718
INDUSTRIAL A N D ENGINEERING CHEMISTRY
VOl. 18. No. 7
Pure Hydrocarbons from Petroleum' By G. G. Brown and A. R. Carr UNIVERSITY
OF
MICHIGAN, ANN ARBOR,MICH.
PPARENTLY, no previous attempt to obtain appreciable quantities of pure hydrocarbons by systematic fractional distillation of petroleum has been reported. Many attempts to identify the hydrocarbons in petroleum are recorded, and the basis of all methods employed has been separation by distillation. Usually only simple distillation has been used. In only rare instances has a fractionating column of any kind been adopted. As the purpose was to produce pure hydrocarbons to be used for purposes of research, and not simply to isolate small quantities for the purpose of identification, the use of an efficient type of fractionating apparatus was imperative.
Evansls conducted a distillation of the light fractions of Asiatic petroleum, using three different columns, first a YoungJs 12-bulb pear column, then a 75-cm. glass column containing lead balls 5 mm. in diameter, and finally through a 7 5 - w . column containing a t the lower end 10 cm. of copper gauze, then 30 cm. of steel balls 2 mm. in diameter, and the upper 35 em. filled with steel balls 4 mm. in diameter. The third column was found to be the best. distilled crude oil from a long-neck, Reed and round-bottom flask. The fraction up to 170" C. so obtained was then distilled from a similar flask through a 4-pear stillhead. The column was 58 cm. high and the bulbs were 2.5 cm. The rate of distillation was 2 drops a second. A simiHistorical lar 102-cm. column containing 8 pears was used to distil the The simple stills used bv Pelouze and Schor- fraction under 180' C. a t the same rate. Cuts were made a t l e m m e r ,2 * Beils t e i n and' 30" C. intervals. KurbatowJ3and Mabery4 in Anderson and Erskine16 s e p a r a t i n g the hydrocaridentified the compounds OMMERCIAL straight-run gasoline from the Cabin bons for purposes of identipresent in natural gas gasoCreek Field was systematically fractionated in a fication gave very unsatisline by distillation in "ordicolumn, 18 feet high, packed with modified Raschig factory results. The use of nary columns" and with "a rings '/z inch in diameter, in a similar column 10 feet E n g l e r f l a s k ~ ~ nled~ # ~ high containing Ih-inch rings, and in a glass column regulated temperature stillCoates8 to state: h e a d . " With this proce5 feet high packed with %-inch rings. Five fracd u r e e i g h t fractionations tionations were sufficient to separate cyclopentane, It is exceedingly difficult to were sufficient to separate cyclohexane, benzene, and toluene from the pentanes get the hydrocarbons in a the pentanes into fractions state of purity. Indeed, I do and hexanes and gave 200 cc. of isopentane (30.5' to boiling within 0.5' C., nine not think it can be done by 30.6" C.) and 250 cc. of normal pentane (36.2" to 36.3" distillation. Several hundred fractionations separated the C.). Seven fractionations gave 300 cc. of isohexane (61.1' distillations were made, but hexanes into 1.0" C. fracto 61.2' C.) and 1000 cc. of normal hexane (68.9' to each fraction continued to tions, and seventeen fracbreak up into fractions with 69.0' C.). Nine fractionations separated 2000 cc. of t i o n a t i o n s separated the lower and higher boiling normal heptane (98.6' C. to 98.7' C.). Ten fractionapoints. heptanes into 1.0" C. fractions gave 2500 cc. of isooctane (117.9' to 118.1" C.), tions. These results, rather 1100 cc. of normal octane (124.3' to 124.5" C.) and 600 Warrens introduced the u n s a t i s f a c t o r y in themof normal nonane (150.2' to 150.9' C.). cc. use of a fractional condenser selves, are the best of agy This work with imperfect columns of moderate effior dephlegmator composed so far recorded. ciency indicates that pure hydrocarbons of boiling of a worm surrounded by a points up to 150' C. may be obtained by systematic Fractionating Column the b a t h whose temperature Best Means fractionation at atmospheric pressure. could be controlled. This equipment was also used by From the experience briefly M a b e r ~who , ~ found it more reviewed above it is evident effective than the Hempel column, although demanding more that efficient fractionation of petroleum can be accomplished attention. Young and Thomaslo used a similar dephlegmator only by means of a column. This fact can be theoretically described as a "regulated temperature stillhead" in conjunction deduced by the method given by Leslie,'G who shows that the with a pear column.ll With this equipment they were able to ideal fractionating column is always more efficient than either separate the pentanes from the butanes in one distillation the ideal simple dephlegmator or differential dephlegmator and to separate the pentanes from each other in twelve dis- for making complete separations into pure components, or tillations. by a careful consideration of the fundamentals of distillation. The