‘‘
Pippen and Masahide Konaka, and Others by The infrared determinations made G. F. Bailey and E. L. Gong.
w’
LITERATURE CITED
(1) Bricker, C. E., Johnson, H. R., IND.
ENG.CHEM.,ANAL.ED. 17,400 (1945).
( 2 ) Carson, J. F., Jr., Wong, F. F.,
Western Regional Research Laboratory, Albany, Calif., private communication, May 1959. (3) Evans, P. N., Albertson, J. hl., J . Am. Chem. SOC.39, 456 (1917). (4) Keeney, M., ANAL. CHEM.29, 1489 (1957). (5) Lipsky, S. R., Lovelock, J. E., Land-
owne, R. A., J . Am. Chern. SOC.81, 1010 (1959). (6) Little, A. D., Inc., “Flavor Research and Food Acceptance,” pp. 290, 317, Reinhold, New York, 1958. (7) Meyer, H., Monalsh. Chem. 15, 164 (1894). (8) Nef, J., Ann. Chem. Liebigs 318, 40 (1901). (9) “Organic Syntheses,” Vol. 23, p. 27, Wiley, New York, 1946. (10) Phillips, C., “Gas Chromatography,” Academic Press, New York, 1956. (11) Ralls, J. W., J . Agr. Food Chem. 7, 505 (1959). (12) Ibid., 8, in press (1960). (13) Rodd, E. H., “Chemistry of the Carbon Compound,” Vol. IA, p. 583, Elsevier, Il’ew York, 1951.
(14) Westerfeld, R. W.,Jr., J . Bid. Chem. 161, 495 (1945). RECEIVEDfor review April 20, 1958. Accepted November 2, 1959. A laboratory of the Western Utilization Research and Development Division, Agricultural Research Service, U. S. Department of Agriculture. The author, as collaborator, is employed by the National Canners Association with which this work was conducted cooperatively. Presented in part before the Division of Agricultural and Food Chemistry, 135th Meeting, ACS, Boston, Mass., April 1959. Mention of specific products does not imply endorsement by the Department of Agriculture over others of a similar nature not mentioned.
Gas Chromatographic Analysis of Olefinic Naphthas in the Three- to Six-Carbon Range With the Aid of a Subtraction Technique RONALD L. MARTIN Research and Development Department, Standard Oil
b Analysis of olefinic naphthas for individual components is exceedingly difficult because of the many closeboiling compounds present. A new gas chromatographic technique, involving successive runs in the presence and absence of olefins, greatly simplifies the problem. In one run, olefins are removed with a short absorber containing silica gel coated with sulfuric acid following the chromatographic column; saturated hydrocarbons pass quantitatively through the sulfuric acid absorber and are determined directly. The other run is made without the absorber, and the olefins appearing in the same peaks with saturates are determined by difference. Three typical olefinic naphthas demonstrate wide variation in components of three to six carbons.
N
containing unsaturated as well as saturated hydrocarbons are common in the petroleum industry. Determination of individual components in such olefinic naphthas is exceedingly difficult because of the many closeboiling compounds present. Distillation techniques (3, IO) were valuable when there was no other approach, but are time-consuming and difficult. Conventional gas chromatographic techniques have been of little value beyond the five-carbon region for the analysis of olefinic naphthas. Polar and phases stationary phases (f,6,8,12)
336
APHTHAS
ANALYTICAL CHEMISTRY
Co. (Indiana),
Whiting, Ind.
containing silver nitrate (a, IS) have been used to retain olefins for longer times than saturated hydrocarbons of like boiling point; however, above five carbons, the retarded olefins are usually eluted with saturated hydrocarbons of higher boiling point. The only reported method for determining olefins through six carbons (7) requires that saturated hydrocarbons be absent. A different approach was used in a recently reported method for gaseous olefins and saturates through five carbons (6);olefins were removed by an absorption precolumn, and saturates were determined before and after hydrogenation. The absorption column was about 5 feet long and contained silver sulfate and sulfuric acid on kieselguhr. Disadvantages of this method are that all olefins with the same carbon skeleton are lumped together, and determination above five carbons is almost impossible because of the extreme tailing from the absorption column. A somewhat similar approach has been used in developing a simple technique that aids in the analysis of olefinic naphthas for components of three to six or more carbons. The technique is based on the removal of olefins by a n exceptionally short absorber containing silica gel coated with sulfuric acid following the chromatographic column. The olefins are absorbed as a result of the addition reaction of sulfuric acid to olefins. Saturated hydrocarbons pass through quantitatively and are deter-
mined by difference from a chromatogram obtained with them present. Analyses of three typical naphthas for components of three to six carbons show widely different isomer distributions. EXPERIMENTAL
Conventional gas chromatographic equipment was employed except for a gas density balance of new and simplified design (11), developed in these laboratories. With this detector, calibration of output was not necessary. The chromatographic column was made of copper tubing and packed with causticwashed firebrick of 35 t o 48 mesh, carrying 10% isoyuinoline as the stationary phase. Nitrogen was ased as the eluting gas. Absorber. T h e absorbing mixture was prepared by shaking together 3 parts by weight of concentrated sulfuric acid and 2 parts of Davison Grade TO silica gel of 60 to 200 mesh. By completely covering the surface of the silica gel with t h e large amount of acid, adsorption on the silica gel and the resulting tailing of peaks was eliminated. An absorber 20 mm. long and 4 mm. in diameter, operated a t 20’ to 50’ C., quantitatively removed mono-olefins, diolefins, cyclo-olefins, and acetylenes through at least eight carbons, except for ethylene and acetylene. Little ethylene or acetylene is present in olefinic naphthas, but if they must be determined, a n absorber containing sulfuric acid saturated with silver sulfate can be used. Although sulfuric acid react* ~th
P
The percentages of the conipoiicmts are calculated by comparing peak a r w s of each with the peak area of a reference compound added in known amount to the naphtha. Wherever an olefin and a paraffin are eluted together, thc area for the olefin is determined by difference after subtracting the area for the saturate as determined from the chroniatogram with the absorber. 2-Chloropropane was employed as reference compound; it is unaffected by the absorber, and is eluted between cyclopentane and n-hexane in the run with the absorber, as illustrated by the dashed peak in Figure 1. If two separate runs are made, the reference compound is omitted in the run without the absorber; a component that is conipletely resolved on both chromatograms, such as 2-methylpentane, is then used to relate peak areas in the two chromatograms.
W
3
c 3 a
I3 0
a
e E
I:
b W
WITHOUT
ABSORBER
0
c
8
0
w+ : o
20 Figure 1.
40 TIME, MINUTES
60
E
Chromatograms of catalytically cracked naphtha
many hydrocarbons, the absorber did not catalyze reactions or react with the saturated hydrocarbons; they passed through quantitatively. Also, the absorber did not promote the alkylation of such component,s as isobutane and 1-butene. The opportunity for one hydrocarbon to react u-ith another in the absorber is minimized by placing the absorber a t the tail of the column so t,httt most of the components are separ:ited before passage through it. Benzene also passed quantitatively through the absorber, although at least part of each of the highw aromatics would be irreversibly absorbed because of sulfonation. Silica gel is an excellent support for t'he sulfuric acid. It absorbs more than its weight of acid and still reniains free-flowing. