Analysis of Hydrocarbon Mixtures Application of Barrett Distillation H. E. MORRIS, W. H. LANE, A N D R . B. STILES' Monsan to Chemical Company, Texas Division, Texas City, Texas A simple analytical procedure for the determination of small amounts of impurities in aromatic hydrocarbon mixtures is presented. The method was developed for the analysis of ethylbenzene containing small amounts of benzene and diethylbenzene and is based on the influence of impurities on the true boiling point of the major component where the impurities boil at temperatures considerably removed from that of the main constituent.
OST of the styrene produced is made by the catalytic dehydrogenation of ethylbenzene ( 5 ) ,which may be produced by the catalytic alkylation of benzene with ethylene. The alkylate, containing benzene, ethylbenzene, diethylbenzene, and higher polyethylbenzenes, is fractionally distilled to recover ethylbenzene of high purity. I n this distillation it is vitally important that the diethylbenzene content of the product be kept at the lowest possible value. In the dehydrogenation step, ethylbenzene is partially converted to styrene and any diethylbenzene present is partially converted to divinylbenzene. The presence of divinylbenzene in the dehydrogenated mixture can be harmful if the styrene is recovered by fractional distillation, because divinylbenzene will copolymerize with styrene to form a cross-linked insoluble polymer. The presence of as little as 0.01% divinylbenzene will create this form of polymer (6). I n practice it has been found that if the diethylbenzene content of the ethylbenzene is kept below 0.03Q/, there are no serious consequences in plant operation. Various methods have been developed for the determination of small amounts of diethylbenzene in ethylbenzene. Quantities as low as 0.04570 by weight have been determined by mass spectrometric methods with indications that amounts as small as 0.003% by weight could be detected (3, 9). There is evidence that the sensitivity in spectrographic methods is low, in the range of 0.5 weight % ( I O ) . As little as 0.011% by weight can be determined by a method which combines ebulliometric nieasurements with fractional distillation ( 7 ) . The German ethylbenzene industry employed a Barrett distillation method by which amounts less than 0.05% by weight (4, 8) were determined. Information available indicates that this German method is similar to the method described in this paper. This paper presents a simple analytical procedure which was developed for the analysis of ethylbenzene containing small amounts of benzene and diethylbenzene. The method can be handled by laboratory technicians and is based on the influence of minor components on the boiling range of the mixture where the minor components boil a t temperatures considerably removed from the boiling point of the main constituent. The method a p pears to be capable of general application for the determination of small amounts of impurities in aromatic hydrocarbon mixtures.
Table I.
Temperature Difference between 5 and 507, Distilled, C .
\\-eight % Benzene Added 0.0
0.5 2.0
2.0
3.4
3.0
4.7
5 0
8.0
1 .o
METHOD
A 100-ml. sample of ethylbenzene is measured in the 100-ml. graduate and carefully transferred to the 200-ml. Barrett distilling flask, allowing 15 seconds for draining the graduate. The flask is placed in position on the asbestos board and connected to the condenser so that the side arm of the flask enters into the condenser tube a t least 5 cm. (2 inches). The graduate is placed a t the end of the condenser to receive the distillate. The thermometer is supported in the neck of the flask by means of a cork stopper in such a position that the top of the mercury bulb of the thermometer is just opposite the lower edge of the side arm and central in the neck of the flask.
Table 11. Effect of Diethylbenzene on Distillation Range of Ethylbenzene Weight % Diethylbenzene Added 0.00
0.01
Temperature Difference between 50% and Dry Point, C. 0.1 0.1
0.02
0.2
0.04
0.4 0.5
0.05 0.10
0.20 0.50
1.00
1.0 1.7 3.3 7.7
The flask is heated gently by the burner a t a distance of 5 cm. (2 inches) from the bottom of the asbestos board. Heating is continued gently until ebullition has started and vapors reach the bottom of the thermometer bulb, when the flame is withdrawn. This process is repeated twice to permit complete expansion of the mercury. The flame is then replaced and regulated to give a distillation rate of 5 to 7 ml. per minute, until the dry point is reached. Three temperature readings are recorded during the distillation: 5% distilled, 50% distilled, and dry point. The temperature is followed closely as the dry point is approached and the dry point is recorded as the temperature at which the liquid just disappears from the bottom of the flask.
APPARATUS
The apparatus consists of a 200-ml. Barrett distilling flask, a 600-mm., West, improved-type, Pyrex condenser supported by a ring stand, a 100-ml. graduate, an asbestos board 15 X 15 cm. ( 6 X 6 inches) with a 3.75-0111. (1.5-inch) hole supported on a ring stand, a Tirrill type burner, and an M.C.A. R-3 thermometer with a 70" to 160" C. range graduated in 0.2" divisions. It is assembled by the A.S.T.M. method ( 2 ) for distillation of industrial aromatic hydrocarbons. 1
Effect of Benzene on Distillation Range of Ethylbenzene
The temperature difference between 5 % distilled and 50% distilled is recorded and the per cent benzene is read from a graph
Deceased.
998
999
V O L U M E 2 1 , NO. 8, A U G U S T 1 9 4 9 constructed from values given in Table I. The temperature difference between 50% distilled and dry point is recorded and the per cent diethylbenzene is read from a graph constructed from data given in Table 11. This distillation will indicate the amount and diethylbenzene to *0.0170 in the of benzene to *O.lO’% ethylbenaene, providing reasonable skill is exercised, particularly in the determination of the dry point.
Table IV.
