R EACT IO N 0 F PHOS PHATE-CO M P LEX ED
SULFUR TRIOXIDE WITH ALPHA-OLEFINS A L B I N
F. T U R B A K ' A N D
R. L I V I N G S T O N , J R .
ESSO Research and Engineering Co., Linden, N . J . High molecular weight a-olefins can b e reacted directly with SO3 a t temperatures up to 40" C. if the SOsis first complexed with an alkyl phosphate. Reactions of this type usually give both hydroxy- and alkenesulfonates. However, b y use of the phosphate complexing agent and proper control of conditions, it is now possible to obtain essentially complete conversion of olefin to alkenesulfonate. The phosphate is extremely effective in causing the olefin to react with a second mole of SOs. The phosphate complexing agent can b e recovered in 90 to 93% yields by simple filtration. Procedures are described for separating the product into hydroxy- and alkenesulfonate fractions. The effect of reacticn conditions on product distribution is discussed, and the influence of phosphate is compared with other complexing agents.
HE EXTREMELY VIGOROUS REACTIVITY of so3 makes it Tnecessary to reduce its activity by complexing prior to reaction with organic compounds-particularly with olefins. Many complexing agents have been reported lvhich moderate SO3 activity to varying degrees. The authors have found that complexes of SO3 with trialkyl phosphates are particularly effective for controlling so3 activity. The SOs-phosphate complex reacts lvith olefins at room temperature to give essential1)- quantitative yields of alkenesulfonates. .4t low temperatures, hydroxysulfonates are also prepared. ?'his report describes the reaction of trialkyl phosphatecomplexed so3 with high molecular weight a-olefins, the techniques used for recovery, separation and identification of the products, and the effect of reaction conditions and other complexing agents on product yield and distribution.
but is a strong indication that the phosphoryl oxygen is involved in the complex formation. The carbon-oxygenphosphorus bond absorbs a t 8.6 microns and a t 9.7 microns in both the complexed and uncomplexed triethyl phosphate. This further substantiates the hypothesis that the phosphoryl oxygen is involved in complex formation. The sulfur-oxygen stretching band is shifted from 7.28 microns in the uncomplexed SO3 to the 7.5-micron region in the complex. Sulfonation of Olefins
I n these studies it was found that SO,. complexed in a 1 to 1 ratio with tributyl phosphate, reacts smoothly with high molecular weight a-olefins a t low (0"to 5" C.) temperatures and short reaction timrs to give hydroxysulfonatas and a t moderate temperatures (20' to 40" C.) to gi\e unsaturated
Many reagents have been reported which complex SO3 ( 7 0 ) . Among the most commonly used are dioxane and bis(p-chloroethyl) ether, both of which form medium strength complexes that sulfonate aromatics a t 0" C. (73). Pyridine, on the other hand: forms a tight complex \vhich will not sulfonate benzene even a t reflux temperature (9). Although this wide range of reactivity is available, it is evident that a different complexing agent must be employed for each degree of desired activity. It was reported previously ( 7 4 that S o 3 activity can be substantially reduced and controlled over a \vide range by complexing lvith trialkvl phosphates. For instance, a complex formed from equivalent amounts of SO3 and trialkyl phosphate reacts readily with olefins and compounds containing active hydrogen but will not react with benzene even a t 75" C. Higher ratios of so3 to trialkyl phosphate are sufficiently active to sulfonate aromatics. In all cases, the phosphate complexes are completely soluble in the organic diluents. \vhereas the other agents give complexes which usually precipitate from the diluents and are consequently difficult to handle. Nature of the Complex. The infrared spectra of SO3. triethyl p!iosphate, and the 1 to 1 complex in dichloroethane solvent are given in Figure 1. T h e band at 7.95 micro7s in triethyl phosphate, which is ascribed to the phosphoryl linkage, is shifted to the 7.5-micron region in the complex. This pronounced shift of the phosphoryl band to the left is very unusual
address, Tee-Pak, Inc., D a n d l e , 111.
Figure 1 . Infrared spectra of phosphate complex and components VOL. 2
MIXED HYDROXY AND ALKENE SULFONATES DISSOLVE 40 g . / l L I T E R 50":. 2 - P R O P A N O L
FILTRATE-SATURATE WITH ANHYD. NaTCO-, @ 55'C. TO GET TWO LAYERS
UPPER LAYER83:: 2 - P R O P A N O L SOLN.
LOWER LAYERSAT. Na2C03 SOLN.
I C O O L TO 25'C.
