of Eq. Smith. yielding R R  (4-42) VR ∂(GR/RT)  =  RT ∂P  (4-43) ∂(GRRT) HR  = −T  ∂T RT  (4-44)   and P ZRT RT RT VR V − Vig =  −  =  (Z − 1) P P P  (constant T) Integration from P = 0 to arbitrary pressure P gives dP  (Z − 1)  P P (constant T) 0 0 P (4-45) dP  P (constant T) (4-46) SR HR GR  =  −  R RT RT (4-47) Equations (4-45) through (4-47) provide the basis for calculation of residual properties from PVT correlations. Van Ness. followed by integration. Differentiation of Eq. the equations that follow from Eq. B. (4-10) gives. pp. Thus. (4-44) gives GR  = RT  Pr 0 dP (Z − 1) r Pr  (4-48) Pr ∂Z HR dP  r (4-49)  = −Tr2 ∂Tr P Pr 0 RTc The terms on the right sides of these equations depend only on the upper limit Pr of the integrals and on the reduced temperature at which they are evaluated. values of GR/RT and HR/RTc may be determined once and for all at any reduced temperature and pressure from generalized compressibility factor data. The ratio CPig /R is dimensionless. For a mixture the heat capacity is simply the molar average iyiCigP . GR = HR − TSR. ig Empirical equations relating CP to T are available for many pure gases. and either C or D is zero. Of primary interest are the enthalpy and entropy. For temperature limits of T0 and T and with τ T/T0. New York (2005)] show that it is permissible here to set the lower limit of integration (GR/RT)P=0 equal to zero. McGraw-Hill. (4-36).. In practice. and T Equation (4-43) provides a direct link to PVT correlations through the compressibility factor Z as given by Eq. and fixed reference state values ultimately cancel in their calculation. and D are constants characteristic of the particular gas. (4-45) with respect to T in accord with Eq.

(4-45) gives The residual properties of gases and vapors depend on their PVT behavior. Substitution for B in Eq. and B′. and Bij. truncation is to two or three terms. (2-17). (4-67) by this expression gives P Z = 1 + (B0 + ωB1)r Tr (4-73) By differentiation. j. . They are given by the following equations: 〈CigP 〉H B C D = A +  T0(τ + 1) +  T 20(τ 2 + τ + 1) + 2  R 2 3 τT 0 〈CigP 〉S D = A + BT0 + CT 20 +   R τ 2T 20   τ+1 τ−1    2  ln τ (4-57) (4-58) B = yi yj Bij (4-60) C = yi yj yk Cijk (4-61) j (4-65) (4-66) Values can often be found for B. C. Although the virial equation itself is easily rationalized on empirical grounds. The temperature derivatives of B and C are given exactly by (4-69) Differentiation of Eq. the two-term expansion in pressure. the second is a mixture property. (4-67) yields ∂Z =  −    ∂T  dT T RT dB B P P By Eq. yj. Virial Equations of State The virial equation in density is an infinite series expansion of the compressibility factor Z in powers of molar density ρ (or reciprocal molar volume V−1) about the real gas state at zero density (zero pressure): i or three terms. truncation is to two (4-74) dB1 HR dB0  = Pr B0 − Tr  + ω B1 − Tr  RTc dTr dTr    (4-75) The residual entropy follows from Eq. They may also be reformulated to give P as a function of T and V or V as a function of T and P. for enthalpy and entropy calculations. Similarly for the third virial coefficients: Ciii. For pure chemical species B0 and B1 are functions of reduced temperture only. In practice. which is explicit in volume. (4-36). (4-48) and (4-49). Duan. are cross coefficients. (4-67) is useful. and yk are mole fractions for a gas mixture and i. Generalized correlations for both B and C are given by Meng. 3d ed. Equations (4-45) and (4-46) are therefore applicable. Moreover. . For pressures above the range where Eq. the two sets of coefficients are related: Z = 1 + Bρ + Cρ2 RESIDUAL ENTHALPY AND ENTROPY FROM PVT CORRELATIONS Z = 1 + Bρ + Cρ2 + Dρ3 + · · · 4-9 (4-70) and by Eq. Equation (4-64) is the virial equation in pressure. (4-65). and Ciij = Ciji = Cjii. . but not so often for C. with B′ given by Eq. integration yields dC dCijk  = yi yj yk  dT dT i j k (4-63) GR Pr  = (B0 + ωB1)  RT Tr An alternative form of the virial equation expresses Z as an expansion in powers of pressure about the real gas state at zero pressure (zero density): (4-64) Z = 1 + B′P + C′P2 + D′P3 + . and k identify species. (4-67). dB1dTr ∂Z dB0dTr B0 B1  = Pr  − 2 + ωPr  − 2 Tr ∂Tr P Tr Tr Tr  r   dB dBij  = yi yj  i j dT dT (4-62) Upon substitution of these equations into Eqs. This is often expressed through correlations for the compressibility factor Z. with B′ and C′ depending on temperature and composition only. . Analytical expressions for Z as functions of T and P or T and V are known as equations of state. 3. the mixing rules of Eqs. New York (1995)]. The first is a virial coefficient for a pure species. Cjjj. Particularly useful is the one for the second virial coefficient. . C′. The coefficient Bij characterizes a bimolecular interaction between molecules i and j. Again. the virial expansion in density truncated to three terms is usually suitable: (4-68) GR BP  =  RT RT j k where yi. wherein they are different (i ≠ j ). are the pressure series virial coefficients. . is usually preferred: Z = 1 + B′P = 1 + BPRT (4-67) For supercritical temperatures. (4-47): dB1 SR dB0  = − Pr  + ω  R dTr dTr  (4-76) . (4-46). it is satisfactory to ever higher pressures as the temperature increases. but below the critical pressure. . Two kinds of second virial coefficient arise: Bii and Bjj. called a cross coefficient. Direct substitution for Z in Eq. wherein the subscripts are the same (i = j). . and Ckkk are for the pure species. (4-47). respectively. D. McGraw-Hill. depend on temperature and composition only.. HR P B dB  =   −  RT R T dT  (4-59) The density series virial coefficients B. Consider first Eq. App. defined by Eq. The basic equation is BPc (4-72)  = B0 + ωB1 RTc with the acentric factor defined by Eq. . Thermodynamics. . SR P dB  =−   R R dT (4-71) An extensive set of three-parameter corresponding-states correlations has been developed by Pitzer and coworkers [Pitzer. . and therefore Bij = Bji. (4-60) and (4-61) follow rigorously from the methods of statistical mechanics. and Li [Fluid Phase Equilibria 226: 109–120 (2004)]. The composition dependencies of B and C are given by the exact mixing rules i B′ = BRT C′ = (C − B2)(RT)2 Equations for residual enthalpy and entropy may be developed from each of these expressions. For pressures up to several bars.PROPERTY CALCULATIONS FOR GASES AND VAPORS Equations (4-50) and (4-51) may sometimes be advantageously expressed in alternative form through use of mean heat capacities: H = Hig0 + 〈CigP 〉H(T − T0) + HR (4-55) T P S = Sig0 + 〈CigP 〉 S ln  − R ln  + SR T0 P0 (4-56) where 〈CigP 〉H and 〈CigP 〉S are mean heat capacities specific. D′.

set equal to zero.722  =  dTr T r5. and their derivatives are provided by Eq. New York (1975)]: 0. Eq. they are least accurate for polar and associating molecules. (4-88) and (4-89) for Z by Eq. Van Ness. For a mixture. 87. When i ≠ j. because the virial equations are suitable only for vapors and gases. (4-81) and (4-82) are substituted into Eqs. Although developed for pure materials. all equations reduce to the appropriate values for a pure species. but generalized correlations are available. 216–217. Adv.083 −  Tr1. with i ≠ j. it is a small (usually) positive number evaluated from minimal PVT data or. 132 and 162. 20: 263–272 (1974). 3d ed. ibid. N. Otherwise. ∂Z = ρ + ρ  ∂T dT dT dB dC 2 ρ Substituting in Eqs. As with all generalized correlations. (4-68) and in Eq. has physical significance.675  =  dTr T r2. (4-84). and Abbott [Introduction to Chemical Engineering Thermodynamics. (4-45) and (4-46). (4-72). T and P. Chem. Prentice-Hall. New York (2005)]: GR  = Z − 1 − ln Z + RT  (Z − 1) dρρ ρ (4-88) 0   dρρ ρ ∂Z HR  = Z − 1 − T  0 ∂T RT (4-89) ρ By differentiation of Eq.. Introduction to Chemical Engineering Thermodynamics. (4-60) and (4-62). 143–162 (1979)] and Hayden and O’Connell [Ind. The resulting equations are given by Smith. (4-68) may be written in two equivalent forms: BP CP2 Z 3 − Z 2 −  Z − 2 = 0 RT (RT)  (4-92)   B 1 P ρ3 +  ρ2 +  ρ −  = 0 (4-93) C C CRT In the event that three real roots obtain for these equations. Differentiation produces RTcij dB0 dBij dB1  =   + ωij  dT Pcij dT dT dBij R dB0 dB1  =   + ωij  dT Pcij dTrij dTrij   (4-82) where Trij = T/Tcij. and dB1/dTr are the same functions of Tr as given by Eqs. Eng. ibid. Proc. (4-72) written in extended form: RTcij 0 Bij =  (B + ωij B1) Pcij (4-81) where B0. Tcij. 2d ed. p. upon integration and reduction. kij = 0. Dev. are usually the favored independent variables.. it is incompatible with Eqs. McGraw-Hill. kij is an empirical interaction parameter specific to an i − j molecular pair. these equations define a set of interaction parameters without physical significance. Values of HR and SR are then given by Eqs. (4-70) and (4-71).. and Pcij are given by Prausnitz. BPc (4-95) Bˆ  = B0 + ωB1 RTc The reduced third virial coefficient Cˆ is defined as CP2c Cˆ  R2Tc2 A Pitzer-type correlation for Cˆ is then written as Cˆ = C0 + ωC1 (4-96) (4-97) . Data for third virial coefficients are often lacking. (4-47). (4-89) for the derivative yields. 21: 827–829 (1975). pp. Applications of Eqs. (1986)]: with and ωi + ωj ωij =  2 (4-83) Tcij = (TciTcj)12(1 − kij) (4-84) ZcijRTcij Pcij =  Vcij (4-85) Zci + Zcj Zcij =  2 (4-86) 13 V13 ci + Vcj Vcij =  2  3 (4-87) In Eq. Eng.4-10 THERMODYNAMICS In these equations. When i = j. appropriate for the vapor phase. Because Eq. Equation (4-68) may be rewritten in reduced form as  Pr Pr Z = 1 + Bˆ + Cˆ  Tr Z Tr Z 2 (4-94) where Bˆ is the reduced second virial coefficient given by Eq. pp. GR 3  = 2Bρ +  Cρ2 − ln Z RT 2 (4-90) HR dB T dC  = B − T  ρ + C −   ρ2 RT dT 2 dT   (4-91) The residual entropy is given by Eq.6 (4-79) dB1 0. 14: 209–216 (1975)]. and de Azevedo [Molecular Thermodynamics of Fluid-Phase Equilibria. With Z = P/ρRT. Basic to this extension are the mixing rules for the second virial coefficient and its temperature derivative as given by Eqs. in Chemistry Series 182. Thus by definition.2 (4-80) Although limited to pressures where the two-term virial equation in pressure has approximate validity.J.. 44: 630–633 (2005)]. Lichtenthaler. (4-68) is explicit in P. The following combining rules for ωij. Chem. pp.422 B0 = 0. Englewood Cliffs. dB0 /dTr.6 (4-77) 0. B0 and B1 and their derivatives are well represented by Abbott’s correlations [Smith and Van Ness. Res. B1. 7th ed. (4-60) and (4-62) to provide values of the mixture second virial coefficient B and its temperature derivative. For aqueous systems see Bishop and O’Connell [Ind. absence data. (4-68). When i = j and for chemically similar species. In a process calculation. Des. McGraw-Hill. 24: 1112–1115 (1978). A primary virtue of Abbott’s correlations for second virial coefficients is simplicity.172 B1 = 0. these correlations are applicable for most chemical processing conditions. More complex correlations of somewhat wider applicability include those by Tsonopoulos [AIChE J. and they must be transformed to make V (or molar density ρ) the variable of integration. rather than T and ρ (or T and V).139 −  Tr4. (4-90) and (4-91) therefore require prior solution of Eq. (4-77) through (4-80).. (4-68) for Z or ρ. only the largest Z (smallest ρ)..2 (4-78) dB0 0. Values for the cross coefficients Bij. values of Bij and dBij /dT from Eqs. these correlations can be extended to gas or vapor mixtures.

42748 0. Eng. Fundam. 21: 510–527 (1975)]. 29: 107–113 (1983)]. Iteration follows the same pattern as for Eqs. (4-104) in conjunction with Eqs. AIChE J. 93–94. (4-103a) is the ideal gas value RT/P. and the process continues until the change in V is suitably small. (4-104b) it is Z = β. specific to a particular form of the equation of state. all equal to Vc. and many others have since been proposed. Van Ness. except at the critical pressure. 7th ed.05539 0. Physically meaningful values of V are always real. all functions of composition.7 Tr10. Another class of equations. For T < Tc. Orbey and Vera [AIChE J.08664 0.5 (4-99) The first is given by. p. but for a range of lower pressures there are three. An extensive review is given by Valderrama [Ind.. Suitable estimates of the parameters in cubic equations of state are usually found from values for the critical constants Tc and Pc. independent of substance and determined for a particular equation of state from the values assigned to % and σ.45724 *Smith.54226ω − 0.07780 0. ω)† αPR(Tr. pp. Eng. Res. solution for V at any positive value of P yields only one real positive root. in which the acentric factor ω enters through function α(Tr. with Tr and Pr specified.574ω − 0.00242 C1 = − 0. ω) as an additional parameter. Equations of state. Chem. Of the equations published more recently. and Abbott [Introduction to Chemical Engineering Thermodynamics. and Rubin [J. Equation (4-100) may be rearranged to facilitate its solution either for a vapor or vaporlike volume or for a liquid or liquidlike volume. Chem. and Abbott. written as RT a(T) P =   −  V − b (V + %b)(V + σb) (4-100) For a specific form of this equation. of state α(Tr) σ % Ω Ψ RK (1949) SRK (1972) PR (1976) Tr−1/2 αSRK(Tr. Vapor: Liquid: a(T) RT V−b V =  + b −   P (V + %b)(V + σb) P RT − bP − VP V = b + (V + %b)(V + σb)  a(T)  (4-103a)  (4-103b) 4-11 Solution for V is most convenient with the solve routine of a software package. and the second is inspired by. whereas parameters a(T) and b are substancedependent. New York (2005)]. and Ψ are given in Table 4-2. and for Eq. All are encompased by a generic cubic equation of state. a modification of the RK equation [Chem.08664 0. yield three-parameter corresponding-states correlations. was introduced by Benedict. and Abbott [Introduction to Chemical Engineering Thermodynamics. Introduction to Chemical Engineering Thermodynamics. The SRK equation and the PR equation. % and σ are pure numbers. Eqs. ω) = [1 + (0. 44: 233–244 (1949)]. find application in the petroleum and natural gas industries for light hydrocarbons and a few other commonly encountered gases [see Lee and Kesler. and when combined with Eqs. The numerical assignments for parameters %. in Eq. McGraw-Hill. In either case. 8: 334–345 (1940). the two most popular are the Soave-Redlich-Kwong (SRK) equation. solution for Z is by iteration. As shown by Smith. Eq. The final value of Z yields the volume root through V = ZRT/P. (4-103) are obtained by substituting V = ZRT/P. Equation (4-94) is cubic in Z. the smallest root is a liquid or liquidlike volume.26992ω 2) (1 − Tr1/2)]2 . the same for all substances. ω) = [1 + (0. 7th ed. This equation contains eight parameters. The procedure is discussed by Smith. When T = Tc. for Eq. McGrawHill. In these equations ψ and Ω are pure numbers. σ.5 (4-98) 0. (4-103). Cubic Equations of State The modern development of cubic equations of state started in 1949 with publication of the RedlichKwong (RK) equation [Chem. iteration is initiated by substituting the estimate on the right side. Eng.42748 0. (4-104a) the initial estimate is Z = 1.PROPERTY CALCULATIONS FOR GASES AND VAPORS Correlations for C0 and C1 with reduced temperature are 0. 218–219. 10: 747–758 (1942)]. When T > Tc. Chem. (4-103b) it is V = b. known as extended virial equations. Vapor: Liquid: Z−β Z = 1 + β − qβ  (Z + %β)(Z + σβ) 1+β−Z Z = β + (Z + %b)(Z + σb)  qβ  (4-104a) (4-104b) where by definition bP β  RT (4-105) and a(T) q  bRT (4-106) These dimensionless quantities provide simplification.. and greater than parameter b. 98. Ω. and (4-47) lead to GR  = Z − 1 − ln(Z − β) − qI RT (4-109) HR d ln α(Tr)  = Z − 1 +  − 1 qI RT d ln Tr (4-110)  TABLE 4-2 of State*  Parameter Assignments for Cubic Equations For use with Eqs. Rev. Expressions are also given for α(Tr. New York (2005)].. New York (2005). where three roots exist. 42: 1603–1618 (2003)]. this is also true. and the largest root is a vapor or vaporlike volume. Sci.176ω 2) (1 − Tr1/2)]2 ‡ α PR(Tr. An initial estimate for V in Eq. The resulting value of V on the left is returned to the right side.02432 0.480 + 1. (4-89). positive. ω)‡ 1 1 1 + 2 0 0 1 − 2 0.01407 +  −  Tr Tr10.02676 +  −  Tr2. ω) for the SRK and PR equations.00313 C0 = 0. (4-100) has three volume roots. 15: 59–64 (1976)]. Equations for Z equivalent to Eqs. 7th ed. Webb. (4-100) the appropriate equations are given as α(Tr)R2Tc2 a(T) = ψ  Pc (4-101) RT b = Ω c (4-102) Pc Function α(Tr) is an empirical expression. (4-101) and (4-102). 27: 1197–1203 (1972)] and the PengRobinson (PR) equation [Ind. Van Ness. An initial guess of Z = 1 on the right side usually leads to rapid convergence. pp. the middle root is of no significance. despite their complexity. (4-88). yield two-parameter corresponding-states correlations.37464 + 1. of which two may be complex. As an equation cubic in V.. † α SRK(Tr . such as the Redlich-Kwong (RK) equation. Here. they yield Pr β=Ω  Tr (4-107) Ψα(Tr) q=  ΩTr (4-108) In Eq. only one real positive (liquidlike) root exists at high pressures. (4-104) through (4-106) Eq. McGraw-Hill. Phys.. Van Ness. It and its modifications. which expresses Z as a function of Tr and Pr only.

One such set is given by Eqs. 3. Substitution for Z by Eq. Parameter bi is for pure species i. Pitzer’s Generalized Correlations In addition to the corresponding-states coorelation for the second virial coefficient.43577T 6r Tr (4-122) The value of ω to be used with Eq.6 T c =  K 1 + 21.016T (for H2) (4-125) 51. and ω have no physical significance for the mixture. Generalized correlations are developed here for the residual enthalpy and residual entropy from Eqs. Cubic equations of state may be applied to mixtures through expressions that give the parameters as functions of composition.01 ≤ Pr ≤ 10.J. ω for a particular substance is determined from ln Prsat − ln Pr0(Tr ) ω= (4-123)  ln Plr(Tr ) where Tr is the reduced normal boiling point and Prsat is the reduced vapor pressure corresponding to 1 standard atmosphere (1. analytical refinements to the tables have been developed. 21: 510–527 (1975). Pc. Combination of Eqs.01325 bar). The Lee-Kesler correlations provide reliable data for nonpolar and slightly polar gases. McGrawHill. Acentric factor ω is defined by Eq. New York (1995)] developed a full set of generalized correlations. Prausnitz. so called because the defined values of Tpc. are represented by (HR)1/RTc and (SR)1/R.09648 ln P0r (Tr) = 5.5 Pc  =  bar 1 + 44. 172–173.0 and 0. (4-120) is found from the correlation by requiring that it reproduce the normal boiling point. and Abbott. The interaction parameter aij is often evaluated from pure-species parameters by a geometric mean combining rule aij = (aiaj)1/2 (4-115) These traditional equations yield mixture parameters solely from parameters for the pure constituent species. (4-49) gives r  HR  = − T 2r RTc Pr 0 ∂Z0  ∂Tr Pr dPr  − ωT 2r Pr ∂Z   ∂T Pr 1 0 Pr r dPr  Pr (4-117) GR SR 1 HR  =   −  R RT Tr RTc  By Eq. excluding ω but including the minus signs. and empirical mixing rules are often used to relate mixture parameters to purespecies parameters.4721 ln Tr + 0. App. 2005)]. hydrogen. For hydrogen. N. (HR)1/RTc.169347T 6r Tr (4-121) 15. The quantum gases (e. (4-116) and (4-117) leads to SR  = − R ∂Z  T   ∂T Pr 0 r 0 r Pr  dP + Z 0 − 1 r − ω Pr ∂Z  T   ∂T Pr 1 r 0 r Pr  dP + Z1 r Pr If the first terms on the right sides of Eq.5 Vc  =  cm3mol−1 1 − 9. Since then. (1999)] propose the use of temperaturedependent effective critical parameters. They are most likely to be satisfactory for mixtures comprised of simple and chemically similar molecules. (4-47). Ppc. Upper Saddle River.91/2. Lee and Kesler also included a Pitzer-type correlation for vapor pressures: 0 1 ln P sat (4-120) r (Tr) = ln P r (Tr) + ω ln P r (Tr) i j with aij = aji.016T (for H2) (4-124) 20. errors of less than 3 percent are likely. with extended range and accuracy. Although the tables representing the Pitzer correlations are based on data for pure materials. (2-17).3 ≤ Tr ≤ 4. Use of these effective critical parameters for hydrogen requires the further specification that ω = 0. E.6875 and ln P1r (Tr) = 15. 2. that is. helium. (4-48) and (4-49). (HR)0/RTc. 5th. and de Azevedo [Molecular Thermodynamics of Fluid-Phase Equilibria. (2-80) through (2-82) in the Seventh Edition of Perry’s Chemical Engineers’ Handbook (1997). These tables cover both the liquid and gas phases and span the ranges 0. (4-48) into generalized form:  GR  = RT Pr dP (Z0 − 1) r + ω Pr 0  Pr dP Z1 r Pr 0 (4-116) Differentiation of Eq. they may also be used for the calculation of mixture properties. These equations define pseudoparameters. Correlations for Z appear in Sec. No established theory prescribes the form of this dependence. pp. (SR)0/R. The aij are of two types: pure-species parameters (like subscripts) and interaction parameters (unlike subscripts). Prentice-Hall PTR.g. 3d ed. the recommended equations are n n n n n 43. (4-104a) or (4-104b).2518 −  − 13. Larger errors can be expected in applications to highly polar and associating gases.016T (for H2) (4-126) where T is absolute temperature in kelvins. . see also Smith. New York (1996.8/2. The simplest realistic expressions are a linear mixing rule for parameter b and a quadratic mixing rule for parameter a b = xi bi (4-113) a = xi xj aij (4-114) i Pitzer’s original correlations for Z and the derived quantities were determined graphically and presented in tabular form. (2-63) yields ∂Z ∂Z0 =   ∂T ∂T r Pr r Pr ∂Z1 + ω  ∂Tr  Pr Substitution for (∂Z∂Tr)P in Eq. App. (2-63): Z = Z0 + ωZ1 (2-63) where Z0 and Z1 are each functions of reduced temperature Tr and reduced pressure Pr. They list values of Z0. (2-63) puts Eq. 6th. Pitzer and coworkers [Thermodynamics..92714 −  − 1. then where 6. 2001. and (SR)1/R. and 7th eds.28862 ln Tr + 0. Van Ness. A set of recipes is required relating the parameters Tc. They have as their basis an equation for the compressibility factor. Lichtenthaler.. the quantum gas most commonly found in chemical processing. (4-117) and of this equation (including the minus signs) are represented by (HR)0/RTc and (SR)0/R and if the second terms. McGraw-Hill. Introduction to Chemical Engineering Thermodynamics.0..2/2. and neon) do not conform to the same corresponding-states behavior as do normal fluids. Z1. 3d ed.4-12 THERMODYNAMICS SR d ln α(Tr)  = ln (Z − β) +  qI R d ln Tr Z + σβ 1 I =  ln  σ−% Z + %β  where (4-111) HR (HR)0 (HR)1  =  + ω RTc RTc RTc (4-118) (4-112) SR (SR)0 (SR)1  =  + ω R R R (4-119) Preliminary to application of these equations Z is found by solution of either Eq. and ω for a mixture to the pure-species values and to composition. as given by Eq.. The most popular Pitzer-type correlation is that of Lee and Kesler [AIChE J.

