258
IXDUSTRIAL AND ENGI,VEERING CHEMISTRY
Vol. 22, s o . 3
Raoult’s Law and the Equilibrium Vaporization of Hydrocarbon Mixtures’8z Marvin C. Rogers and George Granger Brown UNIVERSITY
OF
MICHIGAS, ANN
.kRBOR,
hIICH.
A quantitative study has been made to determine given a t 20 per cent of the the extent to which Raoult’s law holds in calculations porization of hydrocarlighter component are: for involving the vaporization of hydrocarbon mixtures. bon mixtures it has genbutane-pentane, 4 per cent; The following mixtures were studied : three natural for butane-heptane, 0 per erally been assumed that gasolines, a synthetic mixture, three mixtures of cent; for pentane-heptane, 7 Raoult’s law held with natural gasoline and a midcontinent absorption oil, per cent; for butane-benzene, sufficient accuracy for engia mixture of a naphtha and natural gasoline, and a 64 per cent; and for butanrneering calculations a t least. mixture of tetralin and natural gasoline. After analyBut apparently no quantitastraw oil, -2.6. per cent. sis of the mixtures by a special fractional distillation No definite conclusions can tive study has been reported procedure and molecular weight determinations on be drawn from these data as to show that this assumption the various residues, vaporizations were made on these to what might be expected has been justified. mixtures and the results compared with the curves Raoult’s law may be stated from a ternary or more comcalculated by Raoult’s law. plex mixture, but in general in ordinary terms as follows: The results show that deviations from Raoult’s law the deviations seem to be “In an ideal solution of like when applied to complex mixtures are dependent on quite erratic. components the partial vapor the relative composition of the mixtures and that the hlaxson (8) made a short pressure exerted by each comvapor pressure of natural gasolines and similar mixtures study of two absorption ponent is proportional to its cannot be calculated by this law when a precision oils, one of which he termed mol fraction in the solugreater than 5 to 15 per cent is desired. a “long-range oil” with an tion.” This means that the i n i t i a l b o i l i n g p o i n t of p a r t i a l pressure i n t h e iapor phase should be equal to the product of the mol 500” F. and an end point a t 700” F.; the other oil, fraction of a component in the liquid by its vapor pres- having an initial boiling point of 442” F. and an end point a t 460” F., he called a “narrow-range oil.” Vaporsure in the pure state. pressure data were taken on samples of oil and gasoline Previous Work varying in composition from 1 to 20 per cent gasoline by Brown and Caine (1) have gone into the experimental weight. The short-range oil showed the lowest vapor determination of vapor-liquid equilibria in complex hydrocar- pressure and might be considered the better oil for absorption bon mixtures. Three heavy naphthas from Pennsylvania, purposes. This is in agreement with Raoult’s law, since the California, and midcontinent crudes were vaporized and the short-range oil had a lower molecular weight and the mol overhead and bottoms mere analyzed in a ‘‘true boiling fraction of gasoline was somewhat lower for a given weight point” column similar to that described by Peters (10). per cent than it was in the high-boiling long-range oil. The From these data and the amounts vaporized in the equilibrium other considerations which he mentions are mechanical vaporization, a method was developed whereby approximate and are concerned principally with the handling of the oil i n over-all deviations from Raoult’s law could be determined an absorber. Davis ( 4 ) , in a study of an ideal absorption oil for natural over the range of the compounds contained in the naphthas. The choice of the vaporization temperature or the amount gasoline, calculated vapor pressures for binary mixtures to be vaporized was apparently optional and not prede- of hydrocarbons by Raoult’s law, and found that the following termined. The results of the investigation indicate that, combinations gave the lowest total pressure for a given in the naphthas studied, the calculated vaporizations varied composition: as much as 10 to 15 per cent from those which would be SOLUTE SOLVENT indicated by Raoult’s law. Inasmuch as this study was Hexane Ethane Hexane Propane based entirely upon a single vaporization of each of three Octane Butane Pentane Decane naphthas, there may be some question as to whether the Decane Hexane variations noted by them apply through wide ranges of vaporization and composition. I n his experimental work he mixed a representative gasoline Somewhat similar studies by Piroomov and Beiswenger whose composition gave it a molecular weight about that of (11) seem to indicate that an equilibrium vaporization curve pentane with several compounds, and found that the ideal may be calculated by means of Raoult’s law with reasonable absorbent for the mixture was a compound with molecular accuracy for any of the types of oils studied. This method weight approximately 28 units higher than decane, which he was used by Podbielniak and Brown ( I S ) with good agree- computed to be the best absorbent for a mixture of this compoment between calculated and experimental results. sition. McLouth (9) made a similar study and came t o the Calingaert and Hitchcock (2) reported vapor-pressure data same conclusions as Davis. and deviations from Raoult’s law for the following pairs I n Europe a considerable amount of interest has been disof compounds: butane-pentane, pentane-heptane, butane- played in the use of tetralin (tetrahydronaphthalene) for gas benzene, and butane-straw oil. The values of the deviations scrubbing and absorption of all sorts of materials from 1 Received October 28, 1929. Presented at the meeting of the American gas. Weissenberger (14) conducted some experiments on Institute of Chemical Engineers, Philadelphia, P a , June 19 t o 21, 1929. the properties of tetralin as a natural gasoline absorbent, 2 Abstracted from a thesis submitted by Marvin C. Rogers to the and has shown its superiority over spindle oil and gas oil. Graduate School, University of Michigan, in partial fulfilment of the reHis work shows a better removal of gasoline from the gas by quirements for the degree of doctor of philosophy.
I
N TREATIKG the va-
INDUSTRIAL A N D ENGINEERING CHEMISTRY
March, 1930
tetralin than by the other oils. He does not, however, give molecular-weight data from which one can compute the relative values of absorption from Raoult's law, and from the values which are given it seems that Raoult's law holds fairly well and that the difference in the amount absorbed is due to the low molecular weight of the tetralin. THERMOCOUPLE 2~
9
Figure I-Low-Temperature
Fractionation Apparatus
259
amount of each component present in the original liquid with reasonable accuracy. I n the case of binary mixtures this offers no particular difficulty. I n studying mixtures of hydrocarbons such as are found in petroleum and natural gasoline, however, the analysis of liquid mixtures is one of the greatest difficulties encountered. The apparatus used in this investigation was similar to that described by Podbielniak (12) and consisted essentially of a fractionating column connected through suitable tubing to an evacuated receiver in which the products of the distillation were collected as a gas (Figure 1). The column was constructed of Pyrex glass and consisted of a tube 3.5 to 4 mm. inside diameter and 70 cm. in length. To one end of the column the still, B , and to the other end the outlet tube, D, and the reflux chamber, C, were attached. The entire column and reflux chamber, as well as part of the still, were surrounded by a vacuum jacket to prevent radiation and make the column as adiabatic as possible. It was packed with a chromel or copper wire helix coiled so as to fit closely to the side and cause a maximum surface film of liquid for contact with the rising vapor. Heating of the liquid in the still was provided by a heating element of KO. 30 chromel wire welded to tungsten leads sealed into the glass, providing outside contacts.
Equilibrium Vaporizer
The apparatus used to obtain the vapor-liquid equilibria in this investigation was practically identical to that described by Podbielniak and Brown (18). Final condensation of the vapors was obtained in the glass trap, immersed in a vacuum bottle full of natural gasoline and carbon dioxide snow a t a temperature of -79" C., as liquid air was found too cold for collecting the distillates as a liquid. With the receivers in place the sample was fed into the vaporizer and the feed governed so as not to exceed 10 cc. per minute, as this rate was known to be below the maximum a t which equilibrium is obtained. This had been determined by Leslie and Good (?) for the same type of apparatus, by vaporizing a chloroform-toluene mixture and checking the analysis of the liquid bottoms against that of the overhead vapor. -
111
PROPANE. I = BUTANE N. BUTANE I = PENTANE N: PENTANE I HEXANE N- HEXANE I s HEPTANE N. H E P T A N E RESIDUE
A-
u
t
I
I 2
I
I
l
I
I
1
l
I
0
-20
0
20 AMOUNT
40
Figure 3-Fractional
,*.&j$. I! 192
Jy I
IW 0
100
b
I . c / BUTANE
I
ZOO 300 RECENER PRESSURE IN MM MERCURI.
400
4
0
Figure 2
The feed was allowed to run until sufficient overhead and bottoms were obtained for analysis. Then the remaining sample in the buret was drained back into the sample can to determine the actual amount used. The liquid bottoms and the overhead vapor condensate were then weighed to determine the material balance and the percentage vaporized. Fractional Distillation Analyses
For any quantitative study of the deviations from Raoult's law in vapor-liquid equilibria, it is necessary to know the
60
80
100
120
140
D I S T I L L E D - R E C E N E R PRESSURE MI M M .
Distillation of Fuel 72
The top of the column was cooled by siphoning liquid air into a brass container placed in a bath of petroleum ether in the reflux chamber. The temperatures a t the top of the column were determined by three copper-constantan thermocouples in series, connected to a millivoltmeter. .F was an open-arm manometer and H a closed-arm manometer. The products of the distillation were collected in the evacuated receiver, L. PROCEDURE I N FRACTIONATINQ SAMPLES-In order to prevent changes in composition of the mixtures under investigation, the samples were kept in sealed cans or drums and cooled in a brine tank maintained at -20" C. Preliminary to an analysis, the column and the receiver were evacuated to about 1 to 2 mm. with a two-stage mechanical vacuum pump. The top of the column was then cooled to a temperature below that a t which the lowest boiling component boils. The sample was then withdrawn from the
I N D C S T R I A L A S D ENGINEERIXG CHEMISTRY
260 I""
1
I
1
ANALYSIS BUTANE I*PENTANE N* PENTANE I = HEXANE NSHEXANE I. HEPTANE N. HEPTANE
28 ZU 18.80
5
80
60
I
1
1
I
ANALYSIS MOL% BUTANE 4 29 I PENTANE 1'2% N PENTANE 2770
1
906 13 47
I=HEPTANE
355 5 . 35 I
,
1
581 18.02
RESlWE 4 a W
I
MOC%
Vol. 22, No. 3
I
I !
d
U
w 80
a
N=HEXANE
t t
I
60
5
40
I I= P E N T A N E
i
20
0
I
I
I
-20
0
20 AMOUNT
40
60
DISTILLED-RECEIVER
Figure 4-Fractional
100
80
120
140
PRESSURE I N M M
Distillation of Fuel 75A
-20
I
0
20 AMOUNT
of Mixture 72
cooled container and admitted to the still. The pressure on the column, as shown on manometer F , was allowed to come t o 760 mm. and the distillation started. The product from thc distillation was removed from the top of the column by cracking the cock G and permitting the gas t o be drawn into the receiver. The closed-arm manometer, H, measured the pressure on the receiver. During the removal of any one compound the temperature a t the top of the column remained practically constant, but when the compound was practically removed from the sample the rate of distillation was decreased and the amount of reflux increased slightly, in order to obtain close fractionation. Inasmuch as the compounds were collected as gases, it was essential that the reflux temperature be below that of the room. When the reflux temperature rose to about that of the room, the pressure on the column was reduced. I n making the analyses the compounds through butane were distilled a t 760 mm., pentanes a t 400 mm., hexanes a t 100 mm., and heptanes a t 30 mm. CALIBRATION OF COLUMN AND RECEIVER-It was necessary to know accurately the volunie of the receiving bottle, L , in order to determine the amounts of the compounds collected. Although the volume of the receiver had been carefully determined a t atmospheric pressure, it was observed during the course of the analyses that the recovery, calculated from the distillate and residue, if any, did not equal the amount of sample originally fed to the column. I n all cases the calculated recovery was less than the amount of sample fed, and the loss was greater in the case of samples containing large amounts of hexane and heptane. This led to an investigation of the properties of the hydrocarbon gases when
80
60
DISTILLED-
Figure 5-Fractional
WEIGHT % VAPORIZED
Figure 6-Vaporization
40
RECEIVER
100
120
140
PRESSURE IN MM.