Effect of Toluene on Distillation Range of Ethylbemzene
Weight % Toluene Added to Ethylbenzene
Temperature Difference between 5 and 50% Distilled, C.
n n
0...~ 2 0.3 0.6 0.8
0.8 1.6 2.4 3.2
0.9
DEVELOPMENT OF METHOD
Various concentrations of benzene and diethylbenzene in ethylbenzene were prepared using Baker’s C.P. thiophene-free benzene, diethylbenzene prepared in this laboratory by fractional distillation, and Monsanto ethylbenzene purified by fractional distillation by 11.R. Fenske of Pennsylvania State College. These mixtures were used to determine the effect of benzene and diethylbenzene on the distillation range of ethylbenzene. Data are presented in Tables I and 11. Because it is difficult to obtain a reproducible value for the initial boiling point, the 5 to 50% valw i i taken for the benzene determination.
Table 111. Effect of lsopropylbenzene on Determination of Diethylbenzene in Ethylbenzene Weight % Isopropylbenaenr Added
Temperature Difference between 60% and Dry Point, ’ C.
0.0 0.1 0.2 0.3 0.4
0.2 0.2 0.4 0.6 0.; O.,
0.5 Ethylbenzene used contained 0.1% diethylbenzene.
dence of the lower limit characteristic is shown by data on the presence of toluene in ethylbenzene presented in Table IV. CONCLU SI03
Utilizing the method described, benzene may be determined in ethylbenzene with an accuracy of *O.lO%, provided the concentration does not exceed 3.oYO. Diethylbenzene may be determined i n ethylbenzene xith an accuracy of *0.01%, provided the concentration is not greater than 0.20%. This method is applicable to the analysis of similar ternary hydrocarbon mixtures. Traces of a compound boiling above the main component can be determined with greater accuracy than one boiling below the main product. The presence of a fourth constituent similar in character to one of the impurities being determined interferes with the determination. The 50% point was steadily depressed with increasing conceutrations of toluene. The upper limit method, however, is applicable to a limited extent for the presence of ethylbenzene in toluene and toluenein benzene, as shown by the data in Tahle V.
Table V.
The method is limited by the quantity of benzene and diethylbenzene in ethylbenzene. If benzene exceeds 3%, the 50% temperature reading may be suppressed, leading to a higher and incorrect value for diethylbenzene. The diethylbenzenr content cannot be determined accurately above a concentration of 0.2%. Above this value the temperature is rising too rapidly at the dry point to give reproducible values. Thermometer corrections, emergent stem corrections, and barometric pressure corrections may be ignored, its all results are based on temperature differences. A comparison between this Barrett distillation and the shielded A.S.T.M. distillation ( 1 ) showed that more reproducible dry points could be obtained using the unshielded flask distillation. I t was necessary to look directly down onto the flame when observing the dry point using the A.S.T.M. shielded distillation equipment, and thus more difficult to observe the true dry point. The presence of small quantities of other impurities may interfere with the determination. This has been shonn by the addition o f isopropylbenzene to the sample. A sample of ethylbenzene contairiing 0.01 % diethylbenzene was used in determining the effect of variable quantities of isopropylbenzene on the distillatior]. Yalues are presented in Table 111. To a sample of pure ethylbenzene were added 0.3% isopropylbenzrne and 0.05% diethylbenzene. The 50% to dry point range was 1.0” C. For 0.37’ isopropylbenzene done the 50% t o dry point range is 0.5’ C. For 0.05% diethylbenzene the 50% to dry point range is 0.5” C. When both isopropylbenzene and diethylbenzene are present together in ethylbenzene they have a cumulative effect on the 50% to dry point range, each exerting its own individual effect just as though the other were not present. The higher boiling impurity of a ternary mixture consisting of a major component and two impurities can be determined with greater accuracy than the lower boiling impurity by this distillation method. This is evident in the temperature spread shown for henzene and diethylbenzene in ethylbenzene. Further evi-
Effect of Ethylbenzene in Toluene and of Toluene in Benzene
Weight % Ethylbenzene Added t o Toluene
Temperature Difference between 50%o and Dry Point, C.
o n
n._.. .n.5
0.1 0.2 0.3 0.4 0.6
0.20 0.50 0.65 0.80
0.95
Weight % Toluene Added te Benzene 0.0 0.5 1.0 2.0
0.2 2.1 3.9 6.0
LITERATURE CITED
Am. 600. Testing Materials, Method D 850-47.
Ibid., D 86-46. Johnsen, S. E. J., ANAL.CHEW,19, 305 (1947). Livingston, J. W., “German Report on Ethylbenzene, Styrene, and Polystyrene, at Ludwigshafen, Germany,” Technical Industrial Intelligence Committee Rept. RM-332,June 1945. Mitchell, J. E., Jr., Trans. Am. Inst. Chem. Engrs., 42, 293 (1946). Staudinger, H., and Heuer, W., Ber., 67, 1164 (1934). Ewietoslawski, W., “Ebulliometric Measurements,” pp. 135-40, New York, Reinhold Publishing Corp., 1945. Technical Industrial Intelligence Committee (Rubber Subcornmittee) “Styrene Produrtion at Huh, Germany,” R e p t . R M 303,Sept. 27, 1945. Washburn, H. W., Wiles, H. F.,and Rock, S. M., IND.ENG. CHEM., ANAL.ED., 15, 541 (1943). Woernley, D., and Munch, R. H., “Analysis of Ethylbenzene for Diethylbenzene by Spectrographic Methods,” report submitted to Rubber Reserve Corp. by Monsanto Chemical Co.’s Central Research Department, July 20, 1943. R E C ~ V EDecember D 27, 1948. Presented at the Fourth Southwest Regional Meeting of the AMERICAN CEEMICAL SOCIETYShreveport, La.,December 10 and 11, 1948.