Latent heats of vaporization are functions of temperature. P). A much more accurate equation is that of Riedel [Chem. Calculation of property changes for the liquid state are usually based on alternative forms of Eqs. C. McGraw-Hill. 7th ed. For example. R.013) ∆H lvn  =  0. Van Ness. Moreover (outside the critical region). New York (2001)]. More satisfactory for general use is the Antoine equation dH = CP dT + (1 − βT)V dP (4-132) dT dS = CP  − βV dP T B ln P sat = A −  T+C (4-133) β dU = CV dT +  T − P dV κ (4-134) The Wagner equation is useful for accurate representation of vapor pressure data over a wide temperature range. and experimental values at a particular temperture are often not available. (4-127) and (4-130) yields (4-139) This equation follows from Eq. they are weak functions of T and P and are often assumed constant at average values. Values of the constants for either the Wagner equation or the Antoine equation are given for many species by Poling. and are often assumed constant at the values for the saturated liquid at temperature T. p. Useful here are the definitons of two liquid-phase properties—the volume expansivity β and the isothermal compressibility κ: 1 ∂V β   V ∂T  (4-127)  ∂V dV =  ∂T ∆Mlv Mv − Ml l (4-128) T ∂V dP  dT +   ∂P P T ∆Hlv = T ∆Slv If V is constant. 221. Trouton’s rule of 1884 provides a simple check on whether values calculated by other methods are reasonable: ∆H lvn  ∼ 10 RTn Here. (4-127) and (4-128) becomes dV  = β dT − κ dP V (4-138) v where M and M are molar properties for states of saturated liquid and saturated vapor. because vaporization at the vapor pressure Psat occurs at constant T. B. and Tn must be chosen so that ∆Hnlv/RTn is dimensionless. The simplest such equation is B ln P sat = A −  T (4-141) where A and B are constants for a given chemical species. Ing. (4-32) through (4-35). A. and D are constants. (4-32) through (4-35) by Eqs. It expresses the reduced vapor pressure as a function of reduced temperature β CV dS =  dT +  dV T κ (4-135)  Integration of these equations is most common from the saturatedliquid state to the state of compressed liquid at constant T. Tech. An alternative treatment of V comes from Eq. (4-131). (4-15).OTHER PROPERTY FORMULATIONS 4-13 OTHER PROPERTY FORMULATIONS LIQUID PHASE Although residual properties have formal reality for liquids as well as for gases. This equation in combination with Eqs. This equation approximates Psat over its entire temperature range from triple point to critical point. New York (2005)] the heat of vaporization is directly related to the slope of the vapor-pressure curve. Some experimental values of the enthalpy change of vaporization ∆Hlv. McGraw-Hill. are listed in Table 2-150. App. shown in Table 4-1.5 + Cτ 3 + Dτ 6 ln P sat r =  1−τ where (4-142) (4-143) τ 1 − Tr and A. Recourse is then made to approximate methods. The enthalpy change and entropy change of vaporization are directly related: P 1 ∂V κ −   V ∂P For V = f (T. (4-129) then gives V ln 2 = β(T2 − T1) − κ(P2 − P1) (4-131) V1 Substitution for the partial derivatives in Eqs.. and ∆Hnlv is the latent heat at this temperature. As shown by Smith. (4-132) and (4-133) in integral form become H = Hsat +  P P (1 − βT)V dP (4-136)  (4-137) sat S = Ssat − P P sat β V dP Again. their advantageous use as small corrections to ideal gas state properties is lost. β and V are weak functions of pressure for liquids. 5th ed.930 − Tr RTn n (4-144) . 26: 679–683 (1954)]: 1. β =  κ V ∂P  ∂T  dPsat ∆Hlv = T ∆V lv  dT (4-129) (4-130) Because liquid-phase isotherms of P versus V are very steep and closely spaced. Integration of Eq. Tn is the absolute temperature of the normal boiling point. this exact thermodynamic relation provides the connection between the properties of the liquid and vapor phases. both β and κ are small.092(ln Pc − 1. The units of ∆Hnlv. In application an empirical vapor pressure versus temperature relation is required. Eqs. Prausnitz. and Abbott [Introduction to Chemical Engineering Thermodynamics. treatment of this transition is facilitated by definition of property changes of vaporization ∆Mlv: (4-140) Known as the Clapeyron equation.. It is also a sound basis for interpolation between reasonably spaced values of T. and O’Connell [The Properties of Gases and Liquids. usually called the latent heat of vaporization. The Aτ + Bτ1. which for this application can be written V = V exp[−κ(P − P )] sat sat LIQUID/VAPOR PHASE TRANSITION The isothermal vaporization of a pure liquid results in a phase change from saturated liquid to saturated vapor at vapor pressure Psat.

MASS. potential. and there may be stirring work. and q measure quantity per unit of time. energy. AND ENTROPY BALANCES FOR OPEN SYSTEMS Mass and energy balances for an open system are written with respect to a region of space known as a control volume. Eq. n⋅. The entropy change of the surroundings results from heat transfer between system and surroundings. and kinetic forms. As discussed with respect to the mass balance. (4-148) is zero. Let Q⋅ j represent the heat-transfer rate at a particular location on the control surface associated with a surroundings temperature Tσ. with zero as a limiting value. it also means that no changes occur with time in the properties of the fluid within the control volume or at its entrances and exits. and entropy balances for open systems. dmcv  + ∆(ρuA)fs = 0 dt (4-148) This form of the mass balance equation is often called the continuity equation. (4-149) is zero are said to occur at steady state. it may be rigid or flexible. becomes 1 ⋅ ⋅ (4-150) ∆ H +  u2 + zg m⋅ = Q + Ws fs 2 Entropy Balance for Open Systems An entropy balance differs from an energy balance in a very important way—entropy is not conserved. and Abbott [Introduction to Chemical Engineering Thermodynamics.. and the first term of Eq. Energy Balances for Steady-State Flow Processes Flow processes for which the first term of Eq. Chem.38 ∆H lv2 =  (4-145)  1 − Tr ∆H lv1 This equation is simple and fairly accurate. (4-151) is therefore the sum of all such quantities. Eng. and g is the local acceleration of gravity. this means that the mass of the system within the control volume is constant. According to the second law. and the time rate of change of the entropy of the surroundings. Although velocity u does not represent quantity of flow. n Watson’s equation [Ind. General Energy Balance Because energy. Flow rates m⋅. defined with respect to the sys− Qj Tσ. The flow is positive when directed into the control volume and negative when directed out. The velocity u in the kinetic energy term is the bulk mean velocity as defined by the equation u = m˙ ρA. The mass balance is expressed mathematically by dmcv  + ∆(m˙)fs = 0 dt (4-147) The operator ∆ signifies the difference between exit and entrance flows.. Mass flow rate is related to velocity by m˙ = uAρ (4-146) where A is the cross-sectional area of a conduit and ρ is mass density. McGraw-Hill. the rate of ⋅entropy change in the surroundings as a result of this transfer is ⋅ . 7th ed. j For any process. This equation is the general rate form of the entropy balance. Smith. (4-146). Energy may also flow across the control surface as heat and work. to a heat rate with respect to the surroundings. New York (2005)] show that the general energy balance for flow processes is   d(mU)cv 1 ˙ (4-149)  + ∆ H +  u2 + zg m˙ = Q˙ + W fs dt 2 ⋅ The work rate ⋅ W may be of several forms. Van Ness. the entropy changes resulting from the process sum not to zero. the time rate of change of the entropy of the fluid contained within the control volume. Streams flowing into and out of the control volume have associated with them energy in its internal. In accord with Eq. (4-3). Estimates of the latent heat of vaporization of a pure liquid at any temperature from the known value at a single temperature may be based on an experimental value or on a value estimated by Eq. In accord with the second law. When the mass flow rate m⋅ is given by Eq. No expansion of the control volume is possible under these circumstances. the time rate of change of energy within the control volume equals the net rate of energy transfer into the control volume. it must be positive. The minus sign converts Q j tem. (4-151) can be written Q⋅ j d(mS)cv ⋅ ∆(Sm⋅ )fs +  (4-152) −  = SG ≥ 0 dt j Tσ. j. and all contribute to the energy change of the system. is conserved. Mass Balance for Open Systems Because mass is conserved. the entropy changes in the system and surroundings as the result of any process must be positive. applicable at any instant. and Eq. This equation provides reasonable approximations. but to a positive quantity called the entropy generation term. 47–48. errors rarely exceed 5 percent. it is an important design parameter. with a limiting value of zero for a reversible process. and the subscript fs indicates that the term encompasses all flowing streams.  2 1 THERMODYNAMICS OF FLOW PROCESSES The thermodynamics of flow encompasses mass. (4-149). (4-144). Thus. i. The third term in Eq. like mass. For the special case of steady-state flow. This surface may follow fixed walls or be arbitrarily placed.e. for systems whose boundaries allow the inflow and outflow of fluids.4-14 THERMODYNAMICS where Pc is the critical pressure in bars and Tr is the reduced temperature at Tn. expressed as rates. bounded by an imaginary control surface that separates it from the surroundings. ENERGY. and the general energy balance. pp. z is elevation above a datum level. The only work of the process is shaft work. the control volume contains a constant mass of fluid. The common measures of flow are as follows: Mass flow rate m⋅ molar flow rate n⋅ volumetric flow rate q velocity u Also m˙ = Mn˙ and q = uA where M is molar mass. The statement of balance. If ρ is molar density. Most commonly there is shaft work Ws. the time rate of change of mass within the control volume equals the net rate of flow of mass into the control volume. then this equation yields molar flow rate. the two kinds of irreversibility are (1) those internal to the control volume and (2) those resulting from heat transfer across finite temperature differences that may exist between the . The three terms on the left are the net rate of gain in entropy of the flowing streams. is therefore        Net rate of change in entropy of flowing streams Time rate of change of entropy in control volume + + Time rate of change of entropy in surroundings  Total rate = of entropy generation The equivalent equation of entropy balance is d(mS)cv dStsurr ⋅ +  = SG ≥ 0 (4-151) ∆(Sm⋅ )fs +  dt dt ⋅ where SG is the entropy generation term.j. Work may be associated with expansion or contraction of the control volume. 35: 398–406 (1943)] has found wide acceptance: 1 − Tr 0.

and entropy balance appear in the preceding sections. (4-155). The derivatives in these equations are rates of change with length as the fluid passes through a duct. its cross-sectional area changes with length in such a way as to make the flow nearly frictionless. and Eqs. The most common restricted case is for steady flow processes. j (4-153) m˙ 1 = m˙ 2 = m˙ (4-154) 1 ˙s ∆ H +  u2 + zg m⋅ = Q˙ + W 2 fs (4-150) ∆u2 ∆H +  + g ∆z = Q + Ws 2 (4-155) Q˙ j ∆(Sm˙ )fs −  = S˙ G ≥ 0 j Tσ . ∆u2 ∆H +  = 0 2 In differential form. and Eq. (4-160) and (4-161) reduce to (4-160)  Mach number M is the ratio of the speed of fluid in the duct to the speed of sound in the fluid. and specific volume within a flowing stream. This is not possible in a pipe of constant cross-sectional area. dS/dx ≥ 0. (4-160) and (4-161) become dP u 2 dA 1  =  2  dx VA 1 − M dx   du dA u 1  = −  2  dx A 1 − M dx The characteristics of flow depend on whether the flow is subsonic (M < 1) or supersonic (M > 1). Here. Equation (4-160) relates the pressure derivative. 255–258. m⋅ is the same for both streams. In the limit as the flow approaches reversibility. and a supersonic stream entering a pipe of constant cross section undergoes a compression shock. and the maximum fluid velocity obtainable is the speed of sound. however. and the equations may be divided through by this rate to put them on the basis of a unit amount of fluid flowing through the control volume. the process is completely reversible. wherein the mass and thermodynamic properties of the fluid within the control volume are not time-dependent. the mass rate of flow decreases so that the sonic velocity is still obtained at the outlet of the lengthened pipe. the result of which is an abrupt and finite increase in pressure and decrease in velocity to a subsonic value. Considered here is the adiabatic. The limit is reversible flow. implying that • The process is internally reversible within the control volume. one-dimensional flow of a compressible fluid in the absence of shaft work and changes in potential energy. pressure increases and velocity decreases in the direction of flow. because these inequalities would reverse if the velocity were to become supersonic. (4-161). duct cross-sectional area. In a properly designed nozzle. Pipe Flow For a pipe of constant cross-sectional area. Abbott. this increase approaches zero. • Heat transfer between the control volume and its surroundings is reversible. the velocity increases and the pressure decreases as the cross-sectional area TABLE 4-4 Nozzle Characteristics Subsonic: M < 1 (4-161) βu2/CP + M2 ds du u = T    dx dx 1− M2 dp T 1 + βu2/CP ds  = −    1 − M2 dx V dx (4-159) Smith. entropy. the velocity derivative. reached at the exit of the pipe. (4-148) here becomes d(ρuA) = d(uA/V) = 0. Thus. j (4-157)   ⋅ system and surroundings. In this event dS/dx = 0. the irreversibilities of fluid friction in adiabatic flow cause an entropy increase in the fluid in the direction of flow. and Eqs. and ∆z all set equal to zero. and dP/dx < 0 and du/dx > 0. (4-15) and other property relations yield two very general equations  Balance equations for single-stream steady-flow processes Balance equations for steady-flow processes dA  dx dP  dx du  dx Supersonic: M > 1 Converging Diverging Converging Diverging − + − + − + + − + − − + . The velocity increase is. In general. Pressure therefore decreases and velocity increases in the direction of flow. McGraw-Hill. the flow becomes sonic and lengthening the pipe does not alter this result. for subsonic flow in a converging nozzle. enthalpy. For a discharge pressure low enough. However. to the entropy and area derivatives. 7th ed. velocity. limited. Nozzles Nozzle flow is quite different from pipe flow. then. energy. M2 < 1.THERMODYNAMICS OF FLOW PROCESSES TABLE 4-3 4-15 Equations of Balance General equations of balance dmcv  + ∆(m˙)fs = 0 dt   d(mU)cv 1  + ∆ H +  u2 + zg m˙ dt 2 ˙ = Q˙ + W (4-147) ∆(m˙)fs = 0 (4-149) (4-152) fs d(mS)cv Q˙ j  + ∆(Sm˙ )fs −  = S˙ G ≥ 0 dt j Tσ . APPLICATIONS TO FLOW PROCESSES Duct Flow of Compressible Fluids Thermodynamics provides equations interrelating pressure changes. In this event. dS/dx reaches its limiting value of zero. In each case important applications require less general versions. whence dV du dA  −  −  =0 V u A dP βu2 dS u2 dA V(1 − M2)  + T 1 +   −   = 0 CP dx A dx dx βu2/ CP + M2 dS du 1 u2 dA u − T    + 2   = 0 dx 1 − M2 dx 1−M A dx   the second law. In the limiting case where SG = 0. steady-state. The appropriate energy balance is Eq. dA/dx = 0. and Van Ness [Introduction to Chemical Engineering Thermodynamics.. this flow regime is unstable. According to the equations for supersonic pipe flow. dH = − u du (4-158) The continuity equation given by Eq. for which the rate of entropy increase is zero. With Q. The possibilities are summarized in Table 4-4. New York (2005)] show that these basic equations in combination with Eq. Ws. all terms on the right in these equations are then positive. According to  When flow is subsonic. pp. Summarized in Table 4-3 are the basic equations of balance and their important restricted forms. A further simplification results when there is but one entrance and one exit to the control volume. Summary of Equations of Balance for Open Systems Only the most general equations of mass. j (4-156) Qj ∆S −  = SG ≥ 0 j Tσ.

The process is adiabatic. Compression Processes Compressors. 7th ed. A compressor efficiency is defined as Ws(isentropic) η  Ws In view of Eqs. Eqs. (4-15) and (4-159) for isentropic flow gives 1 H H P1 (H)S 2 2 P2 S u du = − V dP Integration. 269.7 to 0. although the effect is usually small. the large temperature drop results from evaporation of liquid. Turbines (Expanders) High-velocity streams from nozzles impinging on blades attached to a rotating shaft form a turbine (or expander) through which vapor or gas flows in a steady-state expansion process which converts internal energy of a high-pressure stream into shaft work. Sonic velocity is reached at the throat. isentropically. if the fluid expands reversibly and adiabatically. and irreversibilities cause the path to be directed toward increasing entropy. This second equation establishes the final state of the fluid and allows calculation of H2. some of the liquid vaporizes or flashes. This results in a considerable temperature drop because of the evaporation of liquid. indeed. (4-165). pumps. based on the same assumptions. (4-163) both H2 and Ws are unknown. The energy equations are the same. (4-155) reduces to ∆H = 0 or H2 = H1. without appreciable change in kinetic or potential energy undergoes a finite pressure drop. and the lower the value of η. blowers. because its enthalpy depends on temperature only. A converging subsonic nozzle can therefore deliver a constant flow rate into a region of variable pressure. (4-159) through (4-161) apply to adiabatic compression. where dA/dx = 0. New York (2005). and vacuum pumps are all devices designed to bring about pressure increases. The irreversible path (solid line) starts at point 1 and terminates at point 2 on the isobar for P2. [Smith. The isentropic work of compression. Again. The maximum possible fluid velocity is the speed of sound. In any properly designed turbine. turbine efficiency is defined as Ws η  Ws(isentropic) where Ws is the actual shaft work. The vertical dashed line rising from point 1 to point 2′ represents the reversible adiabatic (isentropic) compression process from P1 to P2. Van Ness. the value of H1 is fixed. γ CP/CV.7 to 0. An analytical expression relating velocity to pressure in an isentropic nozzle is readily derived for an ideal gas with constant heat capacities. (4-163) and (4-165). producing a mixture of saturated liquid and saturated vapor at the lower pressure. an equation valid for ideal gases with constant heat capacities. fans. a reduction in pressure at constant enthalpy results in a decrease in temperature. The process therefore occurs at constant enthalpy. heat transfer and changes in potential and kinetic eneregy are negligible. 4-3. Throttling Process Fluid flowing through a restriction. 4-2 (4-162) where the final term is obtained upon elimination of V by PV γ = const. ∆H η=  (∆H)S (4-166) Values of η usually range from 0. This throttling process produces no shaft work. the farther point 2 lies to the right on the P2 isobar. Throttling of a wet vapor to a sufficiently low pressure causes the liquid to evaporate and the vapor to become superheated. the diverging section acts as a diffuser—the pressure rises and the velocity decreases in the conventional behavior of subsonic flow in a diverging section. and the energy balance alone does not allow their calculation. Introduction to Chemical Engineering Thermodynamics. (4-163) and (4-165). they are the same as for turbines or expanders. with nozzle entrance and exit conditions denoted by 1 and 2. Eq. The HS diagram of Fig. However. If this is not the case. McGraw-Hill. i. Equation (4-155) therefore reduces to Ws = ∆H = H2 − H1 S (4-165) The absolute value |Ws |(isentropic) is the maximum work that can be produced by an adiabatic turbine with given inlet conditions and given discharge pressure. Here. The temperature of an ideal gas is not changed by a throttling process. then S2 = S1. such as an orifice. .4-16 THERMODYNAMICS diminishes. Supersonic velocities characterize the diverging section of a properly designed converging/diverging nozzle. For most real gases at moderate conditions of T and P. Equation (4-164) then gives the isentropic work: Ws(isentropic) = (∆H)S Adiabatic expansion process in a turbine or expander. Combination of Eqs. and a further increase in velocity and decrease in pressure require a diverging cross-sectional area to accommodate the increasing volume of flow. is the minimum shaft work required for compression of a gas from a given initial state to a given discharge pressure. Because the actual expansion process is irreversible.] FIG. in the turbine. In Eq.8. The motive force may be provided by steam (turbine) or by a high-pressure gas (expander). reached at the exit. The greater the irreversiblity. By Eqs. 4-2 compares the path of an actual expansion in a turbine with that of an isentropic expansion for the same intake conditions and the same discharge pressure. If a saturated liquid is throttled to a lower pressure. yields u22 − u21 = − 2 P 2γP V  V dP =  1 −   γ−1  P P2 1 1 P1 2 (γ −1)/γ The rate form of this equation is ˙ s = m˙ ∆H = m˙ (H2 − H1) W 1 (4-163) (4-164) When inlet conditions T1 and P1 and discharge pressure P2 are known.e. and Abbott. The isentropic path is the dashed vertical line from point 1 at intake pressure P1 to point 2′ at P2.8. p. Their energy requirements for steady-state operation are of interest here. as given by Eq. The compression process is shown on an HS diagram in Fig. this becomes (∆H)S η  ∆H (4-167) Compressor efficiencies are usually in the range of 0. Thus... Compression of gases may be accomplished in rotating equipment (high-volume flow) or for high pressures in cylinders with reciprocating pistons. The pressure at the throat must be low enough for the velocity to become sonic. and in the absence of heat transfer.

