Hydrolysis of Propylene-Sulfuric Acid Solutions to Isopropyl Alcohol ROBERT W. SCHHAGE AND E R W N €1. AJIICK, JR. Columbia University, New York, T h e effect of temperature and composition on the per cent of total propylene as isopropyl sulfate i n propylenesulfuric acid solutions was studied. I t was found t h a t for acid concentrations u p to 7070 (on a propylene-free basis) less than 257' of the propylene present was esterified a t all temperatures up to the normal boiling point, and t h a t less than 35570 was esterified for even 80% acid. The effects of acid concentration, propylene-sulfuric acid mole ratio, and operating pressure on the yields of products batchdistilled from propylene-sulfuric acid solutions were also investigated. By reducing the pressure to about 250 mm. absolute during t h e distillation of 61.5% acid, yields were realized which were equivalent to those from 46.0oJo acid distilled a t atmospheric pressure under otherwise similar conditions. Although yields of isopropyl alcohol Crom 69.5% acid can be improved by reducing the pressure during distillation, i t does not appear t h a t they can, a t moderate pressure reductions, he made equivalent to those obtained from 46.0% acid.
N. Y.
it was felt that a suitable choice oi o p e m t h g variables during the hydrolysis and distillation step might avoid the necessity of diluting the propylene-sulfuric acid solution to the extent ordinarily practiced in industry, without forming excessive amounta of regenerated propylene, isopropyl ether, and degradation products. Thus a large part of the expensive and difficult acid concentration step would be avoided. The present research waa a preliminary investigation of this idea.
A
REVIEW of the literature on the production of alcohols
from olefins indicates that two types of processos present the best commercial possibilities. The newer one involves direct catalytic hydration of the olefin at fairly high temperatures and pressures. Although this is superficially perhaps the more attractive process, it presents sufficient operating difficulties to limit its application to specific alcohols, for which some o f the disadvantages of the second and older process are particularly clear. In this older process the olefin is absorbed in sulfuric acid of suitable cmcentration, forming an olefin-sulfuric acid solution which is then diluted and distilled to remove dcohol, leaving behind a n essentially organic-free dilute sulfuric acid. In order t o absorb the olefin again this dilute acid must be concentrated an operation which is a major item in the cost of manufacturing alcohol by this method. In making isopropyl alcohol (2-propanol) by the older process
0
Figure 2. Per Cent of Total Propylene as Isopropyl Sulfate in Propylene-Sulfuric .4cid Solutions
TIIEORKI'ICA L CON SIDERATION S 7 7
1hc possible react,ions of propylenct, sulfuric acid, and w a t e r :%renumerous and there has been :t gencral tendency in erigineei,ing literature to underestirnatc thtair complesity. Alcohol protluction is usually p o r t r a y d as consisting of t.hr two well tleficied steps shpwn in Equatioris 1 :tnd 2:
-+ I I M , =zz iso-PrHSOr Iso-PrHBOc + ILO V iso-PrOH + H&Wd C3€Ic
1.100
'
Figure 1.