ideal distillation process a t constant pressure consists Bushong12 used a copper retort and glass flasks equipped with a Le Bel-Henninger dephlegmator. The hydrocarbons in countercurrently contacting the upward flowing vapor and were heated directly by a coil of German silver wire within the downward flowing liquid so intimately that equilibrium the flask. Fractions were cut every 2' C. Bushong makes is established a t all points and the change in composition of the interesting statement that he "is firmly convinced that a the phases occurs continuously. The dephlegmator is a series of fractional distillations carried out on a factory scale partial condenser operated to serve the double purpose of would richly reward the immense labor involved by opening refluxing or returning part of the distillate to the still and up a new world of possibilities in the way of chemical products washing the rising vapors with this refluxed liquid. In an apparatus of this kind condensation takes place throughout to be manufactured from petroleum." its entire length. A large part of the condensed liquid is 1 Received March 1, 1926. Presented before the Division of Petroformed in the lower parts of the dephlegmator and returns leum Chemistry at the 71st Meeting of the American Chemical Society, to the still, traversing only a small part of the total length Tulsa, Okla., April 6 to 9, 19.26. *Numbers in text refer to bibliography at end of article. of the apparatus. The same total effective washing could be
A
C
,
INDUSTRIAL AND ENGINEERING CHEMISTRY
July, 1926
attained by condensing a much smaller quantity of liquid if it were all returned from the extreme end of the apparatus as is done in the fractionating column. These considerations indicate that the column should be as nearly adiabatic as possible, the heat loss being entirely in the condenser and reflux. Preliminary Tests of Fractionating Columns
The first column constructed was similar to the column described by Dufton,I7 who reported that this type was very satisfactory for distilling small amounts of liquids as the amount of liquid contained in the column was a minimum. The column was made of an 8-foot length of 4-inch iron pipe with standard cast-iron flanges on either end. It contained a helix wound around a closed core constructed of a 2-foot length of 2-inch pipe welded to a 3-foot length of 21/4-in~hpipe topped by 2 feet 8 inches of a 2l/*-inch pipe. The helix was welded to the core, l ' / ~inches between turns, and closely fitted the inside of the column. The column was bolted to the still through a special iron gasket used as the support of the helix core. The still body was made of welded l/rinch steel plate of about 15 gallons capacity. A look box was provided 011 top of the column so that hhe amount of reflux could be observed. The total condenser was made by forming a 20-foot length of 1-inch iron pipe into a coil 1 foot inside diameter and immersed in water, or ice and water, as the occasion demanded. The liquid in the still was heated internally by a coil of resistance ' wire. The coil was mounted on a chrome1 ''-4, wooden platform made fast to the cover of the hand hole. Contact was made through two porcelain spark plugs in the hand hole cover. Current was taken from a 220-volt direct current line and controlled by means of a rheostat of the granular resistance type. The still and column were lagged with 85 per cent magnesia about 2 inches thick. Preliminary work on the separation of benzene-toluene mixtures in this column gave very unsatisfactory results. The helix was removed and the column was filled with modified Raschig rings inch in diameter. This type of packing provided a greater countercurrent contacting surface and free volume than the helix. These rings were made from sheet metal cut to the proper length and width and formed in a mandrel. Preliminary work with this second column also gave unsatisfactory results, although much better than the first column. The length of the column was then increased by adding a 10-foot length of 3-inch pipe above the 8 feet of 4-inch pipe. This 18-foot column was filled with the same Raschig rings separated by the insertion of truncated cones of perforated sheet metal, every 2 feet. These cones served the purpose of directing the downward flowing liquid toward the center of the column, as it had been noticed that the returning liquid had a tendency to flow down the inside walls. Preliminary work on the separation of benzene-toluene mixtures in this 28-foot column gave satisfactory results. The intermediate fraction was fairly constant, containing about 2 liters regardless of the amount of liquid distilled. Gasoline Used
One hundred and thirty-five gallons of straight-run gasoline distilled from crude oil from the Cabin Creek Field, W. Va., were obtained from the Elk Refining Co., Charleston, W. Va. The A. S. T. M. distillation of this gasoline ran as follows: Percent Initial 10
20 30 40 50 60
O F .
144 196 215 229 243 257 268
C. 42.2 91.1 101.7 109.4 117.8 124.9 131.5
Percent
io
OF.
c.
280 138 80 297 146.9 90 321 160.5 95 344 173.5 (end) 9 8 . 5 376 191.4 Specific gravity 0.7260 at 21"/15O C.
719
Two samples tested for sulfur by the "doctor" tests gave negative results; five samples tested by the more thorough lead acetate and sodium nitroprussiate test also gave negative results for sulfur. Six samples tested for nitrogen by treatment with sodium and a ferrous salt gave no ferrocyanide test, showing the absence of nitrogen. The results of two checks in combustion tests gave 86.16 per cent carbon and 13.43 per cent hydrogen. From these tests we may assume that the gasoline used was a mixture of pure hydrocarbons. Preliminary Distillation of Gasoline
In distilling gasoline it was found that the heavy residue remaining in the still after the lighter fractions had been taken off was a t so high a temperature that gasoline could not be charged directly into the still until after it had been cooled for 3 to 5 hours, because the highly volatile components of the fresh charge would vaporize and create such pressures as to prevent the liquid from running into the still body. As this work was conducted, not as an end in itself, but simply to supply raw material for further study, results had to be obtained in the shortest possible time. Accordingly, the still was equipped with a 10-gallon blow case in order to charge the still promptly and not lose time, as the work was carried out on the shift basis, operating 24 hours a day. The air was supplied by a hand-operated diaphragm pump. A gage glass was so arranged that the flow of liquid into the still could be observed and the cock closed promptly to prevent an excess of air from entering the still. Because of the limited capacity of the still it was impossible to fractionate the entire quantity of gasoline in one run. So the still was filled, run down as far as possible, and refilled by raw gasoline. This process was continued until the heavy residue in the still amounted to 5 gallons, when it was drained and another similar run was made. At the end of three runs of this kind all of the gasoline had been distilled once and collected into the following fractions : -Temperatur--
c.
Below 55 55 to 90 90 to 110 110 to 120 120 to 125 125 to 130 130 to 135 135 to 140 140 to 145 145 t o 150 150 to 155 155 to 160 Above 160
F. Below 131 131 to 194 194 to 230 230 to 248 248 to 257 257 to 266 266 to 275 275 to 284 284 to 293 293 to 302 302 to 311 311 to 32Above 320
-Weight-Kgs.
2.5 82.5 57 31.8 22.2 20.4 18.2 27.3 19.1 21.4 14.1 7.75 23.2
TOTAL347.45
Lbs. 5.5 181.75 124.75 70.0 51.0 45.0 40.0 60.0 42.0 47.0 31.0 17.0 51.0
766.0
Systematic Fractionation
FIRSTFRACTIONATION (CURVE1)-These fractions were then systematically redistilled in the same column. The method employed was to charge the first fraction and distil slowly with a large reflux ratio until the temperature a t the top of the column reached the initial temperature of the fraction charged, when the second fraction was charged. The same process was repeated for the latter fractions. This method gave the more volatile components in a purer state than if the temperature a t the top of the column was run up to the initial temperature of the fraction to be charged, as the column would then contain considerable amounts of the heavier hydrocarbons, which would contaminate the lighter, more volatile hydrocarbons entering the column from the still when the succeeding fraction was charged. Working in this manner, the above fractions were redistilled and the distillate received as shown by Curve 1. In this fractionation the pentane (28" to 38" C.), hexane (60' to 70" C.), and iso-
INDUSTRIAL AA'D ENGINEERING C H E M I S T R Y
720
octane (117" to 121" C.) fractions were clearly evident, and there is definite suggestion of isoheptane, heptane, isononane, and nonane. SECOND FRACTIONATION (CURVE 2)-Those fractions collected a t 135" C. (275" F.) or over were then redistilled in the same manner. The distillate was received as shown by Curve 2. The lower boiling fractions were not distilled a t this time, as the intent was to work the more volatile components out of the higher boiling fractions. THIRDFRACTIOXATION (CURVE 3)-LJ7ith the exception of that fraction distilling below 35" C. (95" F.), all fractions were redistilled in the same manner, giving distillate according to Curve 3. The lowest boiling fraction was not run because the losses in distilling material boiling at room temperature are very high. This run shows clearly the presence of the pentanes, hexanes, isoheptane, normal heptane, isooctane, apd normal nonane, while normal octane, two isononanes, and isodecanes are suggested. FOURTH FRACTIONATION (CURVE4)-Those fractions collected above 101" C. (214" F.) were then systematically redistilled, giving distillate as shown on Curve 4. FIFTHFRACTIONATION (CURVE5)-All fractions collected above 97" C. (207" F.) were then systematically redistilled, giving distillate as shown on Curve 5. SIXTHFRACTIONATION (CURVE6)-All fractions were then redistilled, with results as s h a m by Curve 6. Here the presence of butane was first indicated. The pentane and hexane fractions were beginning each to separate into two fractions and the nonane fractions are clearly defined. The fraction collected under 10" C. (50" F.) was composed largely of butanes and evaporated rapidly, even though kept packed in ice. As many of the fractions were becoming small in quantity, a fourth column, 10 feet high and 11/4 inches inside diameter, filled with l/4-inch Raschig rings and mounted on a 20-gallon still. was used for the following distillations.
Vol. 18, Yo. 7
what less than 2" C. (35.6" F.). The isopentane is clearly distinct from the normal pentane fraction and the isohexane is distinct from the normal hexane. EIGHTHFRACTIONATION (CURVE 8)-Pentanes. In the fifth fractionation of pentanes-carried out in a %liter glass still with a &foot column packed with '/S-inch rings200 cc. of isopentane were collected, boiling a t 30.5" to 30.6" C. with a specific gravity of 0.6240 a t 17"/15.5" C. About a liter of similar material boiling between 29.8" and 30.8" C. was obtained, which could have been further fractionated to yield more pure material, as this was obtained by a total of only five fractionations from commercial gasoline. About 250 cc. of material boiling a t 36.2" to 36.3" C. were collected. This fraction had a specific gravity of 0.6340 at 14"/15.5" C. and was presumably normal pentane. About 1 liter was obtained between 36.0" and 36.4" C. in the distillate from the 10-foot column. Cyclopentane. In the seventh fractionation there was a suggestion of a component a t about 50" C. This became more evident in the eighth fractionation and was evidently due to a trace of ~ y c l o p e n t a n e . ~ ~ Hexanes. About 2 liters of isohexane, boiling a t 60.5" to 61.5" C. and 5 liters of normal hexane boiling a t 68.0' to 69.0" C. were collected. These fractions were later separately twice fractionated in the &foot glass column giving 300 cc. of isohexane boiling a t 61.1" to 61.2" C., and having a specific gravity of 0.6620 a t 17"/15.5" C., and 1 liter of normal hexane boiling a t 68.9" to 69.0" C. and having a specific gravity of 0.6622 a t 20"/20" C. and refractive indices of 1.3851 and 1.3810 a t 20" and 25" C., respectively. Benzene Fraction. In the sixth fractionation there was some evidence of a fraction separating from the hexanes and boiling at about 75" to 80" C. During the eighth fractionation this small fraction separated a t about 78" to 80" C., with a specific gravity of 0.8590 a t 20" C. compared to water a t 15.5". These properties indicate that this fraction is com-
/SO'
G
160"
0
I
ts!