Distillation of Fuel 48
WEIGHT % VAPORIZED
Figure 7-Vaporization
of Mixture 75A
collected in glass containers which have been previously eyacuated. The samples used for the determination of the corrections were normal heptane, hexanes, pentanes, and isobutane. The normal heptane was obtained from fractionation and was that part of the sample boiling under 760 mm. pressure at 98.4" =t0.2" C. The hexanes were a mixture of diisopropyl and normal hexane with less than 1per cent of pentane and nothing heavier than hexane in the sample. The pentanes were a mixture of iso- and normal pentane and contained 1e.s than 1.5 per cent of butane and nothing heavier than pentane. The butane contained over 95 per cent isobutane with the remainder normal butane. KO attempt was made to prepare the pure compounds because the correction difference between iso- and normal pentane could not be detected in the apparatus used. of Corrections for Distillation Analysis CORRECTION ROOMRECEIVER WEIGHT WEIGHT RECEIVER VOLUME RECOVERED FACTOR PRESSURE TEMP. COMPOUNDOF SAMPLE cc. Grams M m . Hg C. Grams 1.01 29.4 19,875 488.5 25 29.7 Butane 1.005 6.23 19,875 90.0 25 6.26 Butane 1.02 18.41 19,875 238.0 24 18.78 Pentane 1.01 5.40 19,875 70.0 24 5.46 Pentane 1.03 28.52 19,875 369.0 24 29.29 Pentane 1.042 38.01 19,875 495.0 24 39.60 Pentane 1.036 9.57 19,875 104.0 25 9.90 Hexane 1.042 11.37 19,875 126.3 24 11.87 Hexane 1.021 2.48 19,876 30.6 24 2.53 Hexane 1.020 2.59 19,875 28.0 24 2.64 Hexane 1 .08 2 . 3 1 19,875 21.7 25 2.495 Heptane 1.10 2,. 33 19,875 21.9 25 2.575 Heptane Table I-Determination
Each of these samples was weighed and vaporized in the fractionating column. The change in pressure in the receiver and the receiver temperature were recorded. From
I N D U S T R I A L A N D ENGINEERING CHEMISTRY
March, 1930
261
“1 1 ---
EXPERIMENTAL CALCULATED PRESSURE = 740
0
IO
40 50 80 WEIGHT % VAPORIZED
30
70
80
90
100
Figure 8--Vaporization of Mixture 48
these data the amount of the compound actually vaporized could be compared with that calculated using the molal volume of hydrocarbons as 22,410 cc. The data so obtained are given in Table I, and the correction factors plotted as a function of pressure a t 25’ C. in Figure 2. The corrections shown in Figure 2 have been applied to all the analyses made during this investigation, and the agreement between the recovery and the sample fed has been within 1 per cent in all cases. There is an indication from these curves that eten a t zero pressure a correction must be made. This would seem unreasonable from any ordinary considerations, hecause at zero pressure the gases would be expected to act like ideal gases. h‘lost of the apparent correction a t very low pressure is qrobably due to error or changes in the volume of the receiving system. The quantity of compound distilled is seldom so small as to giye a pressure less than 10 mm. and, even if due to experimental error, these corrections a t extremely low pressures will introduce no noticeable error in the analysis.
Figure 9-Vaporization
The actual procedure followed in making the molecularweight determinations was that given by Findlay (6). Table I1 gives the results of these determinations. Table II- -Molecular-Weight Determinations ITEIGHTOF W E I G H T OF DEPRESSION OF MOLECULAR SOLUTE SOLVENT FREEZING POINT WEIGHT Grams Grams c. SAPHTHA
2oo!---
i
’ 8 9 2 MOL 7, GASOLINE IN SAMPLE ---EXPERIMENTAL -CALCULATED PRESSURE
.
740.
Figure IO-Vaporization
I
!i
,,
of Oil-Gasoline Mixture
0.7098 0.4898 0.2943
22 22 22
0.1826 0.4405 0.7144
22 22 22
0.895 0.628 0.386 Av. at 0 concn.
184.6 180.5 177.5 172.5
0.203 0.464 0.735 Av. at 0 concn.
209.0 221.0 226.0 204.0
OIL KO. 1
RESIDUE FROM SAXPLE 48
0.1939 0.3790 0.5955
22 ’2 22
0.370 0,705 1.073 Av. a t 0 concn.
121.7 125.0 129.0 118.1 ~~~
RESIDUE FROM KO. i5A
0.1685 0.4202 0.6527
D e t e r m i n a t i o n of Molecular Weights
hIolecular weights of the various residues of unknown composition were determined by the method 0; freezingpoint depression as used by Wilson and Wylde (16)for determining the average molecular weights of absorption oils. In order to eliminate the possible errors caused by association, a series of determinations was made using various concentrations of oil in benzene. The molecular weight as a t zero concentration was obtained by plotting the molecular weight as a function of the depression in freezing point and extrapolating to zero depression corresponding to zero concentration. Inasmuch as the oils used in this work were obtained from petroleum, they contained a mixture of a large number of compounds and any determination made would yield only an average value.
of Mixture 4XA
22 22 22
0.311 0.775 1.176 Av. at 0 concn.