(1)
(2)
T h e first reaction forming isopropyl sulfate iY probably rexponsible for the initial solution o f propylene in sulfuric acid. Formation of diisopropyl sulfate t1ot:s not appear to tw possit~le in t,he presence of any appreciable amount of water (6,8). Equation 2 is a simplification of a sequence of somewhat uncertttin ionic reactions. The possibility that alcohols form comples ions in the presencc of acids was originally suggested by Goldschmidt (4, 6) a s a result Of euterificittion r:Lte studiKq, and some o f the practical implications of this idea have been noted by Ilobey ( 7 ' ) . €Iydrolysis according to this throry.might be said to take place. in thc following st,eps: I
0.2
I
I
0.4 0.6 R . MOL RATIO,
1
I
+ 1120 iso-Pr80,- + H,O+ Iso-PrS04- + IT30+F= iso-PrOII?+ + €ISOITso-PrOI12 + 1320 i iso-PrOH + €LO
I
t
0.8 1.0 12 PROPYLENE/SULFURIC
Iso-PrHS04
I.4
ACID
Specific Gravities of PropyleneSulfuric Acid Solutions
+
2550
z=?r
+
(3) (4) (5)
INDUSTRIAL AND ENGINEERING CHEMISTRY
December 1950
I70
2551
t
sulfuric acid solutions prepared from either olefin or alcohol. Although the amount of free ant1 chemically combined water added with the alcohol or acid limits the concentration of the solutions that can be prepared by this method, this was found to be of little consequence o w r the range of interest in this study. Specific gravities of propylenc-sulfuric acid solutions prepared in this manner are shown in Figure 1. Data on the variation of the density of these solutions with temperature were also secured, but are not included in this paper. Figure 1 can be used to determine the propylene-sulfui io acid mole ratio of a solution if its specific gravity a n d :wid concentration ou a propvlene-free basis are known.
loot oc
02
04
Figure 3.
oe
06 R,
MOL
RATIO,
12 14 SULFURIC A C I D
IO PROPYLENE/
16
PER. CENT PROPYLENE A S ISOPROPYL SULFATE
Boiling Points of PropyleneSulfuric Acid Solutions at Atmospheric Pressure
The name “alcoxonium” ion has btun given to the alcohol cornplex because of its analogy to the oxonium ion. Because it was apparently difficult to solvcnt-extract alcohol from butenesulfuric acid solutions in which less than half of the olefin could be accounted for as ester, Robey concluded that a large part of the alcohol presumably present W R S in the form of alcoxonium ions. The preferential formation of isopropyl ether rather than alcohol is favored by greater acid concentrations. Equation 6 is the reaction usually given, but it again is probably the result of vnrious intermediate ionic reactions. Iso-PrOH
+ iso-PrHSO,
iso-Pr2O
+ HaSO,
(6)
I1 propylene is present in acids of still greater concentrations (particularly if warm), there is a very noticeable tendency for irwversible reactions to occur with the formation of polymer, char, and other degradatioa products. In general, both inciustrially and in this investigation, conditions causing such reactions have been avoided, but even in acid of moderate strength they occur to some extent. PREPARATION O F SOLUTIONS
If it is assumed that the componentR of a propylene-sulfuric arid solution are in equilibrium with etLch other, then a t any given temperature and pressure the composition of the solution may be completely defined in tcrms of the relative amounts and compositions of the materials used in forming it. As a matter of convenience in the laboratory, solutions equivalent t o those obtained by absorbing propylene ill sulfuric acid may be prepared by mixing appropriate amounts of isopropyl alcohol, water, and sulfuric acid. Provided no irreversible reactions occur, in approaching equilibrium from either direction the final result should be the same. Robey ( 7 ) found no differewe in butane-
I eo0
Solutions prepared by the method just discussed were held a t a desired temperature in an oil bath until equilibrium was reached. A 3- to $gram sample was then removed from the solution, and within about 1 second or less it was diluted in 20 grams of water kept in a cooling bath a t 5 ” to 10”C. Hates of hydrolysis of the solutions are negligible at these temperatures ( 1 ). The diluted sample was titrated ~ 6 t h1N sodium hydroxide a t room temperature or less, using phenolphthalein as the indicator. When the end point was reached it did not fade on standing, indicating that no significant hydrolysis was taking place, and thus supporting the observations of ot3hersreported in the literature.
