Beruflich Dokumente
Kultur Dokumente
(8) Hinchley and Himus, Trans. Inst. Chem. Engrs. (London), 2, (17) Sherwood, Chemical Engineer’s Handbook, p. 1222, McGraw-
57 (1924). Hill Book Co., New York, 1933.
(91 Hine.
\-, - I~ -
Phvs. Rev.. 24. 79 (1924). (18) Stanton, Brit. Advisory Comm. Xironautics, Rept. & .)fern.,
94 (1913).
(10) Jurges, Beih. Gesundh. In;., 19 (1924).
(11) Kirkbride, IND. ENG.CHEM.,25, 1324 (1933). (19) Stauffer, Roberts, and Whitman, IND. ENG.CHEar., Anal. Ed.,
(12) Ibid.. 26, 425 (1934). 2, 88 (1930).
(13) Kirschbaum and Kranz, Chem. Fabrik, 7, 176 (1934). RECEIVED July 30, 1934. Presented as part of the Symposium on Diffu-
(14) Monrad and Badger, IND. ENG.CHEM.,22, 1103 (1930). sional Processes before the Division of Industrial and Engineering Chemlatry
(15) Prandtl, Physik. Z . , 11, 1072 (1910). a t the 88th AIeetmg of the Amerlcan Chemical Society, Cleveland, Ohio,
(16) Reynolds, Proc. Jlanchestw Lit.Phil. SOC.,14, 7 (1875). September 10 t o 14, 1934.
of 20 kg. per minute per square meter of tower cross section. ciently good to bring the two streams in equilibrium. The gas
The tower reduces the ammonia concentration of the air leaving plate CC, therefore, will be in equilibrium with the
mixture from 20 to 1 per cent ammonia by volume. Using liquid leaving the same plate. Consequently, for a column of
these data, the problem is to estimate the value of ~ G Ufor the theoretical plates, it is possible to represent the relation be-
absorption of acetone from air a t 0.5 atmosphere, using the tween gas and liquid compositions leaving any plate by the
same type of tower packing, tem- equilibrium curve for the system. The relation between the
Lpo 1% perature, gas, and water rates as in gas and liquid compositions between any two plates-i. e., the
#
the test on ammonia. The concen- relation between the composition Y C of the gas leaving CC
A tration is to be reduced from 5 to 1 and the composition X B of the liquid leaving BB, is given by
per cent acetone by volume. the material balance,
r--0 SOLUTION.In calculating values
G(Yi - Yc) = L(X1 - X B )
-&--1
I (3)
’--<- .6----j of ~ B M the
, solutions may be con- where Y = weight ratio of solute to inert or carrier gas
sidered sufficiently dilute so that the X = weight ratio of solute to solvent in the liquid
D D vapor pressures of acetone and of
ammonia may be neglected. In the Figure 2 shows a diagram for this case with the “operating
case of the ammonia, in the main line” PQ based on Eauation 3. and the equilibrium curve OR.
XI YI
gas stream P B varied from 2.0-0.2 The operating line gives the’ relation
FIGUREZ X 2 = 1.6 atmospheres a t the gas between the composition of gas leav- TREATED
SOLVENT tAs
inlet to 2.0 - 0.01 X 2 = 1.98 atmos- ing a plate and the composition of
pheres at the gas outlet. At the liquid interface, P B = 2.0, liquid from the plate next above. The
since the back pressure of ammonia is neglected. At theinlet, vapor leaving any plate, CC, will be
therefore, P E X is the mean of 2.0 and 1.6, or 1.7‘3, and at the in equilibrium with the concentra-
gas outlet p n is~ the mean of 2.0 and 1.98 or 1.99. The tion, X C , of the liquid on the plate,
average value of p B . V is 1.89 atmospheres. For acetone, ~ B M its composition, Y c , will be deter-
at the gas inlet is practically constant a t 0.49 atmosphere mined by referring to point c on the
throughout the apparatus. equilibrium curve. The liquid leav-
From Table I the ratio of the diffusivities‘ for acetone and ing plate BB above mill have a com-
ammonia is 0.097 to 0.201, The ratio of the molecular p o s i t i o n determined by point b‘ on
nTeights is 58 to 17. Hence, for acetone: the operating line, corresponding to
the gas composition, Yc. The com-
~ Q = 0.016 X
0.097 O.56 129 58 - position of the gas leaving BB may
U
(Om) 0.49 17 - then be determined, and the equilib-
0.140 kg./min./cu. m./mm. rium curve and operating line may
be used alternately until point P is
CONCEPTOF THE THEORETICAL
PLATE reached, corresponding to conditions
Both packed towers and plate columns may be used for gas a t the top of the column. With any
absorption, just as both types of equipment are found ser- fixed operating lines, this s t e p w i s e
’ viceable in distillation practice. Rectifying columns for dis- calculation may be employed to de-
tillation are often of the plate type, and the traditional basis termine the composition of liquid and
. for the design of such equipment is the “theoretical plate” gas on the various plates and, what
concept. Absorption equipment is commonly of the packed is more important, t h e n u m b e r of FIGURE 3
tower type, and the design of such towers should be based on theoretical plates required.