One might expect the solution property M to be related solely to the properties Mi of the pure chemical species which comprise the solution. usually at initial values. p. Interpolation in Table 2-282 at 20 bar yields an enthalpy of H = 1026. to a good approximation ∆H = CP ∆T + V(1 − βT) ∆P (4-169) T ∆S = CP ln 2 − βV∆P T1 (4-170) The estimated power which must be supplied from an external source is ˙ = 1.7 = 504.654 × 105 kJ s−1 = 3. V. A total-system property is then nM. application of Eq. TH and TC are the hot and cold temperatures of the heat reservoirs between which the heat engine operates. no such generally valid relation is known. stored as saturated liquid at 115 K and 1. It is proposed to vaporize the LNG with heat discarded from a heat engine operating between 300 K. (4-163) through (4-165) and Eq. Although the chemical species which make up a solution do not have their own individual properties. 274. The enthalpy change of isentropic compression is then ∆HS = 1026. using the work produced by the heat engine to supply part of the compression work.] FIG. a solution property may be arbitrarily . [Smith.. Thus. (4-132) and (4-133). the farther this point lies to the right on the P2 isobar.160)  − 1 = 3. U. New York (2005). or 300 and 115 K.436 kJ kg−1 K−1  Isentropic compression of saturated vapor at 1. dH = V dP (constant S) Ws(isentropic) = (∆H)S = TH W = QC  −1 TC In combination: Combining this with Eq. and the connection must be established experimentally for every specific system.436 kJ kg−1 K−1. PARTIAL MOLAR PROPERTIES For a homogeneous PVT system comprised of any number of chemical species. S.600 kW W SYSTEMS OF VARIABLE COMPOSITION The composition of a system may vary because the system is open or because of chemical reactions even in a closed system.3 kJ kg−1) = 1. and β are constant. Eq.5 − 297. Integration then gives For a compressor efficiency of 75 percent. The more irreversible the process. (4-167) are valid. A.8 kJ kg−1 Interpolation in Table 2-282 of the Seventh Edition of Perry’s indicates an actual temperature of 265.096 × 105 kW. the temperature of atmospheric air.342 × 105 kJ s−1 = 1. where n = Σ i ni and i is the index identifying chemical species.325 bar. usually by rotating equipment. For estimation purposes we need not be concerned with the design of the heat engine. terminating at point 2. and these are seldom available. and Abbott.7 kJ kg−1  V dP P2 P1 The usual assumption for liquids (at conditions well removed from the critical point) is that V is independent of P. The enthalpy and entropy of saturated vapor at 115 K and 1. The equations that apply to Carnot engines can be found in any thermodynamics text. 4-3 The actual irreversible compression process follows the solid line from point 1 upward and to the right in the direction of increasing entropy. H.5 = 223. and is unloaded at the rate of 450 kg s-1. McGraw-Hill. The equations developed here apply regardless of the cause of composition changes.. |QC| is the heat transferred at the cold temperature. respectively. Van Ness. 2 2 H H P2 (H)S W = QH − QC By the first law: 1 P1 S QH TH  =  QC TC By the second law:  S Adiabatic compression process. Because temperature changes in the pumped fluid are very small and because the properties of liquids are insensitive to pressure (again at conditions not close to the critical point). However.3 kJ kg−1 η 0.325 bar are given in Table 2-281 of the Seventh Edition of Perry’s as Hv = 802. or 1.342 × 105 kW (4-168) Also useful are Eqs. The same equations apply to adiabatic pumps as to adiabatic compressors.5 kJ kg−1 and Sv = 9. Liquids are moved by pumps. Thus. say. Eqs. and 115 K. these equations are usually integrated on the assumption that CP. which is quite suitable for its entry into the distribution system. Estimate the work to be supplied from an external source.2 − 802. The saturated-vapor LNG so produced is compressed adiabatically to 20 bar.325 to 20 bar produces superheated vapor with an entropy of 9. |W| is the work produced by the Carnot engine. However. (4-163) requires values of the enthalpy of compressed (subcooled) liquids.5 + 298. The fundamental property relation. to vaporize the LNG. Introduction to Chemical Engineering Thermodynamics. QC = (450)(504.e.SYSTEMS OF VARIABLE COMPOSITION Example 1: LNG Vaporization and Compression A port facility for unloading liquefied natural gas (LNG) is under consideration. The work of compression is W = m ∆H = (450 kg s−1)(298.160 kJ s−1 The equation for work gives 300 W = (227. let symbol M represent the molar (or unit-mass) value of an extensive thermodynamic property. provides an alternative.2 kJ kg−1 at 234. i.9 K for the compressed LNG. 7th ed. and the lower the efficiency η of the process.3 = 1100.342 × 105 − 1. The LNG arrives by ship.654 × 105 kW 115 This is the reversible work of a Carnot engine.8 kJ kg−1 For a flow rate of 450 kg s−1. or G.65 K. and enthalpy values from Table 2-281 of the Seventh Edition of Perry’s Chemical Engineers’ Handbook provide its heat of vaporization: ∆Hlv = Hv − Hl = 802. (4-165) yields 4-17 Here. the actual enthalpy change of compression is Ws(isentropic) = (∆H)S = V(P2 − P1) 223. LNG is essentially pure methane. For an isentropic process. (4-15).8) = 227.75 The actual enthalpy of superheated LNG at 20 bar is then H = 802.7 ∆H ∆H = S =  = 298.096 × 105 = 24. The assumption is that the actual power produced is 30 percent of this. but assume that a suitable engine can be built to deliver 30 percent of the work of a Carnot engine operating between the temperatures of 300 and 115 K. the temperature of the vaporizing LNG.

however. valid for any equilibrium phase.n ∂(nM) dT +  ∂P   T. This parallelism exists whenever the Symbol M may represent the molar value of any extensive thermodynamic property. each weighted by its mole fraction in solution.x ⎯ dP + Mi dni (4-172) i This equation. it provides for the calculation of solution properties from partial properties. a purpose opposite to that of Eq.4-18 THERMODYNAMICS apportioned among the individual species.nj   (4-171) T. for a binary solution an alternative method is useful.n ∂(nM) dP +  ∂ni i   dni T. (4-172). However. S. (4-171). (4-173). Moreover. imposes a constraint on how the partial properties of any phase may vary with temperature. (4-176).nj where subscript n indicates that all mole numbers ni are held constant. H. Gibbs-Duhem Equation Differentiation of Eq. Equation (4-171) applied to the definitions of Eqs. P) i where subscript x indicates that all mole fractions are held constant. (4-174) is a vital new relation. ∂M d(nM) = n  ∂T ∂M dT + n  ∂P P. P. attributes changes in total property nM to changes in T and P and to mole-number changes resulting from mass transfer or chemical reaction. postulate 5 requires that  P. the partial properties are given directly as functions of composition for given T and P. Equation (4-171). (4-28) and (4-29). say.x  and  T. the assigned property values are quite logically treated as though they were indeed properties of the individual species in solution. pressure.x Similar equations are readily derived when M takes on other identities. becomes ⎯ ⎯ x1 dM1 + x2 dM2 = 0 These two equations combine to yield ⎯ dM M1 = M + x2  dx1 (4-180a) ⎯ dM M2 = M − x1  dx1 (4-180b) Thus for a binary solution. When M H. and subscript nj signifies that all mole numbers are held constant except the ith.x ∂M dT +  ∂P  T. this is in accord with Eq. their right sides can be equated. provides a general means by which partial-property values may be determined.P. Known as the summability equation. Eq. The following are mathematical identities: dni = d(xi n) = xi dn + n dxi d(nM) = n dM + M dn Combining these expressions with Eq. (4-8) through (4-10) yields the partial-property relations ⎯ ⎯ ⎯ ⎯ ⎯ ⎯ ⎯ ⎯ ⎯ Hi = Ui + PVi Ai = Ui − TSi Gi = Hi − TSi These equations illustrate the parallelism that exists between the equations for a constant-composition solution and those for the corresponding partial properties. Thus a solution property apportioned according to the recipe of Eq. (4-173) are both valid in general. These ⎯ two equations lead through derivation to useful expressions for Bi.P.x ⎯ dP − xi dMi = 0 (4-175) i This general result. and direct use of Eq. as shown in detail by Van Ness and Abbott [Classical Thermodynamics of Nonelectrolyte Solutions: With Applications to Phase .P. the Gibbs-Duhem equation for a solution at given T and P. (4-171) is appropriate.x ⎯ H = xiHi (4-178) i ∂V CP dT + V − T  ∂T  ⎯   dP − x dH = 0 i i (4-179) i P. n1. yielding The total differential of nM is therefore  ⎯ ⎯ dM = xi dMi + Mi dxi i nM = M(T. which defines a partial molar property. whence ∂M dM =  ∂T ∂M dT +  ∂P P. …) ∂(nM) d(nM) =  ∂T solution properties in the parent equation are related linearly (in the algebraic sense). Once an apportioning recipe is adopted. and Eqs.x ⎯ M = xi Mi ⎯ dP + Mi dxi ∂M  ∂T P.nj The basis for calculation of partial properties from solution properties is provided by this equation. For a homogeneous PVT system. and composition. Partial Molar Equation-of-State Parameters The parameters in equations of state as applied to mixtures are related to composition by mixing rules.x  ∂(nM) dP +  ∂ni T. Equation (4-174) for a binary solution is ⎯ ⎯ M = x1M1 + x2M2 Moreover. and (4-175) specialize to ∂V dH = CP dT + V − T  ∂T  ⎯   dP + H dx i i (4-177) i P.x   T. (4-171) may be recovered simply by adding the properties attributed to the individual species. (4-172) and collecting like terms give ∂M dM −  ∂T  ∂M dT −  ∂P P. (4-174) yields (4-173) i (4-174) i The first of these equations is merely a special case of Eq. n3. For the special case where T and P are constant. For the second virial coefficient B = yiyjBij (4-60) i j The partial molar second virial coefficient is by definition ∂(nB) ⎯ Bi  ∂ni   (4-181) T. and reasoning on this basis leads to valid conclusions. the derivatives (∂H/∂H)P and (∂H/∂P)T are given by Eqs. n2.x ⎯ ⎯ dP − Mi dxi n + M − Mi xi dn = 0   i i  Because n and dn are independent and arbitrary.x i    dni T. the terms in brackets must separately be zero. Eq. the Gibbs-Duhem equation. The derivatives in the summation are ⎯ called partial molar properties. or G. (4-171). For multicomponent solutions such calculations are complex. This equation may also be written ∂M d(nM) = n  ∂T ∂M dT + n  ∂P P. ⎯ (4-176) xi dMi = 0 (constant T. either closed or open. (4-174).x   T. They are given the generic symbol Mi and are defined by ∂(nM) ⎯ Mi  ∂ni i Because this equation and Eq. U.nj Because B is independent of P. V. in which case m replaces n and {xi} are mass fractions. The equations for partial molar properties are valid also for partial specific properties.

(4-196). McGraw-Hill. (4-114)] used for aki. The Gibbs energy. and provides a conceptual basis upon which to build the structure of solution thermodynamics. (4-196) through (4-198) in that P has units. (4-172) Gid = yiΓi(T) + RT ln (yiP) i (4-188) (4-189) (4-190) These equations again illustrate the fact that for every equation providing a linear relation among the thermodynamic properties of a constant-composition solution there exists a parallel relationship for the partial properties of the species in solution. The entropy of an ideal gas does depend on pressure. (4-14). a quantity that takes the place of µi but that does not exhibit its less desirable characteristics. This equation serves as a partial definition of the fugacity fi. the chemical potential exhibits certain unfortunate characteristics that discourage its use in the solution of practical problems. Eq. (4-199). is realistic in the (4-198) i A dimensional ambiguity is apparent with Eqs. µi approaches negative infinity when either P or yi approaches zero. is defined in relation to the internal energy and entropy.n ∂P T.. in view of Eqs. and (4-16). Consistency in the units of pressure is. moreover.n     j ⎯ = − Si (4-186) T. and (4-104): a⎯ = 2 y a − a (4-183) k i k ki This expression is independent of the combining rule [e. Smith. is analytically simple. both primitive quantities for which absolute values are unknown. (4-15). ⎯ Gi nG nV nH d  =  dP − 2 dT +  dni RT RT RT i RT (4-192) (4-196) where integration constant Γi(T) is a function of temperature only. While these characteristics do not preclude the use of chemical potentials. an analogous equation is written as Gi Γi(T) + RT ln fi (4-199) in which a new property fi replaces the pressure P. 137–140. Subtraction of Eq.nj ⎯ ∂µ ∂µ dµi dGi = i dT + i ∂T P. Fugacity and Fugacity Coefficient The chemical potential µi plays a vital role in both phase and chemical reaction equilibria. P). In combination with Eqs. Gigi = H igi − TSiig. New York (1982)]. Equation (4-195) now becomes ⎯ (4-197) µ igi = Gigi = Γi(T) + RT ln (yiP) dP Similarly. The origin of the fugacity concept resides in Eq. both written for the same temperature and pressure. (4-13) produces not only the Maxwell relation. New York (2005)] develop the following property relations for the ideal gas model. pp.n   j Comparison with Eq. (4-103). and here ⎯ig (4-194) Si = Sigi − R ln yi where Sigi is evaluated at the mixture T and P. pp. SOLUTION THERMODYNAMICS Ideal Gas Mixture Model The ideal gas mixture model is useful because it is molecularly based.P.P.SYSTEMS OF VARIABLE COMPOSITION Equilibria.n ∂P ⎯ ⎯ ⎯ dGi = − Si dT + Vi dP  and ⎯ ig RT V i = V igi =  P  T. (4-171) indicates the following identity: ⎯ (4-184) µi = Gi The reciprocity relation for an exact differential applied to Eq. required. (4-13). gives fi Gi − Gigi = RT ln  P . For a real fluid. In practice this is of no consequence. and hence µi. (4-114) for parameter a in the generic cubic equation of state given by Eqs. but also two other useful equations: ∂(nV) ⎯ ∂µi =  = Vi (4-185)  ∂ni T. Integration gives Because µ = f (T. ⎯ ⎯ ⎯ From the definition of the Gibbs energy. (4-10) gives. the result of derivation is simply bi = bi. after algebraic reduction. (4-13) is the relation ∂(nG) µi =  ∂ni T. written for pure species i as an ideal gas: RT dGigi = V igi dP =  dP = RT d ln P P (constant T) G igi = Γi(T) + RT ln P By Eq.n   ∂µi  ∂T ∂(nS) = −  ∂ni P. whereas ln P must be dimensionless. Partial Molar Gibbs Energy Implicit in Eq. The simplest result is ⎯ Bi = 2 ykBki − B (4-182) k An analogous expression follows from Eq. (4-100). 4-19 (4-191) This result is a useful alternative to the fundamental property relation given by Eq. the application of equilibrium criteria is facilitated by introduction of the fugacity. (4-13). McGraw-Hill.  limit of zero pressure. Moreover. of course. Van Ness. and Abbott [Introduction to Chemical Engineering Thermodynamics.n ⎯ ⎯ ⎯ dUi = T dSi − P dVi ⎯ ⎯ ⎯ dHi = T dSi + Vi dP ⎯ ⎯ ⎯ dAi = − Si dT − P dVi (4-187) Because the enthalpy is independent of pressure. However. this becomes ⎯ Gigi = H igi − TSigi + RT ln yi ⎯ or µigi Gigi = Gigi + RT ln yi (4-195) Elimination of Gigi from this equation is accomplished through Eq. along with ratios of the quantities with units of pressure in the arguments of the logarithm. (4-193) and (4-194). Eq.g. (4-113) for b. ⎯ ig H i = H igi (4-193) ig i where S is evaluated at the mixture T and P. the enthalpy rather than the entropy appears on the right side. 391–394. because only differences in Gibbs energy appear. For the linear ⎯ mixing rule of Eq.. (4-21). with µi = Gi . All terms in this equation have units of moles. (4-196) from Eq. 7th ed. an equation valid only for pure species i in the ideal gas state.P. and for G by Eq. The following equation is a mathematical identity:  nG 1 nG d   d(nG) − 2 dT RT RT RT ⎯ Substitution for d(nG) by Eq. (4-17).

(4-207). (4-209) through (4-212): Gid = xiGi + RT xi ln xi (4-213) V id = xiVi (4-214) i i i Sid = xiSi − R xi ln xi (4-215) Hid = xiHi (4-216) i i i A simple equation for the fugacity of a species in an ideal solution follows from Eq. Thus for the special case of an ideal gas. (4-202). Eq. Evaluation of Fugacity Coefficients Combining Eq. is written for species i in a fluid mixture (4-202) µi Γi(T) + RT ln fˆi where the partial pressure yi P is replaced by fˆi. (4-174). (4-196) is recovered from Eq. (4-195). yields fˆi µi − µigi = RT ln  yiP The residual Gibbs energy of a mixture is defined by GR G − Gig. Eq. (4-199). with the understanding that φ here is for a pure species. Thus. known as the Lewis-Randall rule. liquid. (4-77) and (4-78). The definition of the fugacity of a species in solution is parallel to the definition of the pure-species fugacity. provides a useful generalized correlation for the fugacity coefficients of pure species. (4-199) are combined with Eq. and composition. the proportionality constant is the fugacity of pure species i in the same physical state as the solution and at the same T and P. as was done by Lee and Kesler [AIChE J. therefore φˆ iig = 1 and ˆfiig = yiP. A more comprehensive generalized correlation results from Eqs. This is formalized through residual properties. and Eq. (4-201)] give the alternative form (4-218) φˆ iid = φi . Equation (4-200) with Eqs. written for the special case of an ideal solution. may be applied to Eqs. applied to an ideal solution with µi replaced by Gi. Equation (4-203) is the ⎯ analog of Eq. f iig = P. (4-204)] and of φi for fi/P [Eq. shows that the fugacity of each species in an ideal solution is proportional to its mole fraction. (4-209) yields ⎯ ⎯ ∂Gidi ∂Gidi ∂Gi   =  and ∂P T. (4-48) and (4-73) yields ln φ = dP P  (Z − 1)  = (B + ωB )  P T Pr r 0 0 r 1 r (4-206) r This equation.4-20 THERMODYNAMICS According to the definition of Eq. written as ⎯ µigi = Gigi = Gigi (T.x ∂T ∂P T    ln φ1 (4-207) dP Z P Pr 0 r 1 r (4-208) Correlations may therefore be presented for φ0 and φ1. Eq. it is identified by a circumflex rather than an overbar. (4-200) with Eq. and the resulting expression reduces to (4-217) fˆiid = xi fi This equation. (4-200) and (4-116): ln φ = An alternative form is where ln φ0 By Eq. ∂Gi + R ln x =  ∂T i P. Clearly. (4-217) by xi P and substitution of φˆ iid for fˆidi xiP [Eq. all correlations for GR/RT are also correlations for ln φ. P) + RT ln yi P In combination these sets of equations provide ⎯ id (4-210) Vi = Vi ⎯id and Si = Si − R ln xi (4-211) ⎯id ⎯id ⎯id Because Hi = Gi + TSi . Thus the fugacity of species i in an ideal gas mixture is equal to its partial pressure. (4-209). are written as ⎯ ⎯ ∂Gidi ∂Gidi ⎯ ⎯ Vidi =  and Sidi = −  ∂P T. Equations (4-185) ⎯ and (4-186). which relates φi to GRi. The partial molar Gibbs energy or chemical potential of species i in an ideal gas mixture is given by Eq. For the special case of species i in an ideal solution. the Gibbs energy of pure species i in its real physical state of gas. (4-209) and (4-211) yield ⎯id Hi = Hi (4-212) The summability relation. pressure. Subtracting Eq. The dimensionless ratio fi /P is another new property called the fugacity coefficient φi. (4-40). An equation analogous to the ideal gas expression. Gi − Gigi is the residual Gibbs energy GRi. The ideal solution is similarly useful as a standard to which real solution behavior may be compared. Therefore ⎯R (4-203) Gi = RT ln φˆ i (4-204) The dimensionless ratio φˆ i is called the fugacity coefficient of species i in solution. (4-45) gives GR ln φ =  = RT dP  (Z − 1)  P P (4-205) 0 Subscript i is omitted.x P Equation (4-17) implies that ∂Gi  ∂P = Vi ∂Gi = −S  ∂T and i T fˆi φˆ i  yiP where by definition Ideal Solution Model The ideal gas model is useful as a standard of comparison for real gas behavior. x ∂T P. (4-202) becomes ⎯ µidi = Gidi = Γi(T) + RT ln fˆ idi When this equation and Eq. P) is replaced by Gi (T. Γi (T) is eliminated. GRi = 0. The ideal solution is therefore defined as one for which ⎯ (4-209) µidi = Gidi Gi(T. Because it is not a partial property. (4-209). GRi is necessarily zero.  (Z Pr 0 0 dP dP  (Z − 1)  +ω Z  P P Pr Pr r 0 0 0 r r 1 r ln φ = ln φ0 + ω ln φ1 dP − 1) r Pr and φ = (φ )(φ ) 0 1 ω This equation takes on new meaning when Gigi (T. (4-200). φi = 1. (4-197) from Eq. the fugacity of species i in solution. GRi = RT ln φi (4-200) fi φi  P (4-201) where The definition of fugacity is completed by setting the ideal gas state fugacity of pure species i equal to its pressure. substitutions by Eqs. P). 21: 510–527 (1975)]. used in conjunction with Eqs. For an ideal gas. P) + RT ln xi where superscript id denotes an ideal solution property and xi represents the mole fraction because application is usually to liquids. (4-197). and definition of a partial molar residual Gibbs energy is ⎯R the ⎯ analogous ⎯ Gi Gi − Gigi = µi − µigi . or solid at the mixture T and P. This equation is the basis for development of expressions for all other thermodynamic properties of an ideal solution. both written for the same temperature.x   Differentiation of Eq. Division of both sides of Eq.

Fundamental Residual-Property Relation Equation (4-191) is general and may be written for the special case of an ideal gas ⎯ig nGig nVig Gi nHig d  =  dP − 2 dT +  dni RT RT RT i RT Of particular interest is the partial molar excess Gibbs energy. ∆V. Property changes of mixing and excess properties are easily calculated one from the other. (4-191). Thus. E  Similarly. each excess property is zero. For an ideal solution. and therefore γi = 1. which is now to be filled out with important derived relationships. (4-203) relates φˆ i to GRi .P. whereas residual properties exist for pure species as well as for mix⎯E are defined analogously: tures. commonly called the heat of mixing. Gi = RT ln γi (4-222) ⎯E Comparison with Eq. a mixture of isomers conforms very closely to ideal solution behavior. the entropy change of mixing.SYSTEMS OF VARIABLE COMPOSITION ∆S. (4-237) in the upper left quadrant of Table 4-6. (4-223) these may be written as shown in the second column of Table 4-5. (4-40).n This equation demonstrates that ln φˆ i is a partial property with respect to GR/RT. It is therefore the basis for extension to several useful supplementary thermodynamic properties. Limited forms of this equation are particularly useful. (4-203) shows⎯that Eq. and for this special case the equations reduce to those shown in the third column of Table 4-5. and each may be combined with the defining equation for an excess property [Eq. ME M − Mid (4-219) FUNDAMENTAL PROPERTY RELATIONS BASED ON THE GIBBS ENERGY where M represents the molar (or unit-mass) value of any extensive thermodynamic property (say. fˆi γi  (4-221) xifi ⎯E Thus. U. (4-213) through (4-216) is an expression for an ideal solution property. (4-13) has as its special or canonical variables T. However. Equation (4-202) may be written as ⎯ Gi = Γi(T) + RT ln fˆi  In accord with Eq. and ∆H are the Gibbs energy change of mixing. Moreover. These properties are identical to the corresponding excess properties. given the symbol γi. this becomes ⎯id Gi = Γi(T) + RT ln xi fi  ⎯ The left side is the partial excess Gibbs energy GEi . excess properties have no meaning for pure species. and composition. Division by dP and restriction to constant T and composition lead to ⎯ ⎯id fˆi Gi − Gi = RT ln  xifi By difference 4-21 (4-241) Expressions for ∆Mid (4-228) ∆G = RT xi ln xi (4-232) (4-229) ∆V = 0 (4-233) SE = ∆S + R xi ln xi (4-230) ∆Sid = − R xi ln xi (4-234) HE = ∆H (4-231) ∆H id = 0 (4-235) E i E id i id i i i i i i . H. The summability relation therefore applies. Indeed an alternative form has been developed as Eq. providing an experimental entry into the network of equations of solution thermodynamics. where ∆G. These equations are the first two entries in the upper left quadrant of Table 4-6. G). (4-203). P. This definition is analogous to the definition of a residual property as given by Eq. (4-219)]. Partial molar excess properties M i ⎯E ⎯ ⎯id Mi = Mi − Mi (4-220) Of the four fundamental property relations shown in the second column of Table 4-1. (4-222) relates γ⎯ i to Gi exactly as Eq. Applications are usually to liquids. (4-217) for an ideal solution. Ideal solution behavior is often approximated by solutions comprised of molecules not too different in size and of the same chemical nature. Thus the fugacity coefficient of species i in an ideal solution equals the fugacity coefficient of pure species i in the same physical state as the solution and at the same T and P.x ∂(nGRRT) ln φˆ i =  (4-240) ∂ni T. The result is listed as Eq. only Eq. S. P. For an ideal solution. x ∂(GRRT) HR  = −T  ∂T RT where M represents a molar thermodynamic property of a homogeneous solution and Mi is the molar property of pure species i at the T and P of the solution and in the same physical state. Excess Properties An excess property ME is defined as the difference between the actual property value of a solution and the value it would have as an ideal solution at the same T. and {ni}. and by definiton. An alternative form follows by introduction of the fugacity coefficient given by Eq. the result of division by dT and restriction to constant P and composition is i TABLE 4-5 Subtraction of this equation from Eq. Equation (4-236) is the fundamental residual-property relation. yielding the equations of the first column of Table 4-5. and (4-224) i E  Also implicit in Eq. Property Changes of Mixing A property change of mixing is defined by ∆M M − xiMi ∂(GRRT) VR  =  ∂P RT   (4-223)   (4-225) V = ∆V SE = S − xiSi + R xi ln xi (4-226) HE = H − xiHi (4-227) V = V − xiVi j GR  = xi ln φˆ i RT i G = ∆G − RT xi ln xi i (4-239) P. So do mixtures of adjacent members of a homologous series. Thus. The most common property changes of mixing are the volume change of mixing ∆V and the enthalpy change of mixing ∆H. V. the dimensionless ˆ ratio fixi fi on the right is the activity coefficient of species i in solution. and the enthalpy change of mixing. they are directly measurable. (4-191) gives ⎯ nGR nVR GRi nHR d  =  dP − 2 dT +  dni (4-236) RT RT RT i RT ⎯ ⎯ ⎯ where the definitions GR G − Gig and GRi Gi − Gigi have been imposed. In view of Eq. (4-237) is the relation Relations Connecting Property Changes of Mixing and Excess Properties ME in relation to M ME in relation to ∆M G = G − xiGi − RT xi ln xi (4-238) T. the volume change of mixing. GEi = 0. Each of Eqs.

(4-240) and (4-181) yields P ⎯ ln φˆ i =  Bi RT (4-242) P ln φˆ 1 =  (B11 + y22δ12) RT (4-243a) P ln φˆ 2 =  (B22 + y21δ12) RT (4-243b) ⎯ where Bi is given by Eq.4-22 THERMODYNAMICS TABLE 4-6 Fundamental Property Relations for the Gibbs Energy and Related Properties General equations for an open system Equations for 1 mol (constant composition) d(nG) = nV dP − nS dT + µi dni (4-13) dG = V dP − S dT (4-17) i ⎯ Gi nH nV nG d  =  dP − 2 dT +  dni RT RT RT RT i (4-191) H V G d  =  dP − 2 dT RT RT RT (4-253) nHR nV R nGR d  =  dP − 2 dT + ln φˆ i dni RT RT RT i (4-237) HR VR GR d  =  dP − 2 dT RT RT RT (4-254) nG E nVE nHE d  =  dP − 2 dT + ln γi dni RT RT RT i (4-248) HE GE VE d  =  dP − 2 dT RT RT RT (4-255)       Equations for partial molar properties (constant composition) ⎯ ⎯ ⎯ dGi = dµi = Vi dP − Si dT Gibbs-Duhem equations (4-256) V dP − S dT = xi dµi (4-260) i   (4-257) ⎯ Gi H V  dP − 2 dT = xi d  RT RT RT i  (4-258) HR VR  dP − 2 dT = xi d ln φˆ i RT RT i (4-262)  (4-259) HE VE  dP − 2 dT = xi d ln γi RT RT i (4-263) ⎯ ⎯ ⎯ Hi G µi V d i = d  = i dP −  dT RT 2 RT RT RT ⎯ ⎯ ⎯R GRi Vi dP − HiR dT = d ln φˆ i =  d  2 RT RT RT ⎯ ⎯E ⎯E Vi GEi Hi d  = d ln γi =  dP − 2 dT RT RT RT Application of Eq. (4-248) in the upper left quadrant of Table 4-6. x (4-251) T. McGraw-Hill. (4-200) for a pure species becomes ln φi = Zi − 1 − ln(Zi − βi) − qiIi (4-245) For species i in solution Smith. (4-222). Equation (4-247) is the fundamental excess-property relation. implying also the validity of the summability relation GE  = xi ln γi RT i (4-252) The equations of the upper left quadrant of Table 4-6 reduce to those of the upper right quadrant for n = 1 and dni = 0. (4-112). (4-191) becomes ⎯id nHid Gi nGid nVid d  =  dP −  dT +  dni 2 RT RT RT i RT  nGR P  =  (nB) RT RT where  Subtraction of this equation from Eq. these equations reduce to P ln φi =  Bii RT (4-244) For the generic cubic equation of state [Eqs. (4-240) to an expression giving GR as a function of composition yields an equation for ln φˆ i. The following equations are in complete analogy to those for residual properties. (4-109). 562–563. (4-67)] is appropriate. (4-191) yields ⎯ GEi nHE nGE nVE d  =  dP − 2 dT +  dni (4-247) RT RT RT i RT ⎯ ⎯ ⎯ where the definitions GE G − Gid and GEi Gi − Gidi have been imposed. (4-172) and therefore has associated with it a Gibbs-Duhem equation of the form of Eq. In the simplest case of a gas mixture for which the virial equation [Eq. Van Ness. and Abbott [Introduction to Chemical Engineering Thermodynamics. This result is listed as Eq. The equations of Table 4-6 store an enormous amount of information. (4-261) Fundamental Excess-Property Relation Equations for excess properties are developed in much the same way as those for residual properties. Each equation in the upper left quadrant has a partial-property analog. (4-182). and I is given by Eq. These are shown in the lower right quadrant. An alternative form follows by introduction of the activity coefficient as given by Eq.nj This last equation demonstrates that ln γi is a partial property with respect to GE/RT. For the special case of an ideal solution. x   (4-250) P. Each equation of the upper left quadrant is a special case of Eq. For a binary system these equations reduce to δ12 2B12 − B11 − B22 For the special case of pure species i. GR/RT is given by Eq. (4-173). as shown in the lower left quadrant. P. New York (2005)] show that ⎯ bi qiI (4-246) ln φˆ i =  (Z − 1) − ln(Z − β) − ⎯ b Symbols without subscripts represent mixture properties. Eq. which in view of Eq. pp. Eq..  ∂(GERT) VE  =  ∂P RT   ∂(GERT) HE  = −T  ∂T RT  ∂(nGERT) ln γi =  ∂ni  (4-249) T. 7th ed. (4-104)]. but they are so general that their direct . (4-69) provides the relation Differentiation in accord with Eqs.

by inspection one can write a vast array of relations valid for particular applications. the variables usually specified or sought in design calculations. and the van Laar equations are all special cases of a very general treatment based on rational functions. and composition. when x2 = 0. Theoretical developments in the molecular thermodynamics of liquid solution behavior are often based on the concept of local composition. The Redlich-Kister expansion. and composition. 48: 49–61 (1952)]: GE  = A + B(x1 − x2) + C(x1 − x2)2 + · · · x1x2RT In application. Thus. Wilson [J. M.x RT ⎯R Hi ∂ ln φˆ i (4-265)  = − 2 RT ∂T P. The corresponding equations for ln γ1 and ln γ2 are ln γ1 = Ax22 (4-268a) ln γ 2 = Ax21 (4-268b) The symmetric nature of these relations is evident.x ⎯E Vi ∂ ln γi (4-266)  =  ∂P T. most notably the NRTL (Non-Random Two-Liquid) equation of Renon and Prausnitz [AIChE J. . Chem. (4-251) yields ln γ1 = x22 [A12 + 2(A21 − A12) x1] (4-270a) ln γ2 = x [A21 + 2(A12 − A21) x2] (4-270b) 2 1 When x1 = 0. ln γ 1∞ = −ln Λ12 + 1 − Λ21 ln γ 2∞ = −ln Λ21 + 1 − Λ12 Both Λ12 and Λ21 must be positive numbers. 86: 127−130 (1964)] with the publication of a model for GE. An alternative equation is obtained when the reciprocal quantity x1x2RT/GE is expressed as a linear function of x1: x1x2  = A′ + B′(x1 − x2) = A′ + B′(2x1 − 1) GE/RT This may also be written as x1x2  = A′(x1 + x2) + B′(x1 − x2) = (A′ + B′)x1 + (A′ − B′) x2 GE/RT The substitutions A′ + B′ = 1/A′21 and A′ − B′ = 1/A′12 ultimately produce MODELS FOR THE EXCESS GIBBS ENERGY GE  = a + bx1 + cx21 + · · · x1x2RT 4-23 (4-269) A′12 A′21 GE  =  A′12x1 + A′21x2 x1x2RT (4-271) The activity coefficients implied by this equation are given by −2 A′12 x1 ln γ1 = A′12 1 +  A′21 x2  A′21x2 ln γ2 = A′21 1 +  A′12 x1  (4-272a) −2 (4-272b) These are the van Laar equations. The infinite dilution values of the activity coefficients are ln γ 1∞ = ln γ 2∞ = A.. presumed to account for the short-range order and nonrandom molecular orientations resulting from differences in molecular size and intermolecular forces. ln γ 2∞ = A′21. and in accord with Eq. If C = · · · = 0. one can in principle compute from it all other excess properties. on equations for GE given by ratios of polynomials [Van Ness and Abbott. No. 5-7. However. The simplest procedure is to express this quantity as a power series in x1: (constant T) An equivalent power series with certain advantages is the RedlichKister expansion [Redlich. Of primary importance for engineering calculations is the excess Gibbs energy GE. when x2 = 0. P. they are without theoretical foundation. then E G  = A + B(x1 − x2) = A + B(2x1 − 1) x1x2RT and in this case GE/x1x2RT is linear in x1. the Margules equations. When all parameters are zero. New York (1982)]. Kister. and for each truncation specific expressions for ln γ1 and ln γ2 result from application of Eq. GE/RT = 0. Soc. and the solution is ideal. = 0. and the quantity most conveniently represented by an equation is GE/x1x2RT. ln γ 2∞ = A21. ln γ 1∞ = A12. with no basis in theory for their extension to multicomponent systems.x RT ⎯E Hi ∂ ln γi (4-267)  = − 2 RT ∂T P. engineering efforts to model GE center on representing its composition and temperature dependence. P. Knowing GE as a function of T. 14: 135−144 (1968)] and the UNIQUAC (UNIversal QUAsi-Chemical) equation of Abrams and Prausnitz [AIChE J. Introduced by G. like the Margules and van Laar equations. When x1 = 0. If B = C = . The excess volume for liquid mixtures is usually small. different truncations of this expansion are appropriate. GE becomes a function of just x1. (4-258) and (4-259) that ⎯R Vi ∂ ln φˆ i (4-264)  =  ∂P T. (4-249) the pressure dependence of GE is usually ignored. Although providing great flexibility in the fitting of VLE data for binary systems. Ser. 2. ln γ 1∞= A′12. Progr. The Wilson equation. one sees immediately from Eqs.x     Excess properties find application in the treatment of liquid solutions. . Classical Thermodynamics of Nonelectrolyte Solutions: With Applications to Phase Equilibria. For binary systems at constant T. Eng. then GE  =A x1x2RT where A is a constant for a given temperature. McGraw-Hill. this concept prompted the development of alternative local composition models. 21: 116−128 (1975)]. Am.e. contains just two parameters for a binary system (Λ12 and Λ21) and is written as GE  = − x1 ln(x1 + x2Λ12) − x2 ln(x2 + x1Λ21) RT (4−273) Λ12 Λ21 ln γ1 = − ln (x1 + x2Λ12) + x2  −  x1 + x2Λ12 x2 + x1Λ21 (4-274a) Λ21 Λ12 ln γ2 = − ln (x2 + x1Λ21) − x1  −  x1 + x2Λ12 x2 + x1Λ21 (4-274b)   At infinite dilution. .SYSTEMS OF VARIABLE COMPOSITION application is seldom appropriate. Chem. and Turnquist. because its canonical variables are T. (4-251). i. For example. Nor do they incorporate an explicit temperature dependence for the parameters. The substitutions A + B = A21 and A−B = A12 transform this expression to the Margules equation: E G  = A21x1 + A12x2 x1x2RT Application of Eq. Symp. Sec.

The temperature dependence of the parameters is given by −aij Vj Λij =  exp  i≠j (4-279) RT Vi where Vj and Vi are the molar volumes of pure liquids j and i and aij is a constant independent of composition and temperature. form ratios that in practice may be taken as independent of T. (e) ethanol(1)/chloroform(2). For example. and a residual term gR (not a residual property). but they are adequate for most engineering purposes.THERMODYNAMICS The NRTL equation contains three parameters for a binary system and is written as   +  (x + x G )  b12 τ12 =  RT 2 1 12 2 G21τ21 2 1 2 21 2 (4-276a) (4-276b) G21 = exp(−ατ21) b21 τ21 =  RT where α. they are implicitly generalizable to multicomponent systems without the introduction of any parameters beyond those required to describe the constitutent binary systems. a combinatorial term gC.] FIG. accounting for molecular size and shape differences.. The infinite dilution values of the activity coefficients are ln γ ∞1 = τ21 + τ12 exp (−ατ12) ln γ ∞2 = τ12 + τ21 exp (−ατ21)  x Λ 2000 T S j (4-277) ij H 0 0 T S H 0 1 x1 G 0 x1 (a) G x1 (d) 1000 H G 0 1 1 x1 (c) 1000 H 0 1000 1000 0 T S T S J mol1 J mol1 0 (4-280) 2000 1000 H 1 x1 (e) TS 0 H G 1000 G 0 1 0 x1 1 (f) Property changes of mixing at 50!C for six binary liquid systems: (a) chloroform(1)/n-heptane(2). p. (d) ethanol(1)/n-heptane(2). Moreover. the Wilson equation for multicomponent systems is written as GE  = − xi ln RT i j j J mol1 G12 = exp(−ατ12) and G12τ12 +  (x + x G )  G12 ln γ2 = x21 τ12  x2 + x1G12 Here (4-275)   2 xkΛki  x Λ −  x Λ j G21τ21 G12τ12 GE  =  +  x1 + x2G21 x2 + x1G12 x1x2RT G21 ln γ1 = x22 τ21  x1 + x2G21 ln γi = 1 − ln and J mol1 4-24 . and all summations are over all species.. Molar volumes Vj and Vi. because Λij ≠ Λji. and this makes parameter determination for the local composition models a task of manageable proportions. For example. parameters specific to a particular pair of species.g. The UNIQUAC equation treats g GERT as made up of two additive parts. all parameters are found from data for binary (in contrast to multicomponent) systems. Λ31. and Abbott. (f ) ethanol(1)/water(2). themselves weak functions of temperature. but it is only approximate. Λ21. 455. Introduction to Chemical Engineering Thermodynamics. 7th ed. (c) acetone(1)/chloroform(2). However. New York (2005). Λ13. All indices in these equations refer to the same species. For each ij pair there are two parameters. Λ32. 4-4 (4-278) kj 0 (b) TS 1000 j 1000 1000 1000 G k g = gC + gR 2000 1000 1000 ij where Λij = 1 for i = j. heats of mixing and liquid/liquid solubility) are in general only qualitatively correct. in a ternary system the three possible ij pairs are associated with the parameters Λ12. etc. accounting for molecular interactions: j J mol1 J mol1 The local composition models have limited flexibility in the fitting of data. are independent of composition and temperature. Computations of properties sensitive to this dependence (e. and Λ23. and are usually evaluated at or near 25°C. b12. [Smith. McGraw-Hill. and b21. The Wilson parameters Λij and NRTL parameters Gij inherit a Boltzmann-type T dependence from the origins of the expressions for E G . (b) acetone(1)/ methanol(2). Van Ness.