SULFATO-SULFONATE M A Y O R M A Y NOT B E PRESENT
HYDROXY S U L F O N A T E
ALKENE SULFONATE IN SOLN. E V A P . TO DRYNESS TO GET PRODUCT
l , 1 , 1 , 1 / 1 1 1 1 1 1 1 1 1 1 1 I I i
Modification of ASTM Method 855-56
sulfonates. Because of the favorable reaction temperatures. extensive cooling equipment is not needed. Reaction Conditions and Product Isolation. In general. the complex was prepared by the slow addition of distilled SO3 to a cooled solution of a n equimolar amount of tributyl phosphate dissolved in 1,2-dichIoroethane. The complex was then adjusted to the desired temperature and the olefin streamed in rapidly. The reaction was quenched a t the desired time by the addition of 5070 S a O H solution to the vigorously stirred mixture. The product and excess sodium sulfate precipitated and were isolated by filtration. T h e filtrate from the reaction was flashed free of dichloroethane and analyzed for phosphate and olefin content by gas chromatography. In all cases, the phosphate recovery was 90 to 93%. The product was separated into fractions by a modification of ASTM Method 855-56 ( 7 ) . the details of which are given in Figure 2. The fractions could be further purified by recrystallization from ~ a t e r . The first fraction isolated (Fraction A) was identified as a hydroxysulfonate and the second fraction (Fraction B) as an unsaturated sulfonate. Identitication of Products. When C12-18olefins were sulfonated, Fraction A was essentially insoluble in water and was identified as a sodium 2-hydroxy-1-sulfonate. The product obtained from 1-hexadecene sulfonation was identical in all respects to a product prepared by the reaction of 1,2-epoxyhexadecane with NaHS03. The infrared spectra, water solubilities, crystal structure, and melting characteristics of both samples were identical. Both samples also had the correct elemental analyses for a hydroxysulfonate. The insolubility of this fraction made it difficult to obtain a good nuclear magnetic resonance (SMR) spectrum, even when converted to the corresponding free sulfonic acid. Fraction B from 1-hexadecene was identified as sodium 4ltrans-1-hexadecenesulfonate. The unsaturation in the product was established by titration with bromine according to the method of Lucas and Pressman (72). An NMR spectrum of the product (as its corresponding free acid) was interpreted as being consistent with a trans-double bond a to a 1-sulfonic acid. The infrared spectrum was also consistent with this assignment of structure. The infrared spectra of the alkenesulfonate and hydroxysulfonate derived from l-hexadecene are given in Figure 3. The hydroxysulfonate has the characteristic (0-H) stretching band in the 3-micron region. This band is absent in the 230
I&EC P R O D U C T RESEARCH A N D D E V E L O P M E N T
1 0 1 1 1 2 1 3 1 4
Infrared spectra of products
alkenesulfonate. Conversely, the alkenesulfonate has a characteristic type I1 olefin band a t 10.4 microns which is absent in the hydroxysulfonate. The slight displacement of the ( S - 0 ) band a t 9.4 microns to longer nave lengths in the hydroxysulfonate is probably due to the hydroxyl group hydrogen bonding \vith the sulfonate group. This bonding could be one of the factors responsible for the almost complete insolubility of the hydroxysulfonate in contrast to the high water solubility of alkenesulfonates. Reaction Variables
The effect of complexing agent on product distribution was studied by sulfonating 1-hexadecene with a 1 to 1 to 1 molar ratio of SOa, complexing agent, and 1-hexadecene. The complex was prepared at 0 ' to 5' C. in dichloroethane and the olefin in dichloroethane added a t 0' to 5' C. The mixture was allowed to warm to room temperature (about 1 hour) and then was neutralized. Table I gives the results of the study. These data indicate that butyl phosphate complexing tends to favor alkenesulfonate formation compared Lvith the other
Distribution of Products as a Function of Complexing Agent Hydroxysulfonate, dlkenesulfonatr, Complexing Agent % ?i 81 19 (Bu0)aPsO
(ClCH?CH?)20 O(CH?CH?)rO Table II. Mole Ratio, SOs/
Effect of Time, Temperature, and SO3 Ratio on Sulfonate Yield and Distribution Sulfonate Product Total Distribution, Temp., Tim?, Yield. Insol. Sol. C. Hr. Mole 7c hydroxy- alkene91 51 49 0 0.5 95 0 100 0 96 25 0.5 95 32 68 25 4 100 0 100 0.5 100 0 100 40 0 0 5 60 57 43 0 96 50 0 100 0.5 57 41 59 25 25 4 54 0 100
25”-30”C. VIGOROUS AGITATION
\ PREC. OF ALKENE SULFONATE AND
(RO)3PF0 CH LOR IN A T E 0 SOLVENT
Figure 6. DRYING BED FOR FILTERED LIQUOR
Figure 4. Flow diagram for production of alkenesulfonates via S03-phosphatecomplex
BUTYL PHOSPHATE. k3
Figure 5. Comparison of phosphate complexes and dioxane complexes of SO3 in formation of carbyl sulfate
agents. Since, to the best of our knowledge, no one had yet produced alkenesulfonates exclusively by this type of reaction, it was important to determine if reaction conditions could be so adjusted with the phosphate complex to give essentially all alkenesulfonate. Thus, a study was made of the effect of time, temperature, and mole ratio of reagents. The results of this study are given in Table 11. When equimolar amounts of SO3 and olefin were reacted, between 40 to 50 mole yo of unreacted olefin was recovered regardless of reaction time and temperature. When a 2 to 1 ratio of SO3 to olefin was used, all of the olefin was sulfonated. These data indicate that alkenesulfonates can now be produced to the virtual exclusion of hydroxysulfonates by the use of phosphate complexing and long reaction times or high temperatures. A flow diagram for the entire process is given i n Figure 4. The precipitate containing alkenesulfonate and a n equimolar amount of sodium sulfonate can be further purified, if desired, by recrystallization from water. Mechanism
The reaction of dioxane-complexed so3 with lower molecular weight linear olefins has been thoroughly studied by Bordwell e t ~ l (2-3, . 5-8). A comparison of their proposal and of the effect of phosphate on the addition of S O 3 to olefins to give carbyl sulfates is given in Figure 5. With dioxane as complexing agent, the first mole of SO3 adds readily to the olefin to give a sultone(1). The sultone then reacts slowly with a second mole of SO3 to give the carbyl sulfate(II1). I n the presence of a sixfold excess of dioxane, the sultone has a more pronounced tendency to react with the second mole of so3 ( 4 ) .