If the weight of the sample and its composition (in terms of sulfuric acid concentration on a propylene-free bagis and the propylene-sulfuric acid mole ratio) are known, the total sulfuric acid and propylene present may readily be calculated. If the total moles of hydrogen ion in the mniple, as determined by titration, are subtracted from the equivalents of the total sulfuric acid, the difference will be the moles of sulfuric acid present as isopropyl sulfate. This @ turn is equal to the moles of propylene as isopropyl sulfate, from which the per cent propylene as isopropyl sulfate can be directly computed. From each prepared propylene-sulfuric acid solution, samples were taken a t five or six temperatures up to the normal boiling point. When the results were plotted as a function of temperature no distinct trend could be seen. Therefore, all analyses for a given solution Composition were averaged and the per cent propylene as isopropyl sulfate was assumed substantially independent of temperature. This procedure gave results with a calculated probable error which was in most cases less than * 1.0% and in no case greater than * 1.7%. Figure 2 shows the variation in the per cent propylene as isopropyl sulfate with the concentration of the sulfuric acid. The greatest number of determinations were made at a propylencsulfuric acid mole ratio of unity. At higher and lower mole
1
I I
I
I
100 400 boo coo TOO ABSOLUTE PRESSURE, UlLLlYEftRS ff YERCURl
I
eo0
R, MOL RATIO, PROPYLENE/
14 SULFURIC ACID
Figure 4. Variation of F ~ i l i n g Point of PropyleneSulfuric Acid Solution w i t h Pressure
Figure 5. Variation of Composition of Distillate with Propylene-Sulfuric Acid Mole Ratio
61.5% aoid, 1.06 propylenoaolfuric acid mole ratio
61.570 acid at atmospheric prcasure
INDUSTRIAL AND ENGINEERING CHEMISTRY
2552
ISOPROPYL ALCOHU.
0
0
6
K
d.2
010
016 d.6 IlO 112 R. MOL R A T I O , PROPYLENE/SULFMIC
d.4
L Figure 6.
114
ACID
’
Variation of Per Cent Yield of Products with PropyleneSulfuric Acid Mole Ratio 61.5%acid at atmospherio pressure
a
I
401
zoo
300
400
so0
600
amrmi VITF P P c ~ s i i R ~U~LLIWF-COS . OF
Figure 7.
700
Vol. 42, No. 12
Uncondensed gases from the distillation, presumably only propylene and the initial air content of the apparatus, were passed successively through Drierite (anhydrous calcium sulfate), silica gel, and 96y0 sulfuric acid catalyzed with 1.0% silver sulfate. The Drierite was used t o remove any residual water vapor from the gases and the propylene was recovered, principally on the silica gel. Propylene on the gel was determined by weighing it (‘wet” and after desorption b y red heat. Because of the varying partial pressures of the propylene in the different runs, and an indeterminate eficiency factor involved in contacting the gases with the various absorbants, the complete recovery of propylene is somewhat uncertain. I n all distillations the initial composition of the propylenesulfuric acid solution was known, because i t was prepared by the method previously given. From the yields of the various products the final composition could be calculated. A linear interpolation between the two gave an average compdsition used in presenting the data. Fortunately, although the distribution of products varied considerably over the runs made, the amounts of total propylme and water (in all forms) removed in each run were about the same. For example, for the data presented as applying to 61.501, acid, in all runs the initial acid concentration was 60.0% and the final about 63.0%. Similarly for the runs at varying pressure, the average mole ratio of propylene to sulfuric acid during the runs was in all cases 1.06 * 0.02. The data, therefore, approximate the yields of products. which might be expected in distilling a differential amount of solution of specific composition, although the data were actually secured over a small, but finite, interval in a range of compositions.
mo
USPC~IW
.--I
Variation of Composition of Distillate with Pressure
61.5% acid, 1.06 propylenegulfuric acid mole ratio
r h o s there seems to be some divergence from this curve. There was some indication in further results, discussed later, that this divergence might be of some significance. DISTILLATION TEMPERATURES
The operating temperatures for distillation of propylenesulfuric acid solutions are a function of their composition and the pressure. Figures 3 and 4 show the effect of both these variables. The temperatures given were measured in the solutions.