Chemistry Data Series. parts 1–8. and Arlt [VaporLiquid Equilibrium Data Collection. (4-280) through (4-282)]. Values for the parameters of the commonly used models for the excess Gibbs energy are given by Gmehling. The influence of temperature on g enters through the interaction parameters τji of Eq. xi qi θi  xj qj and (4-286) (4-284) j ri Ji =  rj xj (4-289) qi Li =  qj xj (4-290) si = θ l τli (4-291) j Subscript i identifies species.. The result is given by the following equations: Φi θi g C = xi ln  + 5 qi xi ln  xi Φi i i ln γ i = ln γ iC + ln γ Ri g R = − qi xi ln i (4-281)  j θ τ j ji xiri Φi  xjrj where (4-282) Ji Ji ln γ iC = 1 − Ji + ln Ji − 5qi 1 −  + ln  Li Li (4-283) τij ln γ iR = qi 1 − ln si − θj  sj j j   (4-287) (4-288) where in addition to Eqs. DECHEMA. Again subscript i identifies species. In these equations ri (a relative molecular volume) and qi (a relative molecular surface area) are pure-species parameters. 4-5 . all summations are over all species. 7th ed. 420. which are temperature-dependent: j l − (uji − uii) τji = exp  RT (4-285) Parameters for the UNIQUAC equation are therefore values of uji − uii. (4-284) and (4-285). 0 TS E GE J mol1 500 H J mol1 J mol1 HE E GE 500 TS E GE 1000 TS E HE 0 0 1 0 1 x1 x1 (a) (b) (c) GE HE GE J mol1 500 0 TS 500 GE 500 0 E H E 0 500 1 J mol1 J mol1 1000 1 x1 HE TS E 1000 TS E 0 0 1 0 1 x1 x1 x1 (d ) (e) (f ) Excess properties at 50!C for six binary liquid systems: (a) chloroform(1)/n-heptane(2). (b) acetone(1)/methanol(2).] FIG. 1. Onken. then τii = τjj = 1. however. (4-282). (f ) ethanol(1)/water(2). (d) ethanol(1)/n-heptane(2). (c) acetone(1)/chloroform(2). when i = j. and j is a dummy index. [Smith. New York (2005). An expression for ln γ i is found by application of Eq. whereas function gR incorporates two binary parameters for each pair of molecules. and Abbott. p. (e) ethanol(1)/chloroform(2). Note that τji ≠ τij. and j and l are dummy indices. McGraw-Hill. Frankfurt/Main (1974–1990)]. For a multicomponent system. (4-251) to the UNIQUAC equation for g [Eqs. vol. Van Ness. Introduction to Chemical Engineering Thermodynamics.SYSTEMS OF VARIABLE COMPOSITION 4-25 Function gC contains pure-species parameters only.

the heat of mixing (HE = ∆H). (f) ethanol(1)/water(2). HE ( ∆H). For example. derived from Eq.. and ∆G as functions of composition for six binary solutions at 50°C.P ≤ 0 . ∆S. at temperatures well below the critical temperature of each constituent species.P ≤ 0 or dGtT. then changes in the system are irreversible and necessarily bring the system closer to an equilibrium state.e. The properties shown here are insensitive to pressure and for practical purposes represent solution properties at 50°C and low pressure (P ≈ 1 bar). and the excess Gibbs energy GE are experimentally accessible. McGraw-Hill.4 In 1 0.6 0. (4-230). [Smith. The activity coefficient is also of special importance because of its application in phase equilibrium calculations. Constraints put on this expression produce alternative criteria for the directions of irreversible processes and for the condition of equilibrium. The corresponding excess properties are shown in Fig. i. dW = −PdVt. ∆S. New York (2005). If a closed PVT system of uniform T and P. whereas the equality holds at the equilibrium state where change is reversible. appear in Fig.5 In 1 1. Figure 4-4 displays plots of ∆H. EQUILIBRIUM dUt − dW − TdSt ≤ 0 CRITERIA Combination gives The equations developed in preceding sections are for PVT systems in states of internal equilibrium.] FIG. 7th ed. 4-6 Behavior of Binary Liquid Solutions Property changes of mixing and excess properties find greatest application in the description of liquid mixtures at low reduced temperatures. ∆V and ∆H by direct measurement and GE (or ln γ i ) indirectly by reduction of vapor/liquid equilibrium data.0 In 2 1. p. (4-251). (e) ethanol(1)/chloroform(2).2 0. this produces t t d(Ut + PVt − TSt)T. the activity coefficients. (d) ethanol(1)/n-heptane(2). 445. The criteria for phase and chemical reaction equilibria are less obvious. Knowledge of HE and GE allows calculation of SE by Eq.4-26 THERMODYNAMICS 0 0. The first and second laws written for the entire system are dQ dUt = dQ + dW dSt ≥  T Because mechanical equilibrium is assumed. (b) acetone(1)/methanol(2).8 0 1 x1 0 (a) In 1 1. is in thermal and mechanical equilibrium with its surroundings.6 0. SE.5 x1 1 x1 (c) 1. Van Ness.5 0.4 0. and ∆G.V ≤ 0. and Abbott.. The properties of interest to the chemical engineer are VE ( ∆V). but is not at internal equilibrium with respect to mass transfer or chemical reaction. dUSt .4 In 2 In 1 0. The criteria for internal thermal and mechanical equilibrium simply require uniformity of temperature and pressure throughout the system. The volume change of mixing (VE = ∆V). GE.2 In 2 In 2 In 1 0.2 0. whence dUt + PdVt − TdSt ≤ 0 The inequality applies to all incremental changes toward the equilibrium state. (c) acetone(1)/chloroform(2). Particularly important is fixing T and P.5 0 0 (b) 3 2 1 x1 1 0 1 x1 (d ) 0 1 x1 (e) (f) Activity coefficients at 50!C for six binary liquid systems: (a) chloroform(1)/n-heptane(2).0 In 1 In 2 In 2 1 0. 4-6. (4-10).6 0. Introduction to Chemical Engineering Thermodynamics. 4-5. written for excess properties as HE − GE SE =  (4-292) T with ∆S then given by Eq. either homogeneous or heterogeneous.

b. CH4 + H2O → CO + 3H2 CO + H2O → CO2 + H2 CH4 + 2H2O → CO2 + 4H2 The result. Combine these reaction equations so as to eliminate from the set all elements not present as elements in the system. etc. material balances therefore require that dn″i = −dn′i. Thus the equilibrium state of a closed system is the state with the minimum total Gibbs energy attainable at the given T and P. for equilibrium states not all these variables are independent. giving r equations The total number of independent equations is therefore (π − 1)N + r.. reducing the preceding equation to i (µ′ − µ″)dn′ = 0 i i Because the dn′i are independent and arbitrary. PHASE RULE The intensive state of a PVT system is established when its temperature and pressure and the compositions of all phases are fixed. or P and x1. because they have nothing to do with the intensive state of the system. capable of exchanging mass with the other. Equation (4-296) for each independent chemical reaction. it applies to both closed and open systems. (4-13) is written for each phase: d(nG)′ = −(nS)′ dT + (nV)′ dP + µ′i dni′ phase. (4-13) and (4-293) leads immediately to i µ dn = 0 i i F=2−π+N−r (4-297) The number of independent chemical reactions r can be determined as follows: 1. the equilibrium state is completely .P d(nG)T. and fixing a limited number of them automatically establishes the others. although two or more elements may be simultaneously eliminated. Write formation reactions from the elements for each chemical compound present in the system. for example. and all pairs are equivalent. This number is the difference between the number of variables needed to characterize the system and the number of equations that may be written connecting these variables. an equation not taken into account in the derivation of the phase rule. It is readily generalized to multiple phases by successive application to pairs of phases. (4-202) produces the equivalent result: (4-295) fˆi ′ = fˆi″ = fˆi″′ = · · · These are the criteria of phase equilibrium applied in the solution of practical problems. However.P = 0 (4-293) This equation may be applied to a closed. one cannot in addition require that the system form an azeotrope (assuming this is possible). is a rule relating to closed systems only: For any closed system formed initially from given masses of prescribed chemical species. no further specification is possible that would restrict the phase rule variables. The equilibrium equations that may be written express chemical potentials or fugacities as functions of T. For the case of equilibrium with respect to chemical reaction within a single-phase closed system. giving (π − 1)N phase equilibrium equations 2. for this requires x1 = y1. F = 2 + (N − 1)π − (π − 1)N − r i d(nG)″ = −(nS)″ dT + (nV)″ dP + µ″i dn″i or i where the primes and double primes denote the two phases. making F = 1. and N − 1 mole fractions in each The 2 degrees of freedom for this system may be satisfied by setting T and P. This number of independent variables is given by the phase rule. P. Equation (4-295) for each species.P = µ′i dni′ + µi″dn″i = 0 i i The system is closed and without chemical reaction. CO2. Example 2: Application of the Phase Rule a. A systematic procedure is to select one equation and combine it with each of the other equations of the set so as to eliminate a particular element. this sum is given by d(nG) = (nV)dP − (nS)dT + µ′i dn′i + µ″i dni″ i i If the two-phase system is at equilibrium. means that one is free to specify. Once the 2 degrees of freedom are used up. depending on how the combination of equations is effected. differential variations may occur in the system at constant T and P without producing a change in Gt. F = 4. but all sets number r and are equivalent. P. the phase rule variables: 1. The resulting set of r equations is a complete set of independent reactions. nonreactive. (4-293) yields dGtT. provided nothing else is arbitrarily set. The masses of the phases are not phase rule variables. Thus for equilibrium at a particular T and P. Three possible pairs of equations may result. This usually reduces the set by one equation for each element eliminated. H2. For a system containing N chemical species distributed at equilibrium among π phases. at fixed values. Eq. Duhem’s theorem. When each total-system property is expressed by an equation of the form nM = (nM)′ + (nM)″. The change in the Gibbs energy of the two-phase system is the sum of these equations. Duhem’s Theorem Because the phase rule treats only the intensive state of a system. on the other hand. or T and y1. This is the criterion of two-phase equilibrium. 2. F=2−π+N−r=2−2+2−0=2 i µi′ = µi″ = µi″′ = · · · 4-27 (4-296) For a system in which both phase and chemical reaction equilibrium prevail. At the equilibrium state. Thus it cannot simultaneously be required that the system be prepared from specified amounts of particular constituent species. and CH4 in chemical reaction equilibrium. For a gaseous system consisting of CO. Thus the requirement that the system form an azeotrope imposes a special constraint. Each phase taken separately is an open system. The general result is (4-294) Substitution for each µi by Eq. presumed uniform throughout the system. it follows that µ′i = µ″i. For a system of two miscible nonreacting species in vapor/liquid equilibrium. this state (if possible at all) exists only at one liquid and one vapor composition. the presumption is that T and P are uniform throughout the two phases. then application of Eq. and it is called the number of degrees of freedom of the system. The number of these variables is 2 + (N − 1)π. This is the meaning of the equilibrium criterion dGtT. T. the criteria of Eqs. For example. It is the number of variables that may be arbitrarily specified and that must be so specified in order to fix the intensive state of a system at equilibrium. and the phase compositions. H2O. two-phase system. Any pair of the following three equations represents a complete set of independent reactions. F=2−π+N−r=2−1+5−2=4 The value of r = 2 is found from the formation reactions: C + 12O2 → CO C + O2 → CO2 H2 + 12O2 → H2O C + 2H2 → CH4 Systematic elimination of C and O2 from this set of chemical equations reduces the set to two.EQUILIBRIUM This expression shows that all irreversible processes occurring at constant T and P proceed in a direction such that the total Gibbs energy of the system decreases. More than one such set is often possible. combination of Eqs. (4-295) and (4-296) are superimposed. the phase rule variables are T and P. and two mole fractions in an equilibrium mixture of these five chemical species. Because the degrees of freedom of the system F is the difference between the number of variables and the number of equations. or x1 and y1.

This is the incentive for application of thermodynamics to the calculation of phase equilibrium relationships. most commonly by Eq. . . (4-251) applied to an expression for GE/RT. on the other hand. . it is useful as a standard of comparison. the remaining N variables can be determined by simultaneous solution of the N equilibrium relations (4-298) fˆi l = fˆiv i = 1. (4-303) in conjunction with Eq. First. the pressure is low enough that a relatively simple equation of state. 2. When Eq. The remaining N variables are then subject to calculation. (4-303) reduces to yiP = xiP sat i i = 1. . and N − 1 vapor-phase mole fractions. . . P. (4-305) may be solved for N unknown phase equilibrium variables. VAPOR/LIQUID EQUILIBRIUM Vapor/liquid equilibrium (VLE) relationships (as well as other interphase equilibrium relationships) are needed in the solution of many engineering problems. respectively. The required data can be found by experiment. and the right side of Eq. γi xi fi = φˆ iyi P i = 1. The fugacity of species i in the liquid phase is given by Eq. For ideal solutions. . 2. N (4-306) an equation which expresses Raoult’s law. . (4-205). not only are T. 2. These are here written as fˆil = γi xi fi and fˆiv = φˆ vi yiP By Eq. but well suited to computer solution. this reduces to ⎯ l sat PBi − Psat i Bii − Vi (P − Pi ) Φi = exp  RT (4-305) ⎯ where Bi is given by Eq. For this isothermal change of state from saturated liquid at Pisat to liquid at pressure P. fugacity coefficients φˆ i and φsat i are unity. (4-182). Liquidphase behavior. N (4-299) Identifying superscripts l and v are omitted here with the understanding that γi and fi are liquid-phase properties. Equation (4-298) written for pure species i becomes fiv = fil = fisat (4-300) where fisat indicates the value both for saturated liquid and for saturated vapor. For a multicomponent system the calculation is formidable. whereas φˆ i is a vaporphase property. (4-244) and (4-243). . For the systems of interest here. (Note that Σixi = 1 and Σiyi = 1. is a very weak function of P at temperatures well below Tc. Applications of Eq. an excellent approximation is usually obtained when evaluation of the integral is based on the assumption that Vi is constant at the value for saturated liquid Vil: fi Vil(P − Psat i ) ln  =  sat fi RT sat sat Substituting f sat i = φ i P i and solving for fi give Vil(P − P sat i ) sat fi = φ sat i P i exp  RT (4-302) The exponential is known as the Poynting factor. . Eq. This is done in two steps. provided only that the two properties are independently variable at the equilibrium state. (4-298). The meaning of completely determined is that both the intensive and extensive states of the system are fixed. as described in the section “Models for the Excess Gibbs Energy. N (4-307) . N where superscripts l and v denote the liquid and vapor phases. (4-304) reduces to the Poynting factor. the result is the modified Raoult’s law: yiP = xi γi P sat i i = 1. Real solution behavior is reflected by values of activity coefficients that differ from unity. P. either T or P and either the liquid-phase or vapor-phase composition are specified. such as the two-term virial equation.” The fugacity fi of pure compressed liquid i must be evaluated at the T and P of the equilibrium mixture.4-28 THERMODYNAMICS determined by any two properties of the system. the activity coefficients γi are also unity. provided that sufficient information is available to allow determination of all necessary thermodynamic properties. (4-204). and application of the phase rule shows that exactly N of these variables must be fixed to establish the intensive state of the system. and the vapor-phase fugacity is given by Eq. the one least often reasonable is the supposition of solution ideality for the liquid phase.) Thus there are 2N independent variables. (4-299) represent what is known as the gamma/phi approach to VLE calculations. and they become ever more difficult as the number of species increases. Subtraction provides another expression for Gi − Gisat: fi Gi − Gisat = RT ln  fisat Equating the two expressions for Gi − Gisat yields 1 fi ln  =  fisat RT  P Vi dP sat Pi Because Vi. . this factor is always very close to unity. even for binary systems. . (4-242) evaluated for pure species i at temperature T and the corresponding vapor pressure P = Pisat. N − 1 liquid-phase mole fractions. (4-17) is integrated to give Gi − Gisat =  P Vi dP sat Pi Equation (4-199) is then written twice: for Gi and for Gsat i . For ideal gases. respectively. Evaluation of φˆ i is usually by Eq. is satisfactory for the vapor phase. (4-303) is applied to VLE for which the vapor phase is an ideal gas and the liquid phase is an ideal solution. . but measurements are seldom easy. it reduces to a very simple expression. 2. . The general VLE problem treats a multicomponent system of N constituent species for which the independent variables are T. . When γi of Eq. Gamma/Phi Approach For many VLE systems of interest. (4-221). and for practical purposes Φi = 1. . (4-243). . This means that once N variables have been specified. one calculates the fugacity coefficient of saturated vapor φvi = φisat by an integrated form of Eq. 2. . . Division by Pisat yields corresponding fugacity coefficients: fisat fiv fil =  =   sat sat Pi Pi P sat i or φvi = φ li = φ sat i (4-301) The second step is the evaluation of the change in fugacity of the liquid with a change in pressure to a value above or below Pisat. It is the simplest possible equation for VLE and as such fails to provide a realistic representation of real behavior for most systems. (4-303) is retained in the equilibrium equation. Nevertheless. Modified Raoult’s Law Of the qualifications that lead to Raoult’s law. is described by an equation for the excess Gibbs energy. . Equation (4-299) may now be written as where yiPΦi = xi γi Psat i = 1. from which activity coefficients are derived. the liquid-phase molar volume. where xi and yi represent liquid and vapor mole fractions. and the phase compositions established. The N equations represented by Eq. N i −Vil(P − P sat φˆ i i ) Φi = sat exp  RT φi (4-303) (4-304) If evaluation of φ isat and φˆ i is by Eqs. and Eq. but so also are the masses of the phases. The activity coefficient γi is ultimately based on Eq. thus fixing 1 + (N − 1) = N independent variables. based on the two-term virial equation of state. In practice.

the test of whether an azeotrope exists at a given T or P is provided by values of α12 in the limits of x1 = 0 and x1 = 1. Amsterdam (1977)]. Shown in the accompanying table are calculated azeotropic states for a temperature of 46!C and for a pressure of 101. 20: 331–339 (1981)]. Not only do they provide a wide temperature range of applicability. neither the vapor pressures nor the activity coefficients can be initally calculated. Eq. given x1 and T. an iteration scheme based on Eqs. Ind. Process Des. Ind. With x1 an unknown..616) x1γ1Psat 1 y1 =  =  = 0. and Gmehling [Ind.8053)(87. Dev. Its most recent revision and extension is treated by Wittig. including phase equilibria.61 When T and x1 are given. Eng. P.05 a21 cal mol−1 453. DEW P calculation: Find x1 and P. results are listed in the accompanying table. [Ind. given y1 and P.22 2696.57 V1 cm3 mol−1 74. [Ind.33 kPa.5851 P 108.33 kPa.134 kPa (0. (4-307)] may therefore be applied to the calculation of any pair of phase rule variables. For the system acetone(1)/n-hexane(2). Dev. (4-307) may be solved for xi. solid/liquid equilibria [Anderson and Prausnitz. Fluid Phase Equilib.32 and P = 80 kPa are listed in the accompanying table. an azeotrope necessarily exists at the given T or P. 25: 22–31 (1986)]. and excess enthalpies [Dang and Tassios. 41: 1678–1688 (2002)]. However. vapor pressures are given by Eq. Ind.616) + (0. infinite dilution activity coefficients. the Wilson equation: ln γ1 = −ln(x1 + x2Λ12) + x2λ (B) ln γ2 = −ln(x2 + x1Λ21) − x1λ (C) Λ12 Λ21 λ  −  x1 + x2Λ12 x2 + x1Λ21 where − aij Vj Λij =  exp  RT Vi By Eq. BUBL P calculation: Find y1 and P.4 and T = 325. Chem. If both values are either > 1 or < 1. Eng. Ind. in which case summing over all species yields y1/x1 α12  y2/x2 1 P =  yi /γiP sati Because yi = xi for the azeotropic state. The most recent revision and extension of the modified UNIFAC (Dortmund) model is provided by Gmehling et al. At 46°C. and calculations require trial or iteration. Chem. . Skjold-Jørgensen. and (C). Equations (4-308) and (4-307) then become sat P = x1γ1P sat 1 + x2 γ2 P 2 = (0. Ind. Vapor-Liquid Data Collection. and is part of any solve routine of a software package. regardless of the formulation of the equilibrium equations. defined as i Alternatively. With T unknown. e. The most widely applied such correlations are based on the UNIQUAC equation. given x1 and P. 41: 3260–3273 (2003)]. wherein activity coefficients are found as sums of contributions from the structural groups that make up the molecules of a solution.60)(1. (4-307) may be summed over all species to yield P = xi γi P sat i (4-308) 4-29 with parameters i 1 2 Ai 14. In all other cases either T or x1 or both are initially unknown. (B). Substitution for the two ratios by Eq. (A). Chemistry Data Series.105) = 108. Frankfurt/Main (1983)] a12 cal mol−1 985. Calculation here is direct. proposed by Fredenslund. Eng. Res. no azeotrope exists.40)(1. solvent activities in polymer solutions [Oishi and Prausnitz. α12 = 1. DEW T calculation: Find x1 and T. Dev. Eng. and given detailed treatment by Fredenslund. the activity coefficients cannot be immediately calculated. For x1 = 0.15 K (52!C). The range of applicability of the original UNIFAC model has been greatly extended and its reliability enhanced. But if one value is < 1 and the other > 1. Modified Raoult’s law [Eq. Again. Subsequent development has led to a variety of applications. Rasmussen. (A). and they have their origin in the UNIFAC method (UNIQUAC Functional-group Activity Coefficients). vapor pressures of pure species [Jensen. 2 (A) i /kPa = Ai −  TK + Ci γ 1Psat 1 α12 =  γ 2P sat 2 Because α12 is a monotonic function of x1. two must be fixed to allow calculation of the other two. (4-307) provides an equation for calculation of α12 from the thermodynamic functions: (4-309) i Example 3: Dew and Bubble Point Calculations As indicated by Example 2a.40 and T = 325. The UNIFAC model has also been combined with the predictive Soave-Redlich-Kwong (PSRK) equation of state. For y1 = 0. Fundam. Eq. 17: 269–273 (1978)]. given the other two. Res.833 Activity coefficients are given by Eq.40)(1. Chem. (4-142). Activity coefficients are functions of temperature and liquid-phase composition and are correlated through equations for the excess Gibbs energy. results are listed in the accompanying table. Fredenslund.8053)(87. Eqs. Eng. DECHEMA.223 at x1 = 1. Azeotrope calculations: As noted in Example 1a. Both papers contain extensive literature citations.2869)(58. Chem. When an appropriate correlating equation for GE is not available. 21: 1086–1099 (1975)]. An iteration scheme or a solve routine with starting values for the unknowns is required. given P. and Prausnitz [AIChE J.EQUILIBRIUM This equation is often adequate when applied to systems at low to moderate pressures and is therefore widely used. Eng. (B). Bubble point and dew point calculations are only a bit more complex than the same calculations with Raoult’s law. Gmehling. a. x1. only a single degree of freedom exists for this special case. 42: 183–188 (2003)].289 at x1 = 0 and 0. including liquid/liquid equilibria [Magnussen. This is the “solution of groups” approach.8193 Bi 2756. and Rasmussen. The necessary vapor pressures and activity coefficients are supplied by data correlations. have come into use. Because Σ iyi = 1. Given T. and (C) yield the values listed in the table on the following page. Starting values result from setting each γi = 1. The procedure is most completely described (with background literature citations) by Horstmann et al. Res. a binary system in vapor/liquid equilibrium has 2 degrees of freedom. (4-279) i≠j with parameters [Gmehling et al. Thus of the four phase rule variables T. Fundam. with final values for vapor pressures and activity coefficients given immediately by Eqs. its application is largely restricted to 0 to 150°C. part 3. gas solubilities [Sander. the limiting values of α12 are 8. wherein are cited earlier pertinent papers. [Fluid Phase Equilibria 227: 157–164 (2005)]. the Antoine equation: Bi ln P sat i = 1. and Rasmussen. vol. and excess properties. given y1 and T. Jones. 1.04 Ci −45.3145 13.134 b. Chem. the calculation is direct. based on temperaturedependent parameters. The most sensitive quantity for identifying the azeotropic state is the relative volatility. BUBL T calculation: Find y1 and T. (4-274). and y1. d. Process Des.60 and P = 101. c. Elsevier. For y1 = 0. Eng. An extended UNIFAC model called KT-UNIFAC is described in detail by Kang et al. Chem. Two modified versions of the UNIFAC model. but also they allow correlation of various kinds of property data. 11: 105–126 (1983)]. Similarly. suitable estimates of activity coefficients may often be obtained from a group contribution correlation. an iteration scheme or a solve routine with starting values for the unknowns is required. Process Des. (4-309) and (4-307) is readily devised. and Rasmussen [Vapor-Liquid Equilibrium Using UNIFAC.090 −48. Because it is based on temperature-independent parameters. and Fredenslund. the azeotropic composition and pressure is found by seeking the value of P that makes x1 = y1 or that makes α12 = 1.15 K (52!C). Results for x1 = 0. 20: 239–246 (1981)]. Eng. Lohmann. Chem.05 V2 cm3 mol−1 131. Chem. one finds the azeotropic composition and temperature. 17: 333–339 (1978)].