125.4 126.2 129.1 124 3
Materials Used (1) -4natural gasoline, designated as KO. 7 2 . The composition of this mixture was obtained by averaging the analyses of the original sample with the sums of the overhead vapors and corresponding residues from the vaporization tests. This method was used in order to reduce the errors in a single analysis. In most cases the differences in the analyses are within the experimental error of the analytical method used. Figure :3 gives the distillation curve of this mixture. ( 2 ) A similar mixture, designated as h-0. 75A. Analytical results and the distillation curve of this mixture are given in Figure 4. (3) A similar mixture containing the same compounds as 7SA but relatively less pentane, and identified as No. 48. Analytical data are shown in Figure 5 . (4) A synthetic mixture (48A) made by adding to sample 48 2oo
73 2 MOL% GASOLINE IN SAMPLE ---EXPERIMENTAL
Figure 11-Vaporization
of Oil-Gasoline Mixture
INDUSTRIAL A N D ENGINEERING OHEMISTRY
262
a sufficient amount of pentane to give a mixture with approximately the same percentage of pentanes as was found in sample 75A, as given in Table 111. The pentane sample used to prepare 48A had the following composition: Butane .............................................. Isopentane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . n-Pentane.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Mol per cent 2.82 31.20 66.10
1 -CALCULATED
Vol. 22, No. 3
T a b l e IV-Properties of T a b l e V-Properties of Absorotion Oil No. 1 N a o h t h a Samole Molecular weight 204 Moleeular weight 172.3 Specific gravitr 6Oo/6O0’F. (15.6’/ Specific gravity 6 0 ° / 6 d . F. (15.6’/ 15.6 C.), 0.8490 15.6’ C.), 0.7968 DATA A. S. T. M. DISTILLATION A. S. T. M. DISTILLATION DATA c. Per cent F. C. Per cent F. 232 Initial boiling point 285 141 5 8.16 255 ... 169 257 184 10 364 270 201 20 394 277 30 415 213 284 40 435 224 289 50 444 229 300 60 467 242 70 4xn 312 ._ 249 331 80 496 258 360 90 518 270 379 95 544 284 404 98 (end point) 562 294 Recovery, 98.5 cc. Recovery, 99 cc.
___
(9) A mixture of tetralin (tetrahydronaphthalene) prepared by the Eastman Kodak Company and natural gasoline No. 7 2 , containing 50 mol per cent of gasoline.
Experimental Results o 2
0
4 WEIGHT
Figure 12-Vaporization
CO>WOUND Butane Isopentane n-Pentane Isohexane n-Hexane Isoheptane n-Heptane Residue
.
6
% VAPORIZED
8
10
of Oil-Gasoline Mixture
Table 111-Preparation of S a m p l e 48A I N 491 Ih. 32.5 GRAMS GRAMS No. 4 8 PEITAIBS TOTAL ANALYSIS Mols Mols Mols Mol der cent 0.242 0,012 0.284 4.14 0,706 0.133 0.839 13.66 1.568 0,281 1.849 30.03 0.680 0.680 11.09 1,000 1.000 16.31 0.224 0.224 3.66 0.423 0,423 6.91 0.872 0.872 14.21
( 5 ) A mixture of natural gasoline (89.3 mol per cent) and midcontinent absorption oil (10.7 mol per cent). The oil is designated as No. 1 and its properties (stripped) are given in Table IV. This oil was received as a rich oil from the field with the following composition:
T‘aporizations were made on each of the samples at various temperatures, with results shown as dash lines in Figures 6 to 15, inclusive. Data for the vaporization of the individual hydrocarbons in sample 72 are given in Figure 15. The data taken in vaporization tests on the absorption oils enriched by the addition of this same natural gasoline indicate the same vaporization characteristics as possessed by the natural gasoline itself. Data on the vaporization of the individual hydrocarbons of the enriched absorption oil are not given because the integral vaporization curves (Figures 6, 10, 11, and 12) supply all the necessary information. 210
IW
3170
z < Y I I50
COMPOUND Mol per cent Ethane . . . . . . . . . . . . . . . . . . . 0.103 Propane.. . . . . . . . . . . . . . . . . 1.537 Isobutane., . . . . . . . . . . . . . . 1.56 n-Butane . . . . . . . . . . . . . . . . . 4.18 Isopentane.. . . . . . . . . . . . . . 2.63 *-Pentane.. . . . . . . . . . . . . . . 2.28
COMPOUND Mol per cent Isohexane ............ 0.78 n-Hexane . . . . . . . . . . . . 1.20 Isoheptane.. . . . . . . . . . 0.41 n-Heptane.. . . . . . . . . . 0.61 Residue . . . . . . . . . . . . . . 0.88 Absorption oil.. . . . . . . 83.82
(6) A similar mixture of absorption oil No. 1 containing 73.1 mol per cent of gasoline and having the following composition: COMPOUND Mol per cent Ethane.. . . . . . . . . . . . . . 0.032 Propane.. . . . . . . . . . . . . 2 . 7 5 3.07 Isobutane., . . . . . . . . . . ?$-Butane.,. . . . . . . . . . . 16.91 Total butanes.. . . . . . . . 19.98 Isopentane.. . . . . . . . . . 7.40 n-Pentane., . . . . . . . . . . 12.46 Total pentanes.. . . . . . . 19.86
COMPOUND Mol per cent Isohexane.. .......... 5.40 n-Hexane.. ........... 9.37 Total hexanes.. ...... 14.86 Isoheptane.. . . . . . . . . . 4.57 n-Heptane.. . . . . . . . . . 8.50 Total heptanes.. 13.07 Residue.. . . . . . . . . . . . . 3.22 Absorption oil.. . . . . . . 26.90
......