k
/zoo
b L
/0O0
a,
Q
F I2
FRAC TION4 T / O N
60'
OF
__--
_____ 2 oo
SEVENTHFRACTIONATION (CURVE7 k T h e fractions obtained from the sixth fractionation werecarefully fractionated with very large reflux ratios, giving the results indicated by Curve 7. The butane came over almost as soon as the lowest boiling fraction had been charged, before any heat was supplied to the still. The temperature indicated was some-
CURVES
GASOLINE
CABIN CREEK FIELD BROWNann
CARR
nosed Dartlv of benzene and Drobablv an isoheDtane or DOSkbly cyclohexane.'* Heptane Fraction. In the sixth fractionation about 5 liters were collected at 93.8" C. having a boiling range of 93.7" to 94" C. Specific gravity determinations indicated that this cut contained some benzene or naphthene derivaI
"
INDUSTRIAL AND ENGINEERING CHEMISTRY
July, 1926 BOILINGRANCB C. AT 760 Mar. SPSCIFICGRAVITY
SOURCS
RBFERENCB
Isogcnlanc (2-Melhyl B u t a n e ) Fraction Cabin Creek, W. 3 0 . 5 - t o 3 0 . 6 0 . 6 2 2 0 20°/15.50 C. Va. gasoline Pa. Synthetic
30.2 30.4
Ohio and Pa.
2 9 to 30 14' C.
Synthetic
27.95
0.6380
30 28
0 , 6 3 8 5 14' C. 0 . 6 3 9 4 4 o / o o c.
2 7 . 8 to28.3
36.2-to36.3 36.3 36.3 36.3 36.3 37 3 5 , s t o 36 .O
6 1 . 1 to 6 1 . 2 61 61 61.3 61 . 7 to 6 2 . 4
l5,5'/15.5' C. Natural gas gnsoAnderson and line Erskine" A-ormal Pentane Fraction Cabin Creek, W. 0 , 6 3 0 9 20°/15.50 C. Va. gasoline Chavanne and 0 . 6 4 5 4 40/0° C. Simonz3 Thorpe and 0 . 6 4 7 3 00 c. Jones20 Mabery a n d 0 . 6 2 5 0 25'/25' C. Young2* Young and 0 . 6 4 5 4 0°/4" C. Thomaslo Markownikoffz' Russia 0 , 6 2 6 3 170 c . 0 . 6 3 0 9 15.5°/15.50 C. Natural gas gaso- Anderson and Erskine's line Isohexane (2-Methyl Pentane) Fraction 0.6248
0 . 6 6 0 0 20°/15.50 C.
60.55 to 60.85 0 . 6 7 2 8
0.6762 0.6760
Redwood28
Maberyzd Warrens Chavanne and Simon23 American petroYoung and O o / 4 O C. Thomasla leum Zelinsky 26 Synthetic 17'/4' C. Risseghemz' Synthetic 150/150 c. 15.5°/15.50 C. Natural gas gaso- Anderson and Erskine" line 00 c. Ohio a n d Canada Mabery' Thorpe and 00 c. Jones20
0 . 6 7 2 0 0°/40 C. 0.676 4'/15' C. 0.658
6 0 . 5 to 6 1 . 0 0 , 6 5 8 3 60.15to60.2 0.6580 5 9 . 7 to 6 1 . 0 0 . 6 6 1 3 61.2 62
Cabin Creek, W. Va. gasoline Pa. and New Brunswick
0.664
Peru
tives,'O as well as the isoheptanes. This fraction was set aside and lost through a leaky container. The cuts above 95" C., run during the seventh and eighth fractionations, gave 9 liters of a fraction boiling a t 98.7" to 99.1" C., and having a specific gravity of 0.7322 a t Zoo/ 15.5' C. Upon careful refractionation in the 5-foot column 2 liters boiling a t 98.6" to 98.7" C. and having a specific gravity of 0.6893 a t 20"/20" C. and refractive indices of 1.4068 and 1.4058 a t 20" and 25" C., respectively, were obtained. This product was partially distilled a t a pressure of 310 mm. of mercury and the residue had precisely the same vapor pressure as the distillate, indicating that the fraction is a p u r e compound, normal heptane. Toluene Fraction. I n the fifth and sixth fractionations there was evidence of a compound boiling at about 110' C. This evidence persisted through the seventh and eighth fractionations when a small amount of material was collected at about 110" C. having a specific gravity of 0.8385 at 20" C. compared to water at 15.5" C. These properties indicate that this fraction is composed of toluene and probably isooctanes. Octanes and Nonanes. Two further fraction:ttions of the material boiling above 115" C. obtained from the eighth fractionation gave the following material: Volume cc. Compound 2500 Isooctane 1100 Normal octane 600 . Normal nonane
Boiling range
c.