that yield pressures by Eq. Adding the two equations gives Moreover. Chem. (4-252) defines an experimental value of g*: g* x1 ln γ *1 + x2 ln γ*2 Subtraction of Eq.6000 0. the Gibbs-Duhem equation.3828 1.330 89. Such a correlation is unnecessarily divergent.105 58. and the preceding equation becomes γ d ln γ * d ln γ * δ ln 1 = x1 1 + x2 2 γ2 dx1 dx1 (4-317) The right side of this equation is the quantity required by Eq.5851 0. etc. . depend for validity on parameters evaluated from a large base of such data. This is usually accomplished by finding the parameters that minimize the sum of squares of the residuals. Once these parameters are found. Pure & Appl. these residuals fail to scatter about zero and the correlation for g does not properly reproduce the experimental values P* and y*1. Given here is a brief description of the treatment of data taken for binary systems under isothermal conditions.8053 3. The extent to which values of this residual fail to scatter about zero measures the departure of the data from consistency with respect to this equation. as represented by Eq. Equation (4-303) is solved for yi P and written for species 1 and for species 2. β. The usual procedure is to minimize the sum of squares of the residuals δP. etc. The data reduction procedure just described provides parameters in the correlating equation for g that make the δg residuals scatter about zero.4550 0. etc. These equations allow calculation of the primary residuals: δP P − P* and δy1 y1 − y*1 If the experimental values P* and y*1 are closely reproduced by the correlating equation for g. one merely searches for values of the parameters α. The process of finding a suitable analytic relation for g ( GERT) as a function of its independent variables T and x1.616 65. Known as Barker’s method [Austral. T. rather than a valid general relationship.6454 P/kPa 108. (4-312): dg* d ln γ * d ln γ *  = x1 1 + ln γ *1 + x2 2 − ln γ *2 dx1 dx1 dx1 or dg* γ* d ln γ * d ln γ *  = ln 1 + x1 1 + x2 2 dx1 γ *2 dx1 dx1 (4-316) If a data set is reduced so as to yield parameters—α.3200 0.) (4-310) where α.58 THERMODYNAMICS P1sat/ kPa P2sat/ kPa 87. d. The residual on the left is therefore a direct measure of deviations from the Gibbs-Duhem equation.2669 1. it provides the best possible fit of the experimental pressures. b. (4-315) gives γ *1 γ1 dg* dg d ln γ * d ln γ *  −  = ln  − ln  − x1 1 + x2 2 γ *2 γ2 dx1 dx1 dx1 dx1  The differences between like terms represent residuals between derived and experimental values.98 319.779 46. . this is possible because the full P-x1-y1 data set includes redundant information. and the correlating parameters α.105 43.2869 1. written as yi*P* γ i* =  Φi xiPsat i i = 1. many P*-x1 data subsets appear in the literature. thus producing a correlation of VLE data. β. (4-313).0237 1. Presumed in all that follows is the existence of an equation inherently capable of correlating values of GE for the liquid phase as a function of x1: g GE/RT = G(x1. represent adjustable parameters. then the derivative on the left is effectively zero. (4-180) with M g: x1γ1Psat x2γ2Psat 1 2 P=  +  Φ1 Φ2 (4-318) x1γ1Psat 1 y1 =  Φ1P (4-319) whence by Eq. are called derived values. the measurement of y*1 values is not usually worth the extra effort.—that make the δ g residuals scatter about zero.830 81. the Gibbs-Duhem equation. β. is known as data reduction. With experimental error usually concentrated in the y*1 values.125 70.6445 = 0. 27: 113–134 (1995). Data Reduction Correlations for GE and the activity coefficients are based on VLE data taken at low to moderate pressures.134 87. and P. (4-263). Eq. evaluated at the experimental values of x1.000 101.. β. the dependence is so weak as to be universally and properly neglected.6445 0.4000 0. 2 (4-311) where Φi is given by Eq.986 58. Groupcontribution methods. (4-310). may not be consistent with. 67: 859–872 (1995)].1130 0.2700 1.1861 1. e.6473 1. β.15 322. Thus Eq. they can be used for the calculation of derived values of both the pressure P and the vapor composition y1. the calculated y1 values are likely to be more nearly correct. y1. (4-313) applied to sets of experimental values becomes a test of the thermodynamic consistency of the data. sets of values of ln γ *1 and ln γ *2 may not satisfy.616 87. (4-303).24 322. to be zero for consistent data. (4-318) that are as close as possible to the measured values.634 79.9111 x1 y1 0.707 101. A simple summability relation analogous to Eq.9172 1.15 325.. (4-305) and the asterisks denote experimental values. (4-303).1286 1. This is the result obtained when the data approach thermodynamic consistency. Hence. may be written for experimental values in a binary system at constant T and P as dg ln γ1 = g + x2  dx1 γ γ γ* δ ln 1 ln 1 − ln 1 γ2 γ2 γ *2 and puts this equation into the form (4-312) d ln γ* d ln γ* (4-313) x1 1 + x2 2 = 0 dx1 dx1 Because experimental measurements are subject to systematic error. such as UNIFAC. Values of g provided by the equation used to correlate the data. i.5535 2. and produce derived values of the activity coefficients by Eqs. etc.591 53. the Gibbs-Duhem equation. c. f. Chem. Chem.e. Thermodyn.. Defining these residuals as δg g − g* γ1 dδg d ln γ * d ln γ *  = δ ln  − x1 1 + x2 2 γ2 dx1 dx1 dx1  (4-314a) dg ln γ2 = g − x1  dx1 (4-314b) These two equations combine to yield γ1 dg  = ln  γ2 dx1 (4-315) This equation is valid for derived property values. calculated results are boldface. they are of course not subject to a test for consistency by the Gibbs-Duhem equation. Although g is in principle also a function of P. Assuming that the correlating equation is appropriate to the data. are usually best determined without them.790 53. Because Barker’s method requires only the P*-x1 data subset. When experimental y*1 values are not consistent with the P*-x1 data. .4000 0.330 Given values are italic.. The measured variables of binary VLE are x1. J.4-30 T/K a. Barker’s method cannot lead to calculated y1 values that closely match the experimental y*1 values. An alternative is to base data reduction on just the P-x1 data subset. A more comprehensive development is given by Van Ness [J. The corresponding experimental values are given by differentiation of Eq. 6: 207−210 (1953)]. (4-316) from Eq. When they do not. α. scatter about zero.021 γ1 γ2 1.15 317.5605 0. 325.6454 = 0. then these residuals. Experimental values of the activity coefficient of species i in the liquid are related to these variables by Eq.939 80. .

given by a Poynting factor. It does characterize “lightness” of a constituent species. for which an alternative approach is required. (4-319) is Henry’s law x2γ2Psat 2 y2 =  Φ2P f^1 x1(γ1γ1∞)k1 x2γ2P sat 2 Because y1 + y2 = 1. the effect of P on a liquid-phase fugacity. whereas for dew point calculations. and {yi}. representing Henry’s law. T flash. (4-326) reduces to Ki = PisatP. these are alternative expressions of Raoult’s law and modified Raoult’s law. tends to concentrate in the vapor phase whereas a “heavy” species. The slope of the tangent line is Henry’s constant. parts 1–8. tends to concentrate in the liquid phase. and overall composition. Solution for y1 gives x1(γ1γ 1∞)k1 y1 =   φˆ 1P (4-324) The same correlation that provides for the evaluation of γ1 also allows evaluation of γ 1∞. For certain binary systems species 1.] FIG. with K < 1. prepared by DePriester [Chem. P. (4-328). Frankfurt am Main (1979–1990)]. defined as (4-320) This is the definition of k1 for temperature T and for a pressure equal to the vapor pressure of the pure solvent P2sat. but their essential basis is Raoult’s law. Because both k1 and γ 1∞ are evaluated at P2sat. and its fugacity as a pure liquid at the system temperature f1 cannot be calculated by Eq. and Abbott. 7th ed. However. Eng. For modified Raoult’s law. Chemistry Data Series. Progr. The activity coefficient of the solute then becomes fˆ1 y1Pφˆ 1 y1Pφˆ 1γ 1∞ γ1  =  =  x1 f1 x1 f1 x1k1 For the solute. the problem is to find the set of K values that satisfies Eq. the liquid-phase fugacity of pure species 1. K Values. species 2. (4-302). or k1 f1 =  γ 1∞ (4-321) where γ 1∞ represents the infinite dilution value of the activity coefficient of the solute. 1.EQUILIBRIUM 4-31 For the solvent. p. A “light” species. However. designated the solvent. The rigorous evaluation of a K value follows from Eq. (4-327). 7. is either unstable at the system temperature or is supercritical (T > Tc). [Vapor-Liquid Equilibrium Data Collection. designated the solute. (4-303) and (4-304). The nomographs for K values of light hydrocarbons as functions of T and P. The defining equation for K can be rearranged as yi = Ki xi. DECHEMA. therefore y1 ˆ∞ k1 = Psat 2 φ1 lim  x1 → 0 x1 The limiting value of y1/x1 can be found by plotting y1/x1 versus x1 and extrapolating to zero. Figure 4-7 shows a typical plot of the liquidphase fugacity of the solute fˆ1 versus its mole fraction x1 at constant temperature. Considered here is the P. and Flash Calculations A measure of the distribution of a chemical species between liquid and vapor phases is the K value. VLE. This problem is determinate on the basis of Duhem’s theorem: For any closed system formed initally from . Since the curve representing fˆ1 does not extend all the way to x1 = 1. this becomes γ 1∞ = k1 f1. 4-7 The world’s store of VLE data has been compiled by Gmehling et al. in which are calculated the quantities and compositions of the vapor and liquid phases in equilibrium at known T. (4-299): y γi fi Ki i =  xi φˆ iP (4-326) When Raoult’s law applies. New York (2005). Were Raoult’s valid. is not established. [Smith. defined as the equilibrium ratio: y Ki i xi (4-325) It has no thermodynamic content. The flash calculation is a very common application of VLE. with K > 1. vol. With Ki = yi xi. Eq. the problem is to find the set of K values that satisfies Eq. Van Ness. the location of f1. McGraw-Hill. Ser.. but not to the solute. provides alternative information. do allow for an average effect of composition. The tangent line at the origin. 555. the sum Σi xi = 1 yields yi =1 i  K (4-328) i Thus for bubble point calculations. is very small and for practical purposes may usually be neglected. No. For equilibrium fˆ1 fˆ1l = fˆ1v = y1Pφˆ 1 x1 fˆ k1 lim 1 x1→0 x1 (4-323) (4-322) Division by x1 gives fˆ1 y1  = Pφˆ 1  x1 x1 Henry’s constant is defined as the limit as x1 →0 of the ratio on the left. (4-320). Ki = γiPisatP. this pressure also applies to f1. P =  +  φˆ 1 Φ2 f^1 0 1 Plot of solute fugacity fˆ1 versus solute mole fraction. {xi}. where the yi are known. 49: 1–43 (1953)]. Those correlations that do exist are approximate and severely limited in application. The activity coefficient of the solute at infinite dilution is fˆ1 1 fˆ lim γ1 = lim  =  lim 1 → x1→0 x1 f1 x 0 f1 1 x1 x1→0 In view of Eq. but may make for computational convenience through elimination of one set of mole fractions in favor of the other. P. Introduction to Chemical Engineering Thermodynamics. this equation takes the place of Eqs. Equations (4-303) and (4-304) are applicable to species 2. Eq. There remains the problem of finding Henry’s constant from the available VLE data. making convenient and accurate correlation impossible. The sum Σiyi = 1 then yields i K x = 1 i i (4-327) With the alternative rearrangement xi = yi/Ki. (4-326) shows that they are in general functions of T. Solute/Solvent Systems The gamma/phi approach to VLE calculations presumes knowledge of the vapor pressure of each species at the temperature of interest. Symp. where the xi are known. K values could be correlated as functions of just T and P. the analog of Eq. Its vapor pressure cannot be measured.

6306 Equation-of-State Approach Although the gamma/phi approach to VLE is in principle generally applicable to systems comprised of subcritical species. 557. Note that as P → ∞. and it is at best approximate. which may be written as ln φvi = ln φli (4-332) For given T. Vapor pressures are taken from Example 3a or 3b. . . . (4-246). The pure numbers ε. Eq. and for y1 = z1. both special cases of Eq.134 kPa.5190 V = 0.1873 K2 = 0. is shown in Fig.5775 with γ1 = 2. For 1 mol of a system with overall composition represented by the set of mole fractions {zi}. (4-332). The values of K1. The BUBL P and DEW P calculations at T = 325. xiφˆ il = yiφˆ vi i = 1. {xi}. Because P here lies between these values.33 kPa. (4-245): p p p ln φip = Zi − 1 − ln(Zi − βi) − qi Ii p = l. 4-8. (4-331) reduces to φvi = φli. are the Soave-Redlich-Kwong (SRK) equation and the Peng-Robinson (PR) equation. (B) and (C) of Example 3. The gently sloping segment on the right (tu) is characteristic of vapors. Eq. (4-112): 1 Z pi + σβi Iip =  ln  σ− % Zip + %βi p = l. For pure species i. N u 0 V (4-329) Substitute xi = yi /Ki and solve for yi: ziKi yi =  1 + V(Ki − 1) i = 1. A subcritical PV isotherm.15 K and P = 101.939 kPa. If not. and Abbott. then successive trials easily lead to this value.4000 at T = 325. . . and y1. summed over all species. . the chosen value of x1 is correct. but has no physical meaning. T flash problem is to find the value of V which satisfies this equation. but neither is explicit. may be applied to each phase: Liquid: fˆi l = φˆ li xiP Vapor: fˆiv = φˆ vi yi P By Eq.] FIG. (4-204). such as connects points M and W. not only of T and P. v (4-333) Expressions for Z iv and Zil come from Eqs. .0997 K1 = 2. The defining equation for the fugacity coefficient. provides a transition from liquid to vapor. In the unlikely event that the sum is indeed unity. where b is a constant in the cubic equation. Note that the trivial solution giving V = 1 must be avoided. . . (4-331) represents N complex relationships connecting T.V → b. Van Ness.2. Thus. (4-101) and (4-102)]. (4-332) are given by Eq. K2. Moreover.4-32 THERMODYNAMICS given masses of prescribed chemical species. yields ziKi =1 i  1 + V (K − 1) (4-330) A subcritical isotherm on a PV diagram for a pure fluid. Determine V. . here P → 0 as V → ∞. The middle segment (st). P. and systems are formed from given masses of nonreacting chemical species. generated by a cubic equation of state. New York (2005). A generally applicable alternative to the gamma/phi approach results when both the liquid and vapor phases are described by the same equation of state. These points then represent saturated liquid and vapor phases in equilibrium at temperature T. and a flash calculation is appropriate. Pbubl = 108. it is most satisfactory for correlation of constant-temperature data. (4-100) [with Eqs. Pdew = 87. and V that result are substituted into the summation equation. . The answers are x1 = 0. v (4-336) . Example 4: Flash Calculation The system of Example 3 has the overall composition z1 = 0. N Combining these equations to eliminate L gives zi = xi(1 − V ) + yiV P i = 1. The independent variables are here T and P. 2. N s Because Σiyi = 1.. let L represent the molar fraction of the system that is liquid (mole fractions {xi}) and let V represent the molar fraction that is vapor (mole fractions {yi}). . a trial value of x1 then allows calculation of γ1 and γ2 by Eqs. . A temperature dependence for the parameters in expressions for GE is included only for the local composition equations. but values are also implicit in cubic equations of state. The modified Raoult’s law K values are given by (γ1)(Psat 1 ) K1 =  P (γ2)(Psat 2 ) and K2 =  P Equation (4-329) may be solved for V : z1 − x1 V=  y1 − x1 Equation (4-330) here becomes (z2)(K2) (z1)(K1)  +  = 1 1 + V(K1 − 1) 1 + V(K2 − 1) A trial calculation illustrates the nature of the solution. Introduction to Chemical Engineering Thermodynamics. (4-104): Zvi − βi Ziv = 1 + βi − qiβi  v (Zi + %βi)(Ziv + σβi) 1 + βi − Zli Zli = βi + (Z li + %βi)(Zli + σβi)  qiβi  (4-334) (4-335) and Iip comes from Eq. Two widely used cubic equations of state appropriate for VLE calculations. 4-8 i The initial step in solving a P.15 K of Example 3a and 3b show that for x1 = z1. 7th ed. More elegant solution procedures can of course be employed. The material balance equations are L +V=1 and zi = xiL + yiV r t P W M i = 1. The fugacity coefficients in Eq. . . x1. Eq. and {yi}. Ψ. line MW lies at the vapor pressure Pisat if and only if the fugacity coefficients for points M and W satisfy Eq. The actual transition occurs along a horizontal line. The very steep one on the left (rs) is characteristic of liquids. 2. The simplest application of equations of state in vapor/liquid equilibrium is to the calculation of vapor pressures Pisat of pure liquids. The present treatment is applicable to both. the system is in two-phase equilibrium. in practice it has found use primarily where pressures are no more than a few bars. 2. N (4-331) This introduces compositions xi and yi into the equilibrium equations. Three segments are evident. (4-298).5297 γ2 = 1. [Smith. this equation. the equilibrium state is completely determined when any two independent variables are fixed. with both a minimum (note P < 0) and a maximum. The associated expression for φˆ i is given by Eq. Vapor pressures can of course be measured. p. σ. and Ω and expressions for α(Tri) specific to these equations are listed in Table 4-2. Note that V = 1 is always a trivial solution. but of composition. because the φˆ i are functions. McGraw-Hill.2373 y1 = 0. .