(7) A similar mixture of absorption oil No. 1 containing 36.3 mol per cent of gasoline and having the following composition: COMPOUND M o l Be7 cent Rthane.. . . . . . . . . . . . . . . . 0.077 Propane.. . . . . . . . . . . . . . . 1.98 Isohutane.. . . . . . . . . . . . . 2.11 n-Butane.. . . . . . . . . . . . . . 8.76 Total butanes.. ......... 10.87 Isopentane.. . . . . . . . . . . . 4.35 n-Pentane.. . . . . . . . . . . . . 5.95 Total pentanes.. . . . . . . . . 10.30
COMPOUND M o l per cent Isohexane.. . . . . . . . . . . 2.47 n-Hexane.. . . . . . . . . . . 4.13 Total hexanes.. . . . . . . 6.60 Isoheptane.. . . . . . . . . . 1.89 n-Heptane. . . . . . . . . . . 3.45 Total heptanes.. ...... 5.34 Residue.. . . . . . . . . . . . . 1.21 Absorption oil.. 63.65
......
(8) A mixture of a naphtha and natural gasoline containing 50 mol per cent of gasoline. The characteristics of the naphtha are given in Table V.
2
2 IM
P f B 110 w
0
IO WEIGHT
5
Figure 13-Vaporization
%
I5 VAPORIZED
20
25
of 50-50 Tetralin-Gasoline Mixture
Calculation of Vaporization Curves by Raoult’s Law
The method used to compute vaporization by Raoult’s law has been outlined (IS) and requires a series of trials before the correct value for the vaporization can be obtained. Usually three trials are sufficient to determine the correct temperature for any given vaporization. In making these calculations, vapor pressures were taken from the vaporpressure chart for hydrocarbons by Coats and Brown ( 3 ) . Figures 6 to 14, inclusive, give the calculated and experimental vaporization curves for all the mixtures studied. Figure 15 gives the vaporizations of the individual compounds as computed by Raoult’s law from the composition of the original mixture and as obtained experimentally. Figure 15 is plotted on a mol basis, whereas the other figures are on a weight basis. Discussion of Results
It was planned to determine the actual deviations from Raoult’s law for each of the compounds present in natural
INDUSTRIAL AND ENGINEERING CHEMISTRY
March, 1930
gasoline. The investigation was started, using mixture 72 as representative of commercial natural gasoline (grade BB). The plan was to vaporize this mixture and mixtures of this gasoline with absorption oils of various kinds, arid to study their vaporization characteristics, particularly in regard to the application of Raoult’s law. Vaporizations were made on these mixtures and in each vaporization analyses were made of the overhead vapors, and in many cases also of the liquid residues, in order that the actual amounts of each compound vaporized could be experimentally determined and compared with the values calculated from the original composition by means of Raoult’s law.