Effect of phosphate on alkenesulfonate formation
Apparently, the presence of the excess dioxane favors ionization of the carbon-oxygen bond (77), and the second mole of so3 may then add to the incipient carbonium ion(I1). Butyl phosphate (e = 8.0 at 25’ C.) is a much more polar complexing agent than dioxane (e = 2.2 at 25’ C.). It should favor carbonium ion formation and stabilization more than the dioxane and hence hasten carbyl sulfate formation. The data in Table I1 suggest that this is indeed the case, since 2 moles of SO3 add almost immediately to 1 mole of olefin, regardless of the starting SO3 to olefin ratio. In the cases where equimolar amounts of so3 and olefin were employed, 40 to SO%, of the starting olefin remained unreacted, regardless of reaction conditions. I t is important a t this point to consider why the phosphate should be so effective in producing alkenesulfonate to the virtual rxclusion of hydroxysulfonate. There is little doubt from the data that phosphate accelerates addition of the second mole of so3 to give the carbyl sulfate structure. Just as the phosphate was effective in facilitating the cleavage of the C-0 bond of the sultone by aiding incipient carbonium ion formation, SO also the phosphate facilitates the cleavage of the similar C--0 bond of the carbyl sulfate (see Figure 6). The carbonium ion may then revert to starting carbyl sulfate o r eliminate a proton to give a n alkenesulfonic anhydride. T h e anhydride is subsequently converted to alkenesulfonate upon neutralization. Acknowledgment
The authors thank the Esso Reseerch and Engineering (20. for permission to publish this paper. They acknowledge the significant contributions of J. H. Bartlett to the over-all synthesis effort. Further, they acknowledge the assistance of H.H. Brady and Fred Osmer, who performed many of the experiments, and Theodore Melchior and Nicholas Feldman, who were responsible for the NMR evaluation work. Literature Cited (1) Am. Soc. Testing Materials, Philadelphia. Pa., “Book of ASTM Standards,” Pt. 5, Method 855-56, 1955. (2) Bordwell. F. G., Colton, F. B., Knell, M., J . A m . Chem. Sac.
7 6 . 3950 (1954). Bordwdl. F. G.. Osborne. C.E..Zbid..81. 1995 11959). Bordwell; F. G.; Peterson, &f. L’., Zbid., 76,3952‘(1954). Ibid., 76,3957 (1954). Zbid., 81, 2000 (1959). Bordwell, F. G., Rondestvedt, C. S., Jr., Zbid., 70,2429 (1948). Bordwell. F. G.. Suter. C. M., Holbert, J. M., Rondestvedt, 2. S.,Jr., Zbzd.,68, 139 (1946). Burkhardt. G. N.. Lamvorth. 4 . J.. J . Chem. SOC.London 128. 684 (1926). (10) Gilbert, E. E., Chem. Revs. 62, 549 (1962). (11) $odd, E. S., “Mechanism and Structure in Organic Chemistry, pp. 253-4, Henry Holt, New York, 1959. (12) Lucas, H. T., Pressman, D., Znd. Eng. Chem., Anal. Ed. 10, 140-2 (1938). (13) Suter, C. M., J . A m . Chem. SOC. 60, 538 (1938). (14) Turbak, .4.F., IND.ENG. CHEM.,PROD.RES.DEVELOP. 2, 275 (1962). ’
RECEIVED for review May 17, 1963 .ACCEPTED July 10, 1963 Division of Petroleum Chemistry, 144th Meeting, ACS, Los Angeles, Calif., April 1963. VOL. 2