(4-115). With φˆ li a function of T. with a starting value for qi from a generalized correlation as given by Eqs. As defined by Eq. (4-102). and (4-106). q i i iRT. New York (2005). are functions of composition. both with the same composition.nj ⎯ bi a⎯i =q 1+  −  a b  (4-345) Any two of the three partial parameters form an independent pair. are functions of T only and are written here as α(Tr )R2Tc2 ai(T) = ψ  Pc (4-338) RT bi = Ω c Pc (4-339) ai(T) qi  biRT (4-340) i i i i i The eight equations (4-332) through (4-337) may be solved for the eight unknowns Pisat. (4-105). This equilibrium condition is shown by Fig. (4-101). Because they lie at the same P.nj T. and ln φiv.nj These are general equations. B 0 4-33 ∂(nq) ⎯ qi  ∂ni   T. For the parameters of the generic cubic. BUBL T. 560. as shown by Eqs. p. they do not represent phases in equilibrium and do not lie at the same pressure. Iiv. (4-332) and (4-333) may be combined and solved for qi. Iil. (4-331) represents N relations among the 2N variables: T. Liquid and vapor mixtures in equilibrium in general have different compositions. Eq. given by Eqs. 7th ed. and DEW T calculations. Because Zil and Ziv depend on qi. Shown in Fig. point D represents a dew point with the vapor-phase composition. P. Here. βi. The solid line is for a liquid-phase composition. Iiv. This pressure is the phase equilibrium pressure. P. When these points lie at the same P (as shown). allows solution for the remaining N variables by BUBL P.or vaporphase composition. and φˆ Vi a function of T. . Each left segment contains a bubble point (saturated liquid). 4-9. For mixtures the presumption is that the equation of state has exactly the same form as when written for pure species. provide for the evaluation of mixture parameters solely from parameters for the pure constituent species. Parameter q is defined in relation to parameters a and b by Eq. usually either T or P and either the liquid. and (4-340). represented by Eqs. Van Ness. (4-246) are implicit in an equation of state and with Eq. yielding Zvi − Zli + ln [(Zli − βi)/(Ziv − βi)] qi =  Iiv − Iil (4-341) Expressions for Zil.EQUILIBRIUM The equation-of-state parameters are independent of phase. and {yi}. Thus. A common combining rule is given by Eq. they represent phases in equilibrium. Although more complex. specification of N of these variables. and each right segment contains a dew point (saturated vapor). empirical mixing rules are used to relate mixture parameters to pure-species parameters. and {xi}. These equations. Eqs. Iil.. DEW P. and Abbott. McGraw-Hill. Equations (4-104) are therefore applicable. They find application primarily for mixtures comprised of simple and chemically similar molecules. (4-331) allow calculation of mixture VLE. and the composition for the dashed line is that of the equilibrium vapor. and the lines are characterized by the liquid and vapor compositions. l v Values of φˆ i and φˆ i as given by Eq. P. They are displaced from each other because the equationof-state parameters are different for the two compositions. ln φil. (4-339). Point B represents a bubble point with the liquid-phase composition. The PV isotherms generated by an equation of state for these different compositions are represented in Fig. and any one of them can be found from ⎯ the other two. Ziv. they represent phases in equilibrium. (4-113) and (4-114). (4-106). (N −1) xi s. Ziv.] FIG. For a BUBL P calculation. (4-334) through (4-337). The problem then is to locate a second (dashed) line for a vapor composition such that the line contains a dew point D on its vapor segment that lies at the pressure of the bubble point B on the liquid segment of the solid line. a. and (4-106). 4-8: the leftmost segment representing a liquid phase and the rightmost segment. and therefore b and a(T). Perhaps more useful is the reverse calculation whereby an equationof-state parameter is evaluated from a known vapor pressure. 4-9 (4-342) T. a vapor phase. (4-104). and this fixes the location of the PV isotherm for the composition of the liquid phase (solid line). 4-9 is a bubble point B on the solid line and a dew point D on the dashed line. the temperature and the liquid composition are known. Zil. with parameters β and q given by Eqs. known as van der Waals prescriptions. the dashed line is for a vapor-phase composition. Thus. Because q. the same basic principle applies as for pure-species VLE. (4-105) and (4-106). valid regardless of the particular mixing or combining rules adopted for the composition dependence of mixture parameters. βi is a function of P and here becomes biPsat i βi  RT (4-337) The remaining equation-of-state parameters. [Smith. Each line includes three segments as described for the isotherm of Fig. (4-338). and βi are given by Eqs. In the absence of a theory to prescribe the composition dependence of parameters for cubic equations of state. 4-9 by two similar lines: the solid line for the liquid-phase composition and the dashed line for the vapor-phase composition. (4-105). these parameters. and (N−1) yi s. The simplest realistic expressions are a linear mixing rule for parameter b and a quadratic mixing rule for parameter a. the definitions are ∂(na) a⎯i  ∂ni  ∂(nb) ⎯ bi  ∂ni  (4-343) ∂(nq) ⎯ qi  ∂ni  (4-344)    P T. Thus. Useful in the application of cubic equations of state to mixtures are partial equation-of-state parameters. The general mole fraction variable xi is used here because application is to both liquid and vapor mixtures. an iterative procedure is indicated. Introduction to Chemical Engineering Thermodynamics. Because these points for a given line are for the same composition.nj n(na) nq =  RT(nb) D whence V Two PV isotherms at the same T for mixtures. and b ⎯ ≠ a⎯ b are not linearly related.

The idea here is to exploit the connection between fugacity coefficients and activity coefficients provided by their definitions: fˆi/xiP φˆ fˆi γi  =  = i fi/P xifi φi ln γi = ln φˆ i − ln φi Therefore.15 K (50!C). New York (2005)]. Data Mixtures. For application of Eq. it is divided by the product x1x2. and the closer T0 and T1 are to T. Extrapolation of Data with Temperature Liquid-phase excessproperty data for binary systems at near-ambient temperatures appear in the literature. 7th ed. The extensive literature on this subject is reviewed by Valderrama [Ind. GE GE  =  x1x2RT x1x2RT  HE  − 1  −  x x RT  T T T0 T T1 1 2 T1 0 CEP T T T −  ln  −  − 1 1 x1x2R T0 T0 T    The assumption here is that C is independent of T. J. xi = 0 T1 0 −   − 1   ln  T T T xi = 0 T T 0 0 T1 (4-352) Example 5: VLE at Several Temperatures For the methanol(1)/ acetone(2) system at a base temperature of T0 = 323. parameters A12 and A21 relate directly to infinite dilution values of the activity coefficients. Ser. (4-113) through (4-115). since 1990. Sim. (4-270).. Progress 98:(11): 58–65 (Nov. 437. Integrate from T2 to T:  C dT T E P  (4-350) T1 ∂CPE dCEP =  ∂T   1 ∂C     RT ∂T T T 2 T0 T E P T1 T2 (4-351) dT dT dT P.. McGraw-Hill. 7th ed.6557 These values allow calculation of equilibrium pressures through Eqs. (4-245) and Eq. 569–572. Thus. When no information is available for CPE and excess enthalpy data are available at only a single temperature. (4-351) are retained. (4-351) to binary systems at infinite dilution of one of the constituent species. Excess heat capacity data are uncommon.08 kPa. CPE must be assumed zero. A. Van Ness and Abbott [Introduction to Chemical Engineering Thermodynamics. the less the influence of this assumption. 28: 13–37 (1986)]. x HE = H1E + T1 constant P. McGraw-Hill. Experimental VLE data at 372. and the T dependence is rarely known. In this case only the first two terms on the right side of Eq. we have from the VLE data at 323. and excess heat capacity data at T2.15 K: dHE = CPE dT constant P. 42: 1603–1618 (2003)] and by Twu. and it more rapidly becomes imprecise as T increases. The VLE data are well correlated by the Margules equations.0362 .6281 and A21 = ln γ ∞2 = 0. Integral I depends on the temperature dependence of CPE. x and the excess-property analog of Eq.. this equation is written for the liquid phase. The required excess enthalpy values at T0 are HE  x x RT 1 2 T0. the equationof-state approach has found primary application to systems exhibiting modest deviations from ideal solution behavior in the liquid phase. 2002)]. e.. both VLE data [Van Ness and Abbott. As noted in connection with Eq.. They provide for the extrapolation of GE correlations with temperature. New York (2005)]. to systems containing hydrocarbons and cryogenic fluids. the summability equation provides an Because q i exact mixing rule: q = xi ⎯ qi Combining this equation with Eqs. (4-270) and (4-308) for comparison with the measured pressures of the data set. Substituting for ln φˆ i and ln φi by Eqs. extensive research has been devoted to developing mixing rules that incorporate the excess Gibbs energy or activity coefficient data available for many systems. p. Values of Pisat required in Eq. Eng. Int. written as HE −  RT − 1     T T (4-347) ⎯ is a partial property. Sel. Assuming CPE independent of T makes the integral zero. DATA Ser. such as those given by Eqs. Eng. Van Ness. As shown by Smith. (4-346) Because γi is a liquid-phase property. x1=0 = 1. However.g. Solution for ⎯ qi yields Z−β 1 b ⎯ =  q 1 − Zi + i (Z − 1) − ln  + qiIi − ln γi i Zi − βi I b   GE GE  =  RT RT (4-348) T0 E  I where Integration of the first equation from T0 to T gives H dT  −  RT T E T0 (4-349) 2 T0 Integration of the second equation from T1 to T yields In addition. thus confirming the suitability of the Margules equation for this system. Chem. Eng.3636 and HE  x x RT 1 2 T0. (4-352) with CPE = 0. (4-349) and (4-350) leads to dT P. (4-308) are the measured values reported with the data set. The key relations are Eq. (4-246) gives Z−β b ⎯I + qI ln γi = i (Z − 1) − Zi + 1 − ln  − q i i i Zi − βi b Symbols without subscripts are mixture properties. and Tassone [Chem. (4-250).7 K are given by Wilsak et al. x This general equation employs excess Gibbs energy data at temperature T0. E P GE  x x RT xi = 0 1 2 ln γ ∞i The preceding equation may therefore be written as HE ln γ ∞i = (ln γ ∞i )T −  x1x2RT 0  CEP −  x1x2R  − 1   T T T T1. hence no value can be given for rms δy1.8 and 397. making I = 0. [Fluid Phase Equilib. (4-26): GE GE  =  RT RT T1 0  2 E G H d  = − 2 dT RT RT T CPE T T T −  ln  −  − 1 1 − I R T0 T0 T i Application of this equation in the solution of VLE problems is illustrated by Smith.4-34 THERMODYNAMICS Because of limitations inherent in empirical mixing and combining rules. and Abbott [Introduction to Chemical Engineering Thermodynamics. Data 20: 403–405 (1975)] are available. x2 = 0 = 1. Res. Chem. pp. Vapor-phase mole fractions were not reported. excess enthalpy (heat-of-mixing) data at T1. 1978: 67 (1978)] and excess enthalpy data [Morris et al. Values of ln γ ∞i and hence of the Margules parameters for these higher temperatures are found from Eq. The root-mean-square (rms) value of the pressure differences is given in Table 4-7 as 0. x dT A12 = ln γ ∞1 = 0. x CEP = CPE + 2 ∂C   ∂T T T2 E P P.

j ∆ε j j For a phase in which a chemical reaction occurs according to the equation ν1A1 + ν2A2 + · · · → ν3A3 + ν4A4 + · · · i = 1.j ε j i = 1. an expression for GE capable of representing liquid/liquid phase splitting is required. because some cannot represent liquid/liquid phase splitting. . for a system containing N chemical species.22 0. . Results for the higher temperatures indicate the quality of predictions based only on vapor-pressure data for the pure species and on mixture data at 323. N j = I. Chemistry Data Series. . NRTL. a vapor-phase equation of state for computing the φˆ i is also needed. .j = 0. .32 RMS δy1 0. The reaction coordinate has units of moles.08 0.014 (0. thereby giving LIQUID/LIQUID AND VAPOR/LIQUID/ LIQUID EQUILIBRIA fˆαi = fˆβi where j = I. Thus. meaning that reaction j has proceeded to such an extent that the change in mole number of each reactant and product is equal to its stoichiometric number. . 0 } (4-358) All these terms are equal. . N 0 (4-361) j Equation (4-361) is the basic expression of material balance for a closed system in which r chemical reactions occur. (4-295) becomes i = 1. . For vapor/liquid/liquid equilibria. called the reaction coordinate for reaction j. With activity coefficients applied to the liquid phases and fugacity coefficients to the vapor phase. (4-296). . Extrapolations based on the same data to still higher temperatures can be expected to become progressively less accurate. solution of these equations requires a single expression for the composition dependence of GE suitable for both liquid phases. the effect of pressure on the γi can be ignored. as for VLE.3764) 0. The νi themselves are called stoichiometric numbers. It is not an equilibrium relation. for reaction j ∆n1. 2 . (4-354) then constitutes a set of N equations relating equilibrium compositions to one another and to temperature. A special table of parameters for LLE calculations is given by Magnussen et al. . 2. . Chem.6557 0. N (4-354) For most LLE applications.j ∆nN. the parameters are determined for pairs of species exactly as for a binary system. Frankfurt am Main (1979–1980)].5271) 0. Dev. [Fluid Phase Equilib. It shows for a reacting system that at most r mole-number-related quantities ε j are capable of independent variation. valid for tracking the progress of the reactions to arbitrary levels of conversion.15 K. N (4-355) fˆαi = fˆβi = fˆvi where α and β designate the two liquid phases. A comprehensive treatment of LLE is given by Sorensen et al. For a given temperature. r (4-363) Standard Property Changes of Reaction For the reaction aA + bB → lL + mM a standard property change is defined as the property change resulting when a mol of A and b mol of B in their standard states at temperature .7 VLE Results for Methanol(1)/Acetone(2) A12 = ln γ ∞1 A21 = ln γ ∞2 RMS δP/kPa RMS % δP 0. More generally. For such systems. 3: 47–82 (1979).004 (0. CHEMICAL REACTION EQUILIBRIA The general criterion of chemical reaction equilibria is given by Eq. The UNIQUAC equation is suitable. N (4-353) Eliminating fugacities in favor of activity coefficients gives xαi γ αi = xβi γ βi i = 1.j ∆ni. r all i (4-356) As for LLE.4640) 0. . any or all of which can participate in r chemical reactions. even in principle. . parts 1–3. . II. A change in ε j of 1 mol signifies a mole of reaction. [Ind.39) 0. j = νi. . j) = for a product species If species i does not participate in reaction j. . . and associated with them is a sign convention such that the value is positive for a product and negative for a reactant. . . 20: 331–339 (1981)]. 2.EQUILIBRIUM TABLE 4-7 T.j ∆ε j i = 1.j  =  =…=  ν1. II.4607) 0.5177 (0.2: 297–309 (1979). K 323.8 397. . Process Des. Equation (4-295) is the basis for both liquid/liquid equilibria (LLE) and vapor/liquid/liquid equilibria (VLLE). . (4-361) becomes ni = ni + νiε whence dni = νi dε Substitution for dni in Eq.3725 (0. vol. . Eq. . and therefore prediction is possible by UNIFAC models. Eq. 4: 151–163 (1980)]. and UNIQUAC equations are suited to the treatment of multicomponent systems.j Ai j = I.j νN.j ν2.4615 (0. Not all expressions for GE suffice. 2. For a system in which just a single reaction occurrs. . Eq. .46 (1.j resulting from the various reactions. . . Data for LLE are collected in a three-part set compiled by Sorensen and Arlt [Liquid-Liquid Equilibrium Data Collection.j i CHEMICAL REACTION STOICHIOMETRY (4-359) Because the total change in mole number ∆ni is just the sum of the changes ∆ ni. . . .85 (0. . then vi. . . Eng. 5. 2. then 0 ni = ni + νi. and they can be equated to the change in a single quantity ε j.j ∆n2. (4-295) is written as i = 1. the reactions are represented by the equations 0 = νi. but merely an accounting scheme. The values in parentheses are from the gamma/phi approach as reported in the papers cited.6281 0. the 2N equilibrium equations for subcritical VLLE are xαi γ αi f αi = yiφˆ iP xβi γ βi f βi = yiφˆ iP − for a reactant species + sign (νi.12 0. r (4-357) i Results of calculations with the Margules equations are displayed as the primary entries at each temperature in Table 4-7. .006) 0. Only the Wilson. Thus for LLE with superscripts α and β denoting the two phases. ∆ni = ∆ni. and Eq.4465 (0. . . II.83) 2. (4-296) leads to i ν µ = 0 i i (4-362) Generalization of this result to multiple reactions produces i ν µ =0 i.013) 4-35 the νi are stoichiometric coefficients and the Ai stand for chemical formulas. The stoichiometric numbers provide relations among the changes in mole numbers of chemical species which occur as the result of chemical reaction.j = νi. N (4-360) j If the initial number of moles of species i is ni and if the convention is adopted that εj = 0 for each reaction in this initial state. DECHEMA.15 372. 2.

The standard states are always at the equilibrium temperature. (4-202) for species i in the equilibrium mixture. y φˆ iP aˆi = i  P° and Eq. for example. j However. or from an equation of state. the standard state for gases is the ideal gas state at the standard state pressure.jµ°i (4-365) ∆H°j νi.” Kj is a function of T.. Eq. Then. The activities in Eq. by definition. Therefore. the standard state is the pure ideal gas at pressure P°.j and P° is the standard state pressure of 100 kPa. specification of the pressure allows solution for the ε j. e.j = Kj all j (4-374) the “law of mass action. Except in the critical region. and Eq. Eq. i [ν (G°i + RT ln aˆi)] = 0 i ν G°i + RT ln aˆνi = 0 i. (4-372). (4-218)] each φˆ i becomes φi. the fugacity coefficient of pure species i at the mixture T and P. for a particular species it must be the state represented by both G°i and the f°i upon which activity âi is based. whence i aˆ νi.jG°i = νi. In the latter case. but at pressures P and P°. An important special case of Eq. which depend on the xi’s. fˆ y φˆ iP aˆi i = i  f i° f i° (4-364) i i The application of Eq. for fixed temperature Eqs. then [Eq. (4-52)]. respectively. and the ratio fi fi° is often taken as unity. (4-369) becomes i (γ x ) νi. However. the second. In principle. and ∆G°j is the corresponding standard Gibbs energy change of reaction [see Eq.e.jH°i (4-366) ∆C°P νi. The standard state pressure P° is fixed at 100 kPa. The standard heat of reaction ∆H°j is used in the calculation of the heat effects of chemical reaction.and liquid-phase reaction . upon rearrangement. A standard state of species i is its real or hypothetical state as a pure species at temperature T and at a standard state pressure P°. (4-371) relate the ε j to P.jG°i i ln aˆνi =  RT i i. When this is not acceptable. for which f i° = P°. Numerical values for ∆H°j and ∆G°j are computed from tabulated formation data. and ∆C°P is determined from empirical expressions for the T dependence of the C°P [see. (4-369) requires explicit introduction of composition variables. In this event φˆ i = 1. j M°i For species present as gases in the actual reactive system. and its general mathematical definition is ∆M°j νi. If the equilibrium mixture is assumed an ideal solution. on the ε j. j νj  P° P = Kj all j (4-372) In the general case the evaluation of the φˆ i requires an iterative process. (4-371) results for gas-phase reactions when the phase is assumed an ideal gas. i. and Eq. Convenience suggests setting it equal to ln Kj. Substitution of this equation into Eq.. (4-243) or (4-246).j fˆ f aˆi i = γ i xi i fi° fi°  i − νi. with the process continued to convergence. For liquid-phase reactions. Note that the standard states may represent different physical states for different species.4-36 THERMODYNAMICS T react to form l mol of L and m mol of M in their standard states also at temperature T. Eq. and the standard heat capacity change of reaction is used for extrapolating ∆H°j and ∆G°j with T. any or all the species may be gases. The yi’s may be eliminated in favor of equilibrium values of the reaction coordinates ε j (see Example 6). The most commonly used standard property changes of reaction are ∆G°j νi. but only of T. Eq. This problem has not been solved for the general case. pressure has little effect on the properties of liquids. (4-372) and (4-374). for which data are presumed available. (4-369) provide the connection between the equilibrium states of interest and the standard states of the constituent species. (4-369) becomes i (y φˆ ) i i νi. Equilibrium Constants For practical application. The activity is then i.j νj  P° P = Kj all j (4-371) where νj Σiνi. aˆi = γ i xi.j i i = Kj all j (4-373) Here the difficulty is to determine the γ i’s. allowing evaluation of the φˆ i by. j or or i. Two courses are open: the first is experiment. giving j i µi = G°i + RT ln aˆi (4-368) ˆ where by definition aˆi fi/f°i and is called an activity. An initial step is to set each φˆ i equal to unity and to solve the problem by Eq. This quantity does not depend on composition and may be determined from experimental data. (4-363) must be reformulated. γ i = 1. Although called a “constant. assumption of solution ideality. (4-369) is modified by introduction of the activity coefficient γi = fˆi xifi.jC°P (4-367) i i j i i The standard Gibbs energy change of reaction ∆G°j is used in the calculation of equilibrium compositions. Both fi and fi° represent fugacity of pure liquid i at temperature T. (4-373) reduces to i (x ) i νi. expressed in the same units used for P. Equation (4-199) for species i in its standard state is subtracted from Eq. from a generalized correlation. and i (y ) i νi.g. it is usually the state of pure real liquid or solid at P°. Although the standard state need not be the same for all species. the problem may be complicated by the dependence of the φˆ i on composition. or solids. For liquids and solids. For gas-phase reactions this is accomplished through the fugacity coefficient = Kj all j −∆G°j Kj exp  RT  (4-369) (4-370) Quantity Kj is the chemical reaction equilibrium constant for reaction j. this ratio is evaluated by the equation V (P − P°)  V dP   RT P fi 1 ln  =  f°i RT i i P° When the ratio fi fi° is taken as unity. This provides a set of yi values.” The significant feature of Eqs. the simplest expressions for gas. liquids. where xi is the liquid-phase mole fraction. (4-364) yields.j i where. The initial step is elimination of the µi in favor of fugacities. The standard property change of reaction j is given the symbol ∆M°j. (4-365)]. j The right side of this equation is a function of temperature only for given reactions and given standard states. Equation (4-371) can then be solved for a new set of yi values.

