Page 87 - Separation process principles 2
P. 87

52  Chapter 2  Thermodynamics of Separation Operations


                Table 2.9  Empirical and Semitheoretical Equations for Correlating Liquid-Phase Activity Coefficients of Binary Pairs
                Name                              Equation for Species 1                Equation for Species 2

                (1) Margules           log  yl  = AX;                        log y2  = AX:
                (2) Margules (two-constant)   log yl  = x;[AI2 + 2xl(Az1 - AI2)1   1%   = x?[A21 + 2~2(A12 - A2111
                ,,.\   7   ,                          A 12                     1.           A21
                (3) van Laar (two-constant)   lnyl =                           1ny2 =
                                               [I + (xIAIz)~(xzAzI)I~                [1 + (xzAzI)~(xIAI~)I~
                (4) Wilson (two-constant)   111 y1 = - ln(xl + A 12x2)         1ny.r = - ln(x2 + 1221x1)
                                               +x2 (   1212   -                             1\12   -

                                                    xi + A12xz  xz + A21xi                XI + A12~2  ~2  + AZIXI
                                                 x;nl~:,   +  x:.~IzG~z               x:.r12Gh   +  xi721 GZI
                (5) NRTL (three-constant)   lnyl =                             1ny2 =
                                                                                                  (XI
                                               (xi + x2G21I2   (x2 $11   G1d2        (XZ + XIG~~)~ +x2G21I2
                                          Gij = exp(-aij.rij)                   Gij = exp(-aij.rij)

                                                 1         01                         2          02
                (6) UNIQUAC (two-constant)   In yl = In - + -91  In -          Inn = In - + -q21n  -
                                                 XI   2    q1                          12   2    qz
                                                                                                  - 92 ln(02 + 01T12)
                                                                                              T12   -
                                                                                    + 'lq2  (02 + 01 T12  01 + 02T21





                Margules Equations
                                                                   constants, shown as AI2 and Azl in (3) of Table 2.9, are best
                The Margules equations (1) and (2) in Table 2.9 date back   back-calculated  from  experimental  data.  These  constants
                to  1895, and the two-constant  form is still in common use   are, in theory, constant only for a particular binary pair at a
                because of its simplicity. These equations result from power-   given temperature. In practice, they are frequently computed
                series expansions in mole fractions for jf  and conversion to   from isobaric data covering a range of temperature. The van
                activity  coefficients  by  means  of  (2-59). The one-constant   Laar theory expresses the temperature dependence of Aij as
                form  is  equivalent  to  symmetrical  activity-coefficient                 A!
                curves, which are rarely observed experimentally.                     A,. - '.'               (2-70)
                                                                                        " - RT
                van Laar Equation
                                                                   Regular-solution  theory  and  the  van  Laar  equation  are
                Because of its flexibility, simplicity, and ability to fit many   equivalent for a binary solution if
                systems well, the van Laar equation is widely used. It was
                derived from the van  der Waals equation of  state, but the       A. - --(& V~L  - 6.)2
                                                                                             1   J            (2-7 1)
                                                                     The  van  Laar  equation  can  fit  activity  coefficient-
                Table 2.10  Classical Partial Molar Excess Functions of   composition  curves  corresponding  to  both  positive  and
                Thermodynamics                                     negative deviations from Raoult's  law, but cannot fit curves

                Excess volume:                                     that exhibit minima or maxima such as those in Figure 2.1%.
                                                                     When data are isothermal, or isobaric over only a narrow
                (I) (ztL -n)p) =ifL RT(+)                          range of temperature, determination of van Laar constants is
                                =
                                             T,x
               Excess enthalpy:                                    conducted  in a straightforward manner. The most accurate
                                                                   procedure is a nonlinear regression  to obtain the best fit to
                (2) (hlL - iip) E ipL = -RT~ -
                                       (a      p,x                 the data over the entire range of binary composition, subject
               Excess entropy:                                     to minimization of some objective function. A less accurate,
                                                                   but  extremely rapid,  manual-calculation  procedure  can be
                (3) (5  -   = iFL = - R  T   -                     used when experimental data can be extrapolated to infinite-
                       i;:)
                                                 + 1nYtL]
                                    [  (a  ;;L),,~
                                                                   dilution  conditions. Modem  experimental techniques  are
               ID = ideal mixture; E = excess because of  nonideality.   available  for  accurately  and  rapidly  determining  activity
   82   83   84   85   86   87   88   89   90   91   92