INDUSTRIAL AND ENGINEERING CHEMISTRY
February. 1930
165
Graphical Solution of Problems Involving Solvent Recovery by Scrubbing’ Irvin L. Murray CARBIDE AND CARBON CHEMICALS CORPORATION, SOUTH CHARLESTON, W.VA.
I
N MANY chemical proc-
om+
A method is given for the calculation of the overflow
+
1
+
Zn = Xn+1 and the number of plates which should be used in a on+1 (P - 0 ) esses it is necessary to ZPP recover a relatively small column for the recovery of a volatile compound from (3) 0”+1 (P - 0 ) amount of valuable volatile a permanent gas. The transposition of the liquidBut material from a large volume vapor curve from the boiling point to the scrubber 1 0,+1 = 0 x ___ of inert gas. An inlportant temperature is accomplished by employing a correction 1 - Xn+1 factor to Raoult’s law. method of accomplishing this Substituting and rearrangobject is the empioynieni of a ing, 0 Z,P(l - X,+J scrubber fed with a liquid of low vapor pressure in which the zn= Xn,l p - ( P - O)X,+t + P - ( P - 0)Xn+1 (4) valuable compound is soluble. Heretofore it has been difficult to determine the proper size of the equipment and the While this equation is a curvilinear function, the curvature rate of feed, and the apparatus has usually been designed by is normally so slight that it may be assumed to be a straight guess or after a long series of difficult experiments. It is obvious that the operation bears a certain relationship line with a slope equal to (W - 0 ) p (substituting TV for to distillation, and advantage has been taken of this fact to ow 1). apply the recently developed graphical methods of analysis The basic equation is thus of distillation problems to this special case (1.1. W
+
+
2,
Nomenclature
= mols of gas entering scrubber
o
=
P
xn+l(W - 0 ) + p +
w =
x = mol fraction of more volatile component in liquid z = mol fraction of more volatile componerit in gas z, = mol fraction of more volatile component in gas leaving
D
= =
PB
=
I.’= P’* = PI* =
leaving scrubber mol fraction of more volatile component in gas entering scrubber total pressure on system vapor pressure of more volatile component a t scrubber temperature vapor pressure of less volatile component a t scrubber temperature mol fraction of more volatile component in vapor a t boiling point and 760 mm. vapor pressure of more volatile component a t boiling point and 760 mm. vapor pressure of less volatile component a t boiling point and 760 mm.
Assumptions
(1) The heat of absorption of the more volatile component
has a negligible effect on the temperature of the scrubbing liquid. (2) The scrubbing liquid obeys Raoult’s law. (3) The more volatile component obeys Henry’s law, written Y = S P A X , where S is variable with Y but is unaffected by temperature. Derivation
Considering the whole scrubber above a point just below 1, a complete material balance may be written as plate n follows:
+
vn+o=o,+l+
P
(1)
A balance of the more volatile component may also be written: Z n K = Xn+1 On+1 ZpP (2) Eliminating ITn,
+
Received Sovernber 22, 1929.
+P
(5)
z = p- AX
scrubber
PA
(W - P0 )
The stepwise progressive method may be used. According to Raoult’s law,
D
= mol fraction of more volatile component in liquid =
zp
Graphical Solution
mols of liquid entering scrubber
= mols of gas leaving scrubber mols of liquid leaving scrubber
x, z,
=
Raoult’s law does not hold, however, in the majority of cases. It is necessary to apply a correction factor for each value of X. I n the range where the mol P fraction of the more volatile component is low-the range where scrubber problems are i m p o r tant-the less volatile component acts as the solvent and approximately obeys Raoult’s law; while the more volatile component obeys Henry’s law:
z = -SPAX D where S is an experimentally determined variable dependent on t h e v a l u e of X. I n the low range of concentrations the actual partial pressure of the more volatile component is abnormally high, while that of the less volatile is normal. S can be determined from a consideration of the liquid-vapor curve a t the boiling point and 760 mm. by the following method:
W
V
Figure 1-Diagram of Scrubber Column
Plot the X - Y curve of the two components from experimental data a t the boiling point and 760 mm. Plot Raoult’s law for the two components under the same conditions. Call Y the ordinate on the true X Y curve and Y’ the ordinate on the curve calculated by Raoult’s law.