See also Sec. but with parameters A. C6H6 H2 C6H12 Tc Tr Pc Pr ω B0 B1 φ 562.008 0. ∆B. (4-370). Tex. 103: 11–22 (1995). (4-372). Combining this equation with Eqs. A feed stream containing 3 mol H2 for each 1 mol C6H6 is the basis of calculation.368 0. T = 600 K. yielding dT d∆Sj° = ∆C°Pj  T and Integration of these equations from reference temperature T0 (usually 298. and a molar feed ratio H2/C6H6 = 3.EQUILIBRIUM equilibrium. and ∆D. This requires application of Eq. (4-218) indicates that φˆ idi = φi. namely. (4-53) and (4-54).78 40.. for an exothermic reaction. ∆C. written for a single reaction. and for this single reaction. LlanoRestrepo. Prausnitz. supp.02874 1 − ε 3 − 3ε  and ε = 0. and (4-378) yields 1 −∆G°0 ∆H°0 τ − 1 ln K =  +   −  T RT0 RT0 τ 0 (4-377) T ∆CP° dT ∆S° = ∆S°0 + R   T0 R T 1 −∆G° −∆H°0 ln K =  =  −  RT RT T is carried out over a catalyst formulated to repress side reactions. serial publications of the Thermodynamics Research Center. rearranging. 6. The Properties of Gases and Liquids.” J. (4-371). and D replaced by ∆A.084 48. is that the temperature-. [See in particular TRC Thermodynamic Tables—Hydrocarbons and TRC Thermodynamic Tables—Non-hydrocarbons. College Station. Fluid Phase Equilib. 2 (1982). (4-371) for a single reaction becomes i (y φ )   P° i i νi P ν =K P φi = exp(Bi0 + ωBi1) r Tr i i T P cyclohexane (4-380) τ+1  ∆RC° dTT = ∆A lnτ + ∆BT + ∆CT + τ∆DT   (τ − 1) (4-381) 2  T0 hydrogen nC = ε For purposes of illustration we evaluate the pure-species fugacity coefficients by Eq. rather than just at the reference temperature T0 = 298.02874. However. and apply Eq.009 1. New York. and O’Connell. with subscript j omitted and ν = − 3: ε  4 − 3ε 15 −3 P ν νi = 3  = K = 0. 5th ed. Note that Tc and Pc for hydrogen are effective values as calculated by Eqs. but because the reaction is exothermic the equilibrium conversion decreases with increasing T.8 553. chap. in the case of an ideal solution Eq. P = 15 bar. 219: 181–193 (2004). An alternative assumption is that the equilibrium mixture is an ideal solution. pressure-.2972 0. (4-52). (4-124) and (4-125) and used with ω = 0. Chem. A comprehensive study of the effect of operating variables on the chemical equilibrium of this reaction has been published by J.5 percent. and compositiondependent terms are distinct and separate. 2 0 Equations (4-379) through (4-381) together allow an equation to be written for lnK as a function of T for any reaction for which appropriate data are available.919 1. written here as ∆C − 1) +  T 30(τ 3 − 1) 3 benzene nH = 3 − 3ε i n = 4 − 3ε (4-378) where for simplicity subscript j has been supressed. these are reviewed by Poling. the assumption of ideal gases leads to a calculated conversion of 81.. “The NBS Tables of Chemical Thermodynamic Properties.758 0. Assume first that the equilibrium mixture is an ideal gas. Carrero-Mantilla and M. Because the standard state pressure is constant. ∆Hj° is positive and K j increases with increasing T. 2.02874 i yi  1 − ε 3 − 3ε P° 1   4 − 3ε 4 − 3ε  whence    4 − 3ε 3 ε   (15)−3 = 0.004 0. Phys. Fluid Phase Equilib. Eq.139 0. The definition of G leads directly to ∆G° = ∆H° − T ∆S°. (4-377).00 0. (4-364). Texas A & M Univ.2 42.0768 − 0. in which case Eq. methods of estimation are available. (4-206).908 . and defining τ TT0 give finally   ∆CP°  dT + T0 R T  ∆C°P dT   T0 R T T (4-379) When heat capacity equations have the form of Eq. For an estimation procedure based on molecular structure. B. the integrals are evaluated by equations of exactly the form of Eqs.067 14.306 0. C.016 0. Carrero-Mantilla and Llano-Restrepo express ln K as a function of T by an equation which for 600 K yields the value K = 0.21 − 0. it is negative and Kj decreases with increasing T. Presented here are calculations for a single set of operating conditions. Thus for the ideal gas standard state ∆B  ∆RC° dT = ∆A T (τ − 1) +  T (τ 2 T P 2 0 0 2 T0 ∆D τ − 1 +   T0 τ  0 i Each mole fraction is therefore given by yi = ni (4 − 3ε).98 19.73 0. Operating conditions cover a pressure range from 10 to 35 bar and a temperature range from 450 to 670 K.2880 0. yielding nB = 1 − ε ∆C °  dT R T In the more extensive compilations of data. McGraw-Hill. values of ∆G° and ∆H° for formation reactions are given for a wide range of temperatures. For these conditions we determine the fractional conversion of benzene to cyclohexane. Ref.815 Thus. System. see Constantinou and Gani. Example 6: Single-Reaction Equilibrium The hydrogenation of benzene to produce cyclohexane by the reaction In view of Eq. The effect of temperature on the equilibrium constant follows from Eq. (4-370) this may also be written as d∆Hj° = ∆CP°j dT 4-37 2 0 2 The following table shows values for the various quantities in this equation.21 0. (4-365) and (4-366) this equation easily extends to relate standard property changes of reaction: d(∆Gj°RT) − ∆Hj°  =  dT RT 2 (4-375) C6H6 + 3H2 → C6H12 d ln Kj ∆Hj°  =  dT RT 2 (4-376) For an endothermic reaction. (4-28) may be extended to relate standard properties of reaction.15 K. (4-41) written for pure species j in its standard state (wherein the pressure Po is fixed): d(Gj°/RT) − Hj°  =  dT RT 2 With Eqs. Eq. the stoichiometric value. in accord with Eq.15 K) to temperature T gives ∆H° = ∆H°0 + R P T0   ∆RC° dT + ∆RS° +  ∆RC° dTT T T P 0 T0 P T0 Substituting ∆S°0 = (∆G°0 − ∆H°0)/T0. (4-361) becomes ni = ni + νiε. Data 11.6 1.] Where data are lacking. Reaction rate increases with increasing T. 2000.

14. New York (1996. In any event. 1. . (4-293). . Form a function F by addition of this sum to Gt: F = Gt + λk k  n a i ik − Ak i t Function F is identical with G . because of their predominance in chemical technology. Thus an iterative procedure is indicated.. these equations give k λ  i n a k i ik − Ak = 0 3. multiphase systems is treated by Iglesias-Silva et al. then each φˆ i is unity. However. the standard Gibbs energy change of formation of species i. Example 13. and Abbott [Introduction to Chemical Engineering Thermodynamics. Formulate the constraining material-balance equations. Ghajar. 2. However.n  ∂n i j  T.919 4 − 3ε P ν 15 ν = 3  i (yiφi)  3 − 3ε 1−ε P° 1  0. and different assumptions produce different results. and define Ak as the total number of atomic masses of the kth element in the feed. N (4-383) k However. In no case are these calculated conversions significantly divergent. (4-384)].. as there are always products and reactions. . . and Gasem [Ind. . The fugacity coefficients in the equilibrium equation clearly cancel one another. N (4-384) If species i is an element. each φˆ i is a function of the set {yi}. the fugacity is eliminated—in favor of the fugacity coefficient by Eq. The treatment here is limited to gas-phase reactions for which the problem is to find the equilibrium composition for given T and P and for a given initial feed. (4-204). Then a new set {yi} is determined.4-38 THERMODYNAMICS The equilibrium equation now becomes: ε  0. a maximum amount is . each φˆ i becomes φi and can at least be estimated. a minimum amount must be expended to bring about a specific change of state in the flowing fluid. This is not uncommon in reaction equilibrium calculations. 2. 511–513. w − Ak = 0 (4-382) k = 1. (4-382)]. CALCULATION OF IDEAL WORK In any steady-state steady-flow process requiring work. . A more direct procedure (and one suitable for general computer solution) is based on minimization of the total Gibbs energy Gt in accord with Eq. 2. i THERMODYNAMIC ANALYSIS OF PROCESSES Real irreversible processes can be subjected to thermodynamic analysis. (4-384). . w Summed over k. because the summation term is zero. . Chem. Example 13. . In addition. The minimum value of both F and Gt is found when the partial derivatives of F with respect to ni are set equal to zero: ∂F ∂Gt =  ∂ni T. then for compounds G°i = ∆G°f i. Eng. and the λk’s. For low pressures or high temperatures this result is usually adequate. both for the ideal gas assumption and for calculations wherein φˆ i values are determined from the Soave-Redlich-Kwong equation of state.. the equation for µi becomes y φˆ iP µi = ∆G°f + RT ln i  P° i Combination with Eq. initiated with each φˆ i set equal to unity. pp. 2 . 42: 3786–3801 (2003)]. (4-368). The unknowns in these equations are the ni’s (note that yi = ni /Σi ni). In a process producing work. . let aik be the number of atoms of the kth element present in each molecule of chemical species i. w 2. . 5th ed. Multiply each element balance by λk. the question of what chemical reactions are involved never enters directly into any of the equations. 210: 229–245 (2003)] and by Sotyan. . With these substitutions. [Fluid Phase Equilib.nj µi + λkaik = 0 = K = 0. 2005)]. the partial derivatives of F and Gt with respect to ni are different. the chemical potential is given by Eq.P.02874 + λkaik = 0 k i = 1. . based on conservation of the total number of atoms of each element in a system comprised of w elements. pp. this equation becomes fˆi µi = G°i + RT ln  P° If G°i is arbitrarily set equal to zero for all elements in their standard states. one for each element—a total of N + w equations. fˆi = yiφˆ iP.14. . In this procedure. 7th ed. Where it is not satisfactory. of which there are w—a total of N + w unknowns. making the ideal gas assumption far more useful than might be expected. . ∆G°f is zero. If the phase is an ideal solution. The treatment here is limited to steady-state steady-flow processes. therefore. 2001.13. a set of species or an equivalent set of independent reactions must always be assumed. All calculations are well suited to computer solution. 4. Further. Thus the number of equations is sufficient for the determination of all unknowns. . . For real gases. of which there are N. Let subscript k identify a particular atom. Complex Chemical Reaction Equilibria When the composition of an equilibrium mixture is determined by a number of simultaneous reactions. an equation of state with the preliminary set {yi} gives a new and more nearly correct set {φˆ i} for use in Eq. one for each chemical species. Van Ness.908 1. (4-383) gives y φˆ iP ∆G°f + RT ln i  + λkaik = 0 P° k i i = 1. because function F incorporates the constraints of the material balances. General application of the method to multicomponent. pp. 6th ed. Res. Equation (4-384) is derived on the presumption that the set {φˆ i} is known. There are N equilibrium equations [Eqs. 2. . the choice of a set of species is entirely equivalent to the choice of a set of independent reactions among the species. the quantities being calculated.816 Solution yields This result is hardly different from that based on the ideal gas assumption. calculations based on equilibrium constants become complex and tedious. . 527–528. 602–604. The goal is to calculate the efficiency of energy use or production and to show how wasted energy is apportioned among the steps of a process. The material balance for element k is then i n a i ik i n a or i ik = Ak k = 1. for gas-phase reactions and standard states as the pure ideal gases at Po. The process is repeated to convergence. .P. McGraw-Hill. −3  ε = 0. Solution of the equations then provides a preliminary set {yi}. Example 15. A detailed example of a complex gas-phase equilibrium calculation is given by Smith.004 4 − 3ε 4 − 3ε  i   The first term on the right is the definition of the chemical potential. Carrero-Mantilla and Llano-Restrep present results for a wide range of conditions. a Lagrange multiplier:  n a λk i ik − Ak = 0 i k = 1. and there are w material balance equations [Eqs. If the phase is an ideal gas.

(4-394) are identical. The second item means that heat exchange between system and surroundings must occur at the temperature of the surroundings. For irreversible processes the inequality holds. the energy that becomes unavailable for work. The rate form is   fs − Tσ ∆(Sm˙)fs (4-386) In most applications to chemical processes. Writing Eq. When ˙ ideal (or Wideal) is positive. Thus reversible process. the limiting value obtains when the specific change of state is accomplished completely reversibly. Eq. this equation becomes Q˙ ∆(Sm˙)fs −  = 0 (4-385) Tσ The energy balance for a steady-state steady-flow process as given by Eq. W ˙ ideal is the maximum work When W obtainable from a given change in the properties of the flowing ˙ s. Thus every irreversibility carries with it a price. When the same property changes occur in an actual ˙ s (or Ws) is given by an energy balance. This may require Carnot engines or heat pumps internal to the system that provide for the reversible transfer of heat from the temperatures of the flowing fluid to that of the surroundings. In this case a thermodynamic efficiency ηt is defined smaller than W as the ratio of the ideal work to the actual work: ˙ eal W ηt(work required) = id (4-390) ˙s W ˙ ideal (or Wideal) is negative. 2. (4-150):   1 2 ˙ s= ∆ H +  u + zg m˙ W 2 − Q˙ fs Subtracting the ideal work rate as given by Eq. ANALYSIS OF STEADY-STATE STEADY-FLOW PROCESSES Many processes consist of a number of steps. and it is ˙ s. the thermodynamic effistreams. (4-386) yields ˙ lost = Tσ ∆(Sm˙ )fs − Q˙ (4-394) W For the special case of a single stream flowing through the control volume. The implications of this requirement are that 1.e. (4-385) to eliminate Q˙ yields   1 ∆ H +  u2 + zg m˙ 2 fs ˙ s(rev) = Tσ ∆(Sm˙)fs + W LOST WORK Work that is wasted as the result of irreversibilities in a process is ˙ lost. devised solely to find the ideal work associated with a given change of state. and lost-work calculations are then made for each step separately. and process. presumed to constitute a heat reservoir at a constant and uniform temperature Tσ . is positive. Because Carnot engines and heat pumps are cyclic. These conditions are implicit in the entropy balance of Eq. The engineering significance of this result is clear: The greater the irreversibility of a process. (4-397) by Tσ gives Tσ S˙ G = Tσ ∆(Sm˙ )fs − Q˙ Because the right sides of this equation and of Eq. (4-156) and (4-157). this becomes Wlost = Tσ SG (4-400) Because the second law of thermodynamics requires therefore S˙ G ≥ 0 and SG ≥ 0 ˙ lost ≥ 0 W and Wlost ≥ 0 When a process is completely reversible. the greater the rate of entropy generation and the greater the amount of energy that becomes unavailable for work. and it is larger than W ciency is defined as the ratio of the actual work to the ideal work: ˙s W ηt(work produced) =  (4-391) ˙ ideal W Wlost Ws − Wideal (4-392) ˙ lost W ˙s−W ˙ ideal W (4-393) The actual work rate comes from Eq. The process is internally reversible within the control volume. (4-399) for each step of the process and summing give W˙ lost = Tσ S˙ G . (4-150) is   1 ∆ H +  u2 + zg m˙ 2 fs ˙s = Q˙ + W (4-150) Combining this equation with Eq. it follows that ˙ lost = Tσ S˙ G W (4-399) For flow on the basis of a unit amount of fluid. This work is called the ideal work W 1 2 ˙ ideal = ∆ H +  W u + zg m˙ 2 4-39 Q˙ S˙ G = ∆(Sm˙ )fs −  Tσ (4-397) Division by m˙ provides an equation based on a unit amount of fluid flowing through the control volume: Q SG = ∆S −  Tσ (4-398) Equations (4-397) and (4-398) are special cases of Eqs. Thus by definition. (4-387) becomes ˙ ideal = m˙ (∆H − Tσ ∆S) W (4-388) Division by m˙ puts this equation on a unit-mass basis: Wideal = ∆H − Tσ ∆S (4-389) A completely reversible process is hypothetical.. In either case. they undergo no net change of state. the kinetic and potential energy terms are negligible compared with the others. Heat transfer external to the control volume is reversible. (4-156) when S˙G = 0. the actual work W comparison can be made of the actual work with the ideal work. Multiplication of Eq. Its only connection with an actual process is that it brings about the same change of state as the actual process. the equality holds and the lost work is zero. (4-386) is written as ˙ ideal = ∆(Hm˙)fs − Tσ ∆(Sm˙)fs (4-387) W For the special case of a single stream flowing through the system. In this case. i. allowing comparison of the actual work of a process with the work of the hypothetical reversible process.THERMODYNAMIC ANALYSIS OF PROCESSES attainable for a specific change of state in the flowing fluid. Equations (4-386) through (4-389) give the work of a completely reversible process associated with given property changes in the flowing streams. The total rate of entropy generation (in both system and surroundings) as a result of a process is ˙ s(rev) indicates that the shaft work is for a completely where W ˙ ideal. it is the minimum work required to bring W about a given change in the properties of the flowing streams. ˙ lost = m˙ Tσ ∆S − Q˙ (4-395) W Division of this equation by m˙ gives Wlost = Tσ ∆S − Q (4-396) where the basis is now a unit amount of fluid flowing through the control volume. If in addition there is but a single surroundings temperature Tσ. in this event Eq. and the lost work. and it is defined as the difference between the called lost work W actual work of a process and the ideal work for the process.

.82)(0.01 55. Mines Tech.407 82.S.01 1. From the tabulated data.8636) + (52.35)(1)] − (−15.1364) + (12.8636) − (12. equal to 300 K P2/P1 = overall pressure ratio. U.485 12.48)(0. bar Thus an analysis of the lost work. Calculation of ideal work: If changes in kinetic and potential energies are neglected.98)(1) − (117.35 *Properties on the basis of Miller and Sullivan.460 12.4 90 300 Air Pure O2 91. assumed to contain 79 mol % N2 and 21 mol % O2. 4-10 Diagram of simple Linde system for air separation.08 75.074)(0.171.08)(1)] − 147 = 7297. This is based on Eq. the work of compression is calculated by the equation for an ideal gas in a three-stage reciprocating machine with complete intercooling and with isentropic compression in each stage. whence x = 0.0 J 2.34)(0.850 5. all these work quantities are positive ˙s>W ˙ ideal. 424 (1928).1364) + (114. (4-387) is applicable.22 1.314 J/(mol·K) The efficiency factor of 0. (4-390) is only 3. J/mol S.01 1. as depicted in Fig.9148 x.4-40 THERMODYNAMICS Dividing Eq.8636) − (117.01 1.4453 JK (γ − 1)nγ P ˙ lost = (300)[(82.773 7. equal to 8.48% N2 Pure O2 Air State H. For the compressor/cooler. ˙ lost = (300)[(118.35)(1)= − 2.8 is already included in the equation. Example 7: Lost-Work Analysis A work analysis follows for a simple Linde system for the separation of air into gaseous oxygen and nitrogen.4)(8.08)(1) W − (75.532 J Q˙ = m˙ (∆H) − W Calculation of lost work: Equation (4-394) may be applied to each of the major units of the process. Compressor/cooler Exchanger Rectifier 589.171 J (0.0 J W Work analysis: Because the process requires work.34)(0.063. ˙s % of W ˙ ideal W ˙ lost: W ˙ Wlost: ˙ lost: W ˙s W FIG.4(3)(1. (4-401) For a work-requiring process. W ˙ ideal > W ˙ s.4) The heat transferred to the surroundings during compression as a result of intercooling and aftercooling to 300 K is found from the first law: ˙ s = (12. by Eq.046 13.48% N2 Air 91. Equation (4-401) in this case is best written as W ˙ ideal = W ˙ s + W ˙ lost W 55.5)0. Heat leaks into the column of 147 J/mol of entering air and into the exchanger of 70 J/mol of entering air have been assumed.4 T1 = initial absolute temperature.532) W = 5221. ˙ lost = (300)[(75. Take Tσ = 300 K. for the rectifier.4 13. By a material balance on the nitrogen.22 1.1364) + (114.34 52. ∆(Hm˙ )fs = (13.98 118.8636) − (83.69)(0.407) −15.8636 mol of nitrogen product 1 − x = 0.171 = −15.4 J Finally.6 J W Calculation of actual work of compression: For simplicity. Eq. written as W˙ lost ˙s−W ˙ ideal =W (4-402) A work analysis then gives each of the individual work terms in the ˙ s. is the same as an analysis of the rate of entropy generation. Pap. 4-10* Point P.0 J 3.297.9 34.6 J 5.0 The thermodynamic efficiency of this process as given by Eq. Eq. The preceding equation is then expressed as and W ˙ s= W ˙ ideal + W ˙ lost W 1 2 3 4 5 6 7 Thus.8636) + (83. and For a work-producing process. indiin the case where a process is so inefficient that W ˙ s is positive.074 5.01 T.79 = 0.8(γ − 1) (4-403) A work analysis here expresses each of the individual work terms on ˙ ideal.48)(0. here taken as 1. (4-393). A lost-work or entropy generation analysis is always possible. Bur.314)(300) ˙ s =  W (54.0 J For the exchanger.6 48. ˙ ideal = −144 − (300)(−2.48 114.0 J 15.1364) − (52. but W cating that the process in fact requires work.9 percent. An alternative to the lost-work or entropy generation analysis is a work analysis.82 83. 0.82)(0.1 100.1364) − (82. (4-387).4 J 7. The various terms of this equation appear as entries in the following table and are on the basis of 1 mol of entering air. J/(mol·K) Superheated Superheated Superheated Superheated Saturated vapor Saturated vapor Superheated 12.69 117. made by expressing each individual entropy generation term as a fraction of the sum of all entropy generation terms. here taken as 3 γ = ratio of heat capacities.2 79. equal to 54. summation on the right as a fraction of W ˙ s and W ˙ ideal are negative. (4-402) is appropriate for a work analysis. 4-10.4) − 1 = 15.221. A work analysis cannot be carried out the right as a fraction of W ˙ ideal is negative.98)(1)] − 70 = 2063. The following equation may be found in any number of textbooks on thermodynamics: nγ RT1 ˙s= W  0.407)(1) = −144 J ∆(Sm˙ )fs = (118.8)(0.5 R = universal gas constant. Table 4-8 lists a set of operating conditions for the numbered points of the diagram. indicating that the process should produce work.1364 mol of oxygen product P2 1 −1  where n = number of stages.4453) = 589. Substitution of the remaining values gives   (3)(1. K Composition 300 295 295 147. The work so calculated is assumed to represent 80 percent of the actual work.69)(0. Significant inefficiencies reside with each of the primary units of the process. made by calculation of the fraction that each individual lost-work term represents of the total lost work.046 − 12. (4-399) by this result yields ˙ lost W S˙ G =   ˙ W lost S˙ G TABLE 4-8 States and Values of Properties for the Process of Fig. The basis for analysis is 1 mol of entering air.460)(0.