0
5
IO
WEIGHT
Figure 14-Vaporization
X
IS VAPORIZED
20
25
of Naphtha-Gasoline Mixture
The first observation was that the vaporization curves for the mixture 72 (Figure 6) indicated entirely different vaporization characteristics than had been previously observed for mixture 48. (Figure 8) The peculiar deviation in the case of mixture 48 led to the belief that the cornpounds more volatile than pentane exerted a greater partial pressure, and those less volatile a lower partial pressure, than would be calculated by Raoult’s law, as had been suggested by Calingaert and Hitchcock in their studies of binary mixtures. It will also be noted in the analysis of mixture 48 that no propane is present and only a small amount of butane, the mixture being principally pentane and heavier material. If this reasoning is correct it would seem, if the mixture having butane as its most volatile component (No. 48) were more volatile a t low vaporizations than would be calculated by Raoult’s law, that certainly a mixture with much larger amounts of butane and considerable propane (No. 72) ought to exhibit this same characteristic to a greater degree. However, this was not the case, as mixture 72 is slightly less volatile for all percentages vaporized than is calculated by Raoult’s law. N o satisfactory explanation could be offered for this difference without further information. I n the hope of obtaining this information another natural gasoline intermediate in volatility (KO. 75A) was vaporized and compared with the calculated vaporizations. (Figure 7 ) I n this case still different vaporization characteristics appeared. It will be seen that the experimental curve follows the calculated curve for practically 50 per cent of the sample vaporized. KO adequate explanation of these data could be offered on the basis of the considerations just discussed. When the analyses of the three mixtures were compared it was observed that mixture 48 contained 43.7 per cent of pentane and butane, mixture 758,50.0per cent of pentane and butane, and mixture 72’56.5 per cent of propane, butane, and pentane. It was also observed from the data on the vaporization of these mixtures that mixture 48 was more volatile a t low vaporizations than was calculated, mixture 75A checked
263
the calculated values, and mixture 72 was somewhat less volatile than would be calculated from Raoult’s law. From empirical considerations it seemed that the characteristics of mixture 48 might be so modified by the addition of the proper amount of pentane as to show the same deviations as mixture 7SA or 72. Accordingly, mixture 48A was prepared by adding to mixture 48 sufficient pentane to give a mixture having about the same per cent of pentane as 75A. This mixture was studied in two runs, which are compared with the calculated vaporization in Figure 9. As was expected, the vaporization characteristics of mixture 48A corresponded to those of 75A for the lower amounts vaporized. These data show that, in so far as Raoult’s law applies to complex mixtures, the properties of the mixture are largely dependent upon its composition. For this reason the vapor pressure of natural gasolines and similar mixtures cannot be calculated with precision or confidence by Raoult’s law. Figure 15 gives one set of curves showing the comparison between the actual amounts of the individual compounds vaporized and as calculated by Raoult’s law. I n the case of mixture 7 2 each compound shorn approximately the same deviation (Figure 15), as is also indicated by the totalvaporization curve. (Figure 6) Consideration of these two curves, as well as other similar tests, indicates that the deviations can be estimated more readily from the over-all vaporization curve. In the study of gasoline mixed with absorption oil S o . 1 vaporizations ivere made a t different temperatures and concentrations of gasoline jn oil. The individual compounds
exhibited the same vaporization characteristics when determined by actual analysis as are indicated by the integral vaporization curves for these mixtures. (Figures 6, 10, 11, and 12) These figures indicate that for this particular mixture of gasoline and absorption oil Raoult’s law may be used with good precision up to about 50 per cent of the gasoline vaporized. Figure 13 shows the vaporization calculated by Raoult’s law and, as obtained experimentally, of the gasoline-tetralin mixture. This figure indicates that Raoult’s law may be applied to tetralin-gasoline mixtures with about the same
264
INDUSTRIAL AND ENGINEERING CHEMISTRY
degree of accuracy as obtainable when applied to the ordinary run of absorption oils or naphthas. Indications of inconsistent deviations from Raoult’s law in previous investigations can be explained by the fact that the composition of the original sample determines the vaporization characteristics. In most cases the number of different samples used was insufficient to bring out this point and conclusions were drawn on results obtained from similar mixtures. Conclusions Deviations from Raoult’s law as applied to complex mixtures are dependent upon the relative composition of a mixture Of substances as as upon the or dissimilarity of the components in the mixture. For engineering calculations, when an accuracy of j to 15 per cent is considered satisfactory, Raoult’s law may be used with fairly satisfactory results. When greater accuracy is desired, little confidence can be placed in results calculated by means of this law.
VOl. 22, No. 3
Literature Cited (1) Brown and Caine, Trans. A m . Inst. Chem. Eng., P i , 21 (1928). (2) Calingaert and Hitchcock, J . A m . Chem. Soc., 49, 750 (1927). (3) Coats and Brown, Dept. of Engineering Research, University of Mich., Circ. P (December, 1928). (4) Davis, Proc. 5th Annual Convention Assocn. Natural Gasoline Mfrs., p. 37 (1926). ( 5 ) Findlay, “Practical Physical Chemistry,” Chapt. VII, p. 112, Longmans. (6) Kallam and Coulthurst, Oil Gas J . , 28, 50 (November 21, 1929). (7) Leslie and Good, IND.ENQ.CHEM.,19, 453 (1927). ( 8 ) Maxson, Proc. 4th Annual Convention Assocn. Natural Gasoline Mfrs., p. 25 (1925). (9) McLouth, Null. Petroleum X e u s , 20, No, 32, 54 (1928). Peters, IND, CHEM,,18, 6g (1926), (11) Piroomov and Beiswenger, Am. Petroleum Inst., Bull. 10, No. 2, 52 (January 3, 1929). (12) Podbielniak, Relrner A’atural Gasoline Mfr., 8, NO. 3, 55 (1929); oil Gas J . , 27, No. 35, 38 (January 17, 1929). (13) Podbielniak and Brown, IND.END.CHRM.,21, 7 i 3 (1929). (14) Weissenberger, Pelroleum, 1817 (1925). (15) wilson and ~ ~ ~ i E M .~,16,947 , (1923). cH (16) Wilson and IYylde, I b i d . , 16, 801 (1923).
The Cooking Process I-Role of Water in the Cooking of Wood1 S. I. Aronovsky2 and Ross Aiken Gortner
The cooking of wood as practiced in the industry for the production of cellulose involves the use of cooking liquors containing various chemicals in various concentrations. The exact effect of the individual constituents upon the cooking process is largely unknown. This paper reports the first of a projected series of studies designed to ascertain t h e role of each constituent of the cooking liquor. Since water comprises the main bulk of the liquor, cooking with water only was selected for the first study. I t is shown t h a t cooking with water a t different temperatures for varying lengths of time has a profound effect upon the various constituents of the wood. Pentoses and
pentosans are rapidly destroyed, resulting in the production of appreciable quantities of furfural; lignin, although apparently the most stable constituent, undergoes partial depolymerization; and the celluloses are broken down to water-soluble constituents and to gaseous products. Approximately 37 per cent of the total celluloses and 46 per cent of the alpha-cellulose were destroyed in the 12-hour (186” C.) cooks. A t the longer times and higher temperatures the rate of destruction of the alpha-cellulose was faster than was t h a t of the total celluloses, indicating a conversion of alpha-cellulose to hydrocelluloses. Water can, therefore, be looked upon as avery active reagent in the cooking process.
HE cooking process in the pulp industry is based more upon practical experience than upon the knowledge of the chemical reactions taking place in the digester. This is due partly to the limited knowledge of the chemical composition of the woods used, and partly to the lack of exact data on the action of the various cooking chemicals on the components of the wood. Empirical studies on the pulping of wood have been, and are being, carried out in various industrial and technological laboratories. Such studies, however, as a rule involve the use of cooking liquors which contain several different cheniicals in various concentrations. The composition of these
cooking liquors has, in general, been arrived a t in part by tradition, and in part by random experimentation, one composition after another having been used or varied until more or less satisfactory results were obtained. There is thus available little or no information as to the exact role which any particular component of the cooking liquor plays in the cooking process. The problem of the nature of the cooking process is further complicated by the fact that wood is a biological product of more or less uncertain and variable composition, and accordingly the various constituents of the cooking liquor probably react differently toward the various organic compounds present in the wood. Following this line of thought, it seemed desirable that a series of fundamental investigations of the cooking process should be carried out by starting with a simple substance as the cooking agent and then adding other reagents, one by one, until the composition of the cooking liquors in the present-day commercial processes was reached. This series of investigations should lead to a better understanding of the reactions taking place in the digester, and might
T
Published with the approval of the 1 Received November 11, 1929. Director a s Paper No. 900, Journal Series, Minnesota Agricultural Experiment Station. Presented under the title “The Chemistry of the Cooking Process. I-Cooking Aspen with Water” before the Division of Cellulose Chemistry a t the 78th Meeting of the American Chemical Society, Minneapolis, Minn., September 9 t o 13, 1929. * Cloquet Wood Products Fellow, University of Minnesota. This study was conducted under an industrial fellowship grant from the Cloquet wood products companies of Cloquet, Minn. Thanks are especially due to various representatives of the Northwest Paper Company, who have cooperated in this study.