-
1NDUSTRIAL AND Eh'GINEERING CHEMISTRY
166 Then and Hence
'E SP'AX = P'B(1 - X ) 1- Y P'AX = - Y' P'B(1 - X ) 1 - Y' Y(l - Y ' ) S= Y'(1 - Y)
VOl. 22, No. 2
to the overflow line, etc., until the point Z = Z, is reached. The number of steps is the number of theoretical plates in the scrubber.
Example
A scrubber for the removal of acetone from air is fed with 1200 pounds of water per hour. 10,000 cubic feet per hour of acetone-free air are blown through the scrubber. The inlet gas analyzes G per cent acetone by volume. The water temperature is 20' C. Assuming that 98 per cent of the acetone is t o be removed, how many plates should be used? Solution
(1) Plot the liquid-vapor curve of acetone and water a t the boiling point, using the data in the literature. The average boiling point in the range 0 to 10 mol per cent acetone is 85' C., a t which temperature the vapor pressure of acetone is 1876 mm. and of water, 440 mm. The volatility ratio is thus 4.27. (2) Plot Raoult's law from these data between 0 and 7.5 mol per cent acetone in the liquid. (3) Calculate S for values of X up to 7 mol per cent. Y 24.1 37.5 44.0 54.8 60.6 66.2 71.0
Y' 5.5 10.7 15.4 19.8 23.8 27.6 31.0
S
The variable S is probably slightly affected by temperature, but this need not be allowed for in the calculations, for the following reasons: (1) The temperature variation is slight. Recovery of volatile solvents is a problem only when the solvent boils near the scrubber temperature, and the transposition of the X - Y curve a t the boiling point to the scrubber temperature is usually only a matter of a few degrees. , (2) S is a correction of the volatility ratio P ~ / P B which itself changes only slightly with temperature.
Therefore, knowing S for each value of X a t the boiling point, Z can be calculated by the equation
z = -SP A X
D where D is the total pressure on the system. The solution may be summarized as follows: (1) .Plot the liquid-vapor curve of the two components a t the boiling point and 760 mm. ( 2 ) Plot Raoult's law under the same conditions. (3) For each desired value of X , determine S, where S =
Y'(1
-- ")Y)'
Y is the ordinate on the true X - Y curve, and Y'
is the ordinate on the curve calculated by Raoult's law. (4) Plot S against X. ( 5 ) Plot 2 against X , using the equation
z = -SP Ax D
(6) Starting a t the point lay off a line having a slope of Note-If
X
= 0 ; 2 = Z,
( W - 0) + p
P ' ( W - 0) P'
+
(the overflow line).
necessary, the overtlow line may be plotted from Equation 4: z = (0 z p m z 2, p (0 P)L. P
-
-
+ +
(7) Lay off a horizontal line Z = Zf.
(8) Starting a t the point where this line intersects the line plotted in step 6, draw a vertical to the X - Z curve, a horizontal
These values are so nearly constant in this range that the average (4.83) will be used. (4) Plot 2 vs. X a t 20" C. and 760 mm., using the equation z = -S P A X (4.83 X 186)X =
D 760 It is next necessary to determine the slope and intercept of the operating line. W The slope is given by the expression ( W - 0) + P W = 1200 0'06 = 68.34 mols per hour 18 (neglecting the small amount of acetone and water leaving in the gas) 1200 0 = - = 66.67 mols per hour
+
("'",,) 18
P = 1o,oOo = 27.87 mols per hour 359 68.34 = 2.31 The slope is thus 68.34 - 66.67 f 27.87
INDI;STRISI; A S D E,YGISEERISG CHEJIISTRY
February, 1930
The tolerance of acetone in the outlet gas is: 10,000 X 0.06 X 0.02 = 12 cubic feet per hour, or 0.12 mol per cent. The intercept on the 2 axis of the operating line is therefore
P
zp
(W - 0)
+P
__
27.87
68.34 - 66.67
- = + 27.87
The operating line is now plotted, and the number of theoretical plates developed in the usual manner. The number of steps on the operating line is 6, and this is the theoretical number of plates.
Literature Cited
0.113
mol per Cent.
167
(1) McCabe and Thiele, IND. ENG.CHEM.,17, 605 (1923).
-
of Comparative Efficiencies of the Components Creosote Oil as Preservatives for Timber' F. I€.Rhodes and F. T. Gardner CORXELLUNIVERSITY, ITHACA,h-.1 ' .