ABSOLUTE PRESSURE. MILLIMETERS O F
Figure 8. YIELDS OF PRODUCTS UNDER VARYING CONDITIONS
In order to determine the yields in which the various components of propylene-sulfuric acid solutions can be removed from the mixture under varying conditions of composition and pressure, simple batch distillations a t approximately equal rates were made in which only a small percentage ( 5 t o 7% by weight) of the solution was distilled. .41cohol, ether, and water were condensed and the distillate was wcighcd and analyzed b y a method developed by Fnzia ( 2 ) . This involves titration of the sample with water or isopropyl ether until a cloud point indicates the formation of a second phase. The specific gravity of the sample a t this point is then determined and, when compared with data given in the reference, it indicates a distinct composition for the ternary. The composition of the original sample can be computed from the amounts of isopropyl ether and/or water added during the titration. Fazia’s unpublished data on the ternary were checked with the recent data of Frere ( 3 )and found t o be in good agreement from the viewpoici of the present research-that is, use of either set of dnia would result in no significant difference in any of the following results.
MERCURY
Variation of Per Cent Yield of Products with Pressure
61.5 9% acid, 1.06 propylene-sulfuric acid mole ratin
Figures 5 and 6 show the effects of the propylene-sulfuric acid mole ratio on the yields of products from 61.5% acid (on a propylene-free basis) distilled at atmospheric pressure. The per cent yield based on propylene, which is shown in Figure 6, is based on all thc propylene in any form which is removed from the solution. The 61.5y0 acid was chosen because its concentration is great enough to cause considerable by-product formation, making the effect of the variables on the yields clearly evident. It was pointed out in discussing the pcr cent propylene as isopropyl sulfate that a t high and low R values (propylene-sulfuric acid mole ratios) the per cent was evidently lower than a t intermediate values. This may have some bearing on the fact that isopropyl ether appears in maximum yields a t intermediate values of R, since ether is formed from isopropyl sulfate. The facts that both operating temperatures (Figure 3 ) and the amount of water available for hydrolysis decrease with increasing R, also affect yields.
December 1950
INDUSTRIAL A N D ENGINEERING CHEMISTRY
presented in Table I. These data were obtained in the same manner as the results in Figures 5 and 6. The effects of pressure on yields from 61.5% acid a t a propylenesulfuric acid mole ratio of 1.06 are shown in Figures 7 and 8. These demonstrate t h a t by sufficiently reducing the pressure, yields of alcohol can be obtained from 61.5y0 acid which are comparable t o the yields a t atmospheric pressure from 46.0oJo acid with the same R value. It was of interest t o determine yields at still higher acid concentrations, approaching the acid used for absorption of the olefm in strength. These data are presented in Table I1 for 69.5’% acid, a t both atmospheric and reduced pressure. For this case, although yields of isopropyl alcohol may be considerably improved by lowering the pressure, it appears unlikely t h a t they cftn be made comparable t o those from more dilute acid a t any economically feasible vacuum.