capacity coefficients such as Koa (weight diffusing per unit
RELATION BETWEEN Kaa AND H. E. T. P.
time per unit volume of packing per unit of driving force ex-
pressed as pressure difference). Just as the equipment is in- I n order to apply the theoretical plate concept to a packed
terchangeable in practice, so the design methods for the two tower, it is convenient to picture a certain height of packing
processes have been interchanged, and the “theoretical plate” as accomplishing the same separation as one theoretical plate.
has been used as the basis for absorption tower design, or The height of packing equivalent to one theoretical plate ha5
capacity coefficients been termed by Peters (3) the H. E. T. P. and its numerical
have been used in value is an indication of the capacity of the apparatus, as is
rectifying c o 1u m n the capacity coefficient KGU. However, large values of KQU
design. correspond to small values of H. E. T. P.
The design proce- Let Figure 1 now represent a packed tower, and the pack-
dure based on the ing between BB and CC perform the same separation as one
theoretical plate theoretical plate. Let A B and CD represent other imaginary
concept may be fol- divisions of the column into theoretical plates. From the
lowed on Figure 1. definition of a perfect plate, the gas leaving section BC a t BB
The horizontal lines will be in equilibrium with liquid leaving the same section a t
A A , BB, etc., repre- CC. Analysis of the process on the basis of the theoretical
sent bubble-cap or X - SOLUTE CONCENTRATION IN LIQUID plate concept assumes the flow to be somewhat as shown in
other plates, L kg. KGS S O L L m / K G . M L V - Figure 3 . On each plate the liquid from the plate above and
per minute of sol- FIGURE 2 the gas from the plate below come together and are separated
vent overflow i n g a t the points circled, where the two streams are assumed to
from one plate to the next down through the column. The be in equilibrium. I n an actual packed column the flow is
gas to be treated, containing G kg. per minute of inert carrier countercurrent throughout, but the analysis based on the
gas, passes up through the column countercurrent to the flow theoretical plate concept does not introduce serious error if
of solvent, being brought into intimate contact with the solvent the sections are sufficiently small-i. e., if the H. E. T. P. is
by bubble caps or other devices on each plate. In a theoreti- small compared to the total height of the column. When the
cally perfect plate the contact between phases would be suffi- gas film resistance is controlling, the potential or driving
1096 INDUSTRIAL AND ENGINEERING CHEMISTRY Vol. 26, No. 10
forces a t BB will be represented on Figure 2 by b‘b. At CC fusing per unit time per unit volume of packing per unit driv-
the corresponding potential will be represented by c‘c. ing force expressed in terms of Y . The variable B is the ratio
Assuming the equilibrium curye to be straight over the short of the mean driving force in pressure units to that expressed
range, bc, the logarithmic mean may be used, and the equa- in stoichiometric units. For very dilute gas:
tion written:
B = M A / M E P where M A , M E = mol. weights of solute and
carrier gases, respectively
P = total pressure
where S, V = cross-sectional area and volume of the column, The derivation given above provides a relation between the
respectively capacity coefficient, Koa, and the H. E. T. P. The vaporiza-
tion data obtained in the wetted-wall column support the
application of the diffusion equation to an extent which indi-
V G In J cates that KGUis the more fundamental factor, and the
Then - = H.E.T.P. = BKGa(J - 1) (5) H. E. T. P. a more involved variable. Equation 7 shows that
S
the H. E. T. P. is dependent, not only on those variables which
Ifzm = ( Y c - Y B ) / ( X C- X E ) represents the slope of the affect Koa, but also on additional effects of gas and liquid
equilibrium curve over the range of concentrations of the plate weight velocities, G and L, and on the slope, m , of the equilib-
in question, and since the slope of the operating line is L;G = rium curve. Because of the influence of the latter variable,
( y o - YC)/(xC - XE), the H. E. T. P. may be expected to vary through the apparatus
with solute concentration. No such variations in Koa may
J = -mG be expected, unless there are large variations in temperature,
L
pressure, or EM. For these reasons it appears logical to dis-
Even when the slopes of the operating and equilibrium courage the reporting of data on gas absorption in terms of
curves differ considerably, it is usually allowable to employ values of the H. E. T. P. The value of the H. E. T. P. concept
the arithmetical in place of the logarithmic mean potential for rectification in packed columns may be somewhat greater,
in Equation 4, and so eliminate the logarithmic mean. The but even there the capacity coefficient Kca would appear to
result is : be a sounder basis for dealing with experimental data.
gv = H. E. T. P. = BKoa(J
2G
+
- 2GL
+
1) - BKGu(L mG) (7) LITERATTJRE
CITED
For the special case where J is substantially unity, both (1) Gilliland, IND. ENO.CHEM.,26,681 (1934).
(2) Gilliland and Sherwood, Ibid., 26, 516 (1934)
Equations 5 and 7 reduce to: (3) Peters, Ibid., 14,476 (1922).
RECEIVEID August 6, 1934. Presented as part of the Symposium on Diffu-
sional Processes before the Division of Industrial and Engineering Chem
istry a t the 88th Meeting of the American Chemical Society, Cleveland,
The term BKGa is defined by Equation 4 as the weight dif- Ohio, September 10 t o 14,1934.