A new method for the determination of the fungicidal T'arious methods have been H E value of c o a l - t a r powers of preservatives for timber is described. suggested for the accelerated creosote oil as a preThe fungicidal powers of various fractions from the testing of timber preservaservative for timber is dead oil, the tar acids, and the tar bases from creosote tives under controlled condiclue in part to its action in exoil, and of some of the pure components of creosote oil tions in the laboratory. Sevcluding moisture from the were measured. eral investigators h a v e dewood. The oil w e t s wood Phenolic compounds vaporize much less readily from termined the minimum conreadily, so that the surface of wood than do aromatic hydrocarbons, probably becentration of the preservative the treated timber is covered cause the phenolic compounds wet the wood more which is required to inhibit with a rather firmly adhering readily and are more strongly adsorbed by it. The the growth of a wood-rotting film of water-repellent matepossible action of tar acids as mordants for creosote fungus in an artificial culture rial. Since moisture is necesoil on wood is discussed. medium and have taken this sary to the growth of the fungi The vapor pressures of various fractions and mixtures minimum inhibiting concenw h i c h d e s t r o y wood, the of fractions of creosote oil were measured. tration as a direct measure formation of such a water-reThe percentages of certain hydrocarbons in the fracof preservative efficiency. pellent film aids in protecting tions from coal-tar creosote oil were determined. W o r k d o n e by this method the timber. I n addition t o Drior to 1915 has been sumthis general sealing a c t i o n , some-of the components of the oil have a specific toxicity for marized by Humphrey and Fleming (20). Kithin recent years fungi. It was formerly thought that the specific fungicidal the method has been employed by Dehnst ( I S ) , Falck (15), power of creosote oil was due largely or entirely to its content Makrinov and Shtrobinder (&3),and Schmitz (29), and by of phenolic compounds. At the present time it is rather Bateman and his co-workers (4 to 9) a t the U. S. Forest generally recognized that some of the other components, Products Laboratory. Bateman and his associates, working with cultures of Foines such as the hydrocarbons, may also be toxic l o fungi, although information as to the exact nature of the compounds annostis on agar-agar media, have modified this general principally responsible for the specific fungicidal properties method to render it suitable for measuring the inhibiting action of solutions which are so dilute or so ineffective that of creosote oil is still rather incomplete. Various methods have been suggested for comparing the growth of the fungus is merely retarded and not comthe efficiencies of different types of creosote oils and for de- pletely prevented. These investigators concluded that only termining quantitatively the preservative powers of the in- those compounds which are a t least slightly soluble in water dividual components of coal-tar creosote oil. Rather ex- have fungicidal properties and that, among similar comtensive service tests (3, 10, 11, 1.2, 14, 16, 18, 18, 22, SO, 32, pounds, the efficiency as a preservative is approximately 33,34,35, 36) have been conducted for many years, and from proportional to the solubility. Phenolic compounds with the data thus collected conclusions have been drawn as to the lower molecular weights than that of naphthol showed a t types of preservatives and the methods of impregnation which least partial inhibiting action; those with higher molecular may be expected to give satisfactory results under various weights were not fungicidal. Alpha-naphthol had no effect conditions of service. Our present specifications for creosote upon the growth of the organism, while beta-naphthol showed oils for various purposes are based partly upon the results of some inhibiting action. Aromatic hydrocarbons with molecuthese tests and partly upon general impressions gained by lar weights less than that of naphthalene completely preexperience in the use of treated timber. While the protection vented the growth of Fomes unnosus, those of higher molecular afforded under actual service conditions must always be the weights showed partial but not complete inhibiting action. final criterion of the value of a preservative, the service test I n mixtures of two or more hydrocarbons, none of which is not a satisfactory guide in the control and improvement of prevented growth entirely, the effects of the individual comthe quality of creosote oil. A very long time is required to ponents were geometrically additive. obtain the final data; tests made on a scale large enough to The use of Fomes annosus as a standardizing fungus was be significant are expensive; and the final results may be so criticized by Dehnst ( I S ) , who pointed out that this organism greatly influenced by unavoidable variations in the conditions is usually found as a parasite on living timber and only of service that direct comparison of the results of different rarely occurs on timber in service. Despite this fact, Fomes series of tests may not be possible. annosus does offer several advantages as a test organism in the comparison of the efficiencies of preservatives. It is rather Received h'ovembt*r 21, 1929. T h e nork described in this article easily obtained in pure cultures and can be propagated in the % a s done under a fellowship maintained a t Cornell University by the American Creosoting Company. laboratory readily and without loss of virility. It appears
T