TABLE I. COMPOSITION OF DISTILLATE AND YIELDS OF PRODUCTS (46.Oye acid, 1.05 propylene-sulfuric acid mole ratio, distilled a t
atmosphenc preasure)
Weight % Compoeition of Distillate 77.9 Isopropyl alcohol IBopropyl ether 2.0 Water 20.1
?+, Yield Based on Propylene Isopropyl alcohol 95.0 Isopropyl ether 2.9 2.1 Propylene
TABLE 11. COMPOSITION OF DISTILLATE AND YIELDS OF PBODUCTS (69.5% acid, 1.06 propylene-sulfuric acid mole ratio)
Weight % Composition of Distillate 1 atm. 271 mm. Iaopropyl alcohol 34.8 56.6 Isopropyl ether 41.2 27.0 24.0 16.4 Water
% Yield Based on Propylene Isopropyl alcohol 26.5 Isopropyl ether 36.8 Propylene 36.7
2553
LITERATURE CITED
49.6 27.9 22.5
The yield of alcohol from 61.5% acid a t any R value is much poorer a t atmospheric pressure than is usually obtained in commercial processes using more dilute acid. For comparative purposes data obtained for 46.0% acid with a propylene-sulfuric acid mole ratio of 1.05 and distilled a t atmospheric pressure are
(1) Evans, P.N.,snd Albertson, J. M., J . Am. Chem. Soc., 39,456-61 (1917). (2) Faaia, A.,M.S. thesis, Colbmbia University, 1948. (3) Frere, F. J., IND.ENO.CHEM.,41,2365-7 (1949). (4) Goldschmidt, H., Haaland, H., and Melbye, R. S., 2. physik. Chem., B143,278-86 (1929). (5) Goldschmidt, H., and Melbye, R. S., Zbid., B143,139-56 (1929). (6) Medwedew, S.S.,and Alexejewa, E. N., Bm., 65B,131-3 (1932). (7) Robey, R.F.,IND.ENG.CEEM.,33,1076-8 (1941). (8) Suter, C. M., and Oberg, E., J . Am. Chem. Soc., 56,677-9 (1934). RECEIVED April 14, 1950. From a thesis submitted by Robert W. Schrage in partial fulfillment of the requirements for the degree of master of science in the School of Engineering, Columbia University. Contribution No. 5 from the Chemical Engineering Laboratories, Engineering Center, Columbia University, New York, N. Y.
Aging Studies on Redox Polymers EFFECT OF IRON C. R. PARKS, J. 0. COLE, AND J. D. D’IANNI The Goodyear Tire 6% Rubber Company, Akron, Ohio T h e stability of GR-S type polymers and vulcanizates prepared by the iron-redox and mutual recipes have been compared by oxygen absorption measurements during accelerated aging. Iron-redox polymers are less stable in general toward oxidation than polymers made by the mutual system. Vulcanizates derived from redox polymers are only very slightly inferior in stability, however, to vulcanizates prepared from mutual type polymers. Evidence shows that soluble iron is a factor in the poor stability of some iron-redox polymers toward oxygen,
sBLE
‘On has a pronounced effect On the aging GR-S p o l p e r but Only a ‘light effect On the
Of
the
1’
Polymerization recipes currently employed for the commercial production of “cold rubber” contain iron in the form of the ferrous pyrophosphate complex. In view of the possible adverse effect of iron on the aging of GR-S type polymers, a comparison was made of the aging of GR-S prepared by the ironmredoxand by the mutual polymerization systems.
~~h~of GRS polymers and vulcanizates was evaluated by oxygen absorption measurements and by the change in physical properties during accelerated aging. In view of the uncertainty as t o the chemical form in which iron is present in iron-redox polymers after coagulation of the latex, the activity of iron in various chemical forms was measured. The amount of soluble iron t h a t
would catalyze decomposition of the polymer was also determined. EXPERIMENTAL
In the discussion which follows, a mutual polymer ( 4 ) refers t o a GR-S type emulsion polymer prepared with potassium persulfate and mercaptan (thiol). An iron-redox polymer refers t o a polymer prepared with a cumene hydroperoxide, ferrous pyrophosphate, and sugar system (6). Oxygen absorption measurements were made in oxygen by the volumetric method employing apparatus similar t o t h a t described by Shelton and Winn ( 8 ) . All pilot plant and laboratory prepared samples contained phenyl-p-naphthylamine 8s antiofidant. Polymers and vulcanizates were compared a t the same antioxidant content. Polymers were analyzed for phenyl-p-naphthylamine by the ultraviolet spectrometric method and the antioxidant content of each sample in a given’ series adjusted t o a predetermined value (1.25% wherever possible). The commercial samples were obtained from various Rubber Reserve plants and contained B-L-E (a high temperature reaction product of diphenylamine and acetone) as an antioxidant. The antioxidant contents were not adjusted as they were within specification limits (1.25 * 0.25%). The polymer samples were prepared by coating a test tube with