1 1 7 . 9 to 1 1 8 . 1 1 2 4 . 3 to 1 2 4 . 5 150.2 to 1 5 0 . 9
BOILINGRANGE O C. AT 760 Mx. SPECIFICGRAVITY
0.6622
SOURCE Normal Hexane Fraction 20"/20" C. Cabin Creek, W.
0.6771 0.6770
0°/4" C. 0°/4' C.
6 8 . 5 to 6 9 6 7 . 9 to 6 9 . 2
0.6767
15.5°/15.5a
9 8 . 6 to 9 8 . 7
0.6893
Norma2 Heptane Fraclion 20°/200 C. Cabin Creek, W. Va.
68.9to69.0 Warren18 Thorpe a n d Jones20 Mabery and Hudson21 Young and Thomas10 Richter22 Chavanne and Simon23
Specific gravity 0 . 7 1 6 5 a t 20°/4' C. 0.7262 a t 2 0 ° / 1 5 . 5 0 C. 0 . 7 3 8 0 a t 2 0 ° / 1 5 . 5 ' C.
Curve 9 shows the results on an enlarged scale of the eighth fractionation of isoijctane and indicates where the cut was made to obtain material for the final fractionation. The normal octane redistilled in the &foot column gave 450 cc. boiling at 124.3" C. with a specific gravity of 0.7123 a t 20°/
72 1
68.95 68.95
98.4
(Refractive index 1 . 3 8 5 1 a t 2 0 ° C . , 1.3840 at 25' C.) Appalachian oil Maberyh American oil Young and Thomaslo KymerZD Natural gas gaso- Anderson and line Erskine'a Va., gasoline
0 . 7 0 1 8 6 0°/4' C.
98.25 t o 98.45 0 . 6 8 2 8 8 200 c. 17.5'C. 9 7 . 5 t o 99 0.709 9 8 . 5 to 9 9 . 5 96 t o 97 98 t o 9 8 . 3 0 . 6 8 7 9 4'/15' C. 98,43 98.2to 99.3
RSFERENCB
American oil
(Refractive index 1.4068 a t 20°C., 1.4058 a t 25' C.) Young and Erskinel0
Rangoon Peru American oil Ohio
Schorlemmer2 Beilstein3 Maberv4 C h a v a k e and Simon28 0 , 7 0 0 4 8 0°/4' C. Thorpeso 0 . 7 1 1 7 15.5°/15.50 C. Natural gas gaso- Anderson and line Erskine'o
Isooctane (4-Methyl Heplane) Fraction 117.9 to 118.1 0 , 7 1 6 6 20°/15.50 C. (Refractive index, Cabin Creek 1.40063 at 25' C. gasoline 0 , 7 2 1 7 (Refractive index, 1 . 3 9 7 8 a t 23' C.) Clarke31 118 Normal Octane Fraction 124.3 to 124.4 0 . 7 1 2 3 20°/200 C. Cabin Creek (Re. fractive index 1.4059 a t 20' C., 1.4037 a t 25' C.) 124.7 0 . 7 0 6 8 15'/15' C. Sumatra Clarke21 121 0.732 1 2 . 1 ' C. -4merican oil Lemoine32 124 to 125 0 . 7 1 3 4 Ohio Mabery and Hudson21 Iiormal X o n a n e Fraction 150.2 to 150.9 0 . 7 3 8 0 20°/15.50 C. Cabin Creek gasoline 150.8 0.7555 Pa. Warren9 151 Ohio Mabery4
20" C., and indices of refraction of 1.4059 and 1.4037 a t 20" and 25" C., respectively. Discussion
Although the fractionating equipment used is not so effective as a properly designed bubble-cap plate column, the results obtained are more satisfactory than any previously reported, and indicate that pure hydrocarbons with boiling points up to 150" C. may be obtained by fractionating commercial petroleum products in efficient columns. For satisfactory results the columns should be well insulated or nearly adiabatic, a large part of the condensate must be returned or refluxed to the top of the column, and distillation must be conducted slowly, with controlled uniform heat input to the still. In addition to the indication of butane, benzene, cyclopentane, toluene, isoheptane, isononanes, and isodecanes, the products given in the accompanying table were obtained. The thermometer used had been calibrated 4 months previously by the Bureau of Standards. All temperatures were corrected to 760 mm. mercury pressure. Specific gravities were taken by a calibrated Westphal's balance and are given as taken. All of these compounds attained by fractionation of gasoline are to a high degree pure. Acknowledgment
The authors wish to acknowledge the aid of Messrs. Benton, Boiney, Fleming, Hunn, Miles, O'Neil, and Rodenberg in conducting fractionations, and the original suggestion of E. H. Leslie. Bibliography I-Pelouze and Cahows, Compt. rend., 64, 1241 (1862); 66, 50.5 (1863); 67, 62 (1863). 2-Schorlemmer, J . Chcm. SOC.( L o n d o n ) , 15, 419 (1863).
INDUSTRIAL A N D ENGINEERING CHEMISTRY
722
3-Beilstein and Kurbatow, Ber., 13, 2028 (1880). 4-Mabery, Proc. A m . Acad. Arts Sci., 81, 10 (1895). 5-Coates and Best, J . A m . Chem. SOC., 2s. 1155 (19031. ti-Marcusson, Milt. kgl. Materialgrtlfungsamt, S1, 301. 7-Mabery, J . Ind. Eng. Chem., 6, 101 (1914). &Coates, J. A m . Chem. Soc., 28, 384 (1906). 9-Warren, Proc. A m . Acad. Arts Sci., 21, 56 (1891); 34, 92 (1898). IO-Young and Thomas, J. Chem. SOC. (London), ‘71, 440 (1897); Young, Ibid., 73, 905, 920 (1998). 11-Young, Chem. News, 71, 177 (1895). 1.2-Bushong, J . Ind. Eng. Chem., 6, 888 (1914). la-Evans, J . SOC. Chem. I n d . , S8, 401T (1919). 14-Reed and Williams, Ibid., 88, 319T (1919); 39, 289T (1920). 15-Anderson and Erskine, I n d . Eng. Chem., 16,263 (1924). l&Leslie, “Motor Fuels,” 1923, p. 114. Chemical Catalog Co. 17-Duftor1, J . SOC.Chem. Ind., 88, 461’ (1919).
Vol. 18, s o . 7
18-Fortey. J . Chem. SOC.(London), 13, 932, 949 (1898). 19-Warren, Mem. A m . Acad., 9, 135 (1867). 20-Thorpe and Jones, J . Chem. SOC.(London), 63, 290 (1893). 21-Mabery and Hudson, Proc. A m . Acad. Arts Sci., 82, 101 (1896). 22--Richter, “Organische Chemie,” 11th ed., 1909, Aufl. I, p. 90. 23-Chavanne and Simon, Comfit. rend., 166, 1324 (1919). 24-Mabery and Young, J. Franklin Inst., 162, 57, 81 (1907). 25--MarkownikotT, B n . , SO, 975 (1897). 26-Zelinsky, Ibid., 40, 4743 (1907). 27-Risseghem, Bull. SOC. chim. Belg., SO,8 (1921). 28--Redwood, “A Treatise on Petroleum,” 3rd ed., 1918, Vol. I, p. 242. Pg-Kymer, J . grakt. Chem., 64, 126. 30--Thorpe, J . Chem. SOC.(London),87, 73 (1880). 31-Clarke, J. A m . Chem. Soc.. 88, 520 (1911). 32-Lemoine, Bull. SOC. chim., 41, 163 (1884). 33-Young and Richardson, J . Franklin I n s t . , 41, 50.
The Aromatic Hydrocarbon Content of Natural Gas Gasoline’ By A. M. Erskine HAMILTON COLLEGE, CLINTON, N. Y.
N T H E study of the com-
were not successful owing to The presence of small amounts of benzene, toluene, position of natural gas the difficulty of salting out and m-xylene in a sample of Pennsylvania absorption gasoline by Anderson and the very small amounts of the natural gas gasoline has been proved by fractionation, Erskine12the specific gravitysulfonates . concentration of each aromatic hydrocarbon by exboiling point curves for the Direct nitration of portions traction with liquid sulfur dioxide, and identification first, third, and fifth prelimiof the crude gasoline fractions as the corresponding dinitro or trinitro derivative. nary fractionations showed by mixed acid was also tried. Quantitative determinations on the fractions using marked and progressive inWhile nitro products were obthe nitrobenzene critical solution temperature method creases of specific gravity in tained which appeared to be showed that the original sample contained 0.6 per cent the ranges 75” to 95” C. and m-dinitrobenzene a n d 2 ,4benzene, 0.6 per cent toluene, and 1.2 per cent m-xylene 100” to 120’ C., indicating dinitrotoluene from the corby weight. the probable presence of benresponding fractions, the very While this work was begun as a matter of scientific z e n e and toluene in these small amounts of these made interest, it has resulted in the development of the nitrofractions. From the identisatisfactory purification diffibenzene critical solution temperature method of deterfication of benzene, toluene, cult. It was also observed mining aromatic hydrocarbons in gasoline, which it is t h a t t h e solubility of the and m- and p-xylene in small believed has possibilities of application in the petroleum nitroaromatic compounds in amounts in Ohio and Canaindustry. dian crude p e t r o l e u m b y the gasoline layer was very Mabery,3 and of benzene and appreciable and this caused toluene in Pennsylvania petroleum by Young,4 and consider- considerable loss. Thole8 calls- attention to this mutual ing the close relationship between natural gas and petroleum, solubility as a factor which cannot be entirely neglected in one should expect to find traces of aromatic hydrocarbons in the usual nitration method for determining aromatics and, the gasoline from Pennsylvania natural gas. of course, it is the basis of the nitrobenzene critical solution It seemed worth while to investigate this matter in some temperature method of analysis for aromatics. This direct detail, and the work described in this paper was carried out nitration method was therefore given up as impracticable with the object of getting qualitative proof of the presence for the purpose in view. An entirely satisfactory method proved to be one in which of the aromatic hydrocarbons in this type of gasoline and also an accurate quantitative determination of the small amount the aromatic hydrocarbons were extracted from the crude fractions by liquid sulfur dioxide. The greater solubility of of each present in a typical sample. aromatic hydrocarbons in this reagent as compared with QualitativeIdentifications nonaromatic has been studied by Rrowery7 and by Egloff .E After evaporation of the liquid sulfur dioxide the extracts Attempts were made to identify the aromatic hydrocarbons containing a high concentration of aromatics were nitrated from the sulfonic acids obtained in the preparation of the and the nitro products purified and identified. This principle aromatic-free fractions which were used in the study of the with a special modification has been used by Tausz and Stunitrobenzene critical solution temperature method of analy- berg in isolating toluene and xylene from petroleum fractions. S ~ S . The ~ method that was tried involved salting out the The sample used for these qualitative tests was a straight sulfonates, conversion to the sulfonyl chlorides, and iden- absorption natural gas gasoline, specific gravity 0.6619 tification as the corresponding amides, but these experiments (15.56”/15.56” C.),10identical in source and composition with that used in the quantitative determinations described later. 1 Received March 24, 1926. Presented before the Division of Petro-
I
leum Chemistry at the 71st Meeting of the American Chemical Society, Tulsa, Okla., April 5 to 9. 1926. 2 THISJOURNAL, 16,263 (1924). a A m . Chem. J . , 18, 43 (1896). 4 J . Chem. Soc., 78, 905 (1898). 6 See page 694, this issue.
J. SOC. Chem. I n d . , 8 8 , 39T (1919). Petroleum Reo.,36, 351, 385, 401 (1917); C. A . , 11, 2540 (1917). 8 Met. Chem. Eng., 18, 396 (1918). 2.angezu. Chem., 32,I, 139 (1919); C. A . , 18, 3310 (1919). 10 The gasoline samples were supplied by the United Natural Gas Co..Oil City, Pa. e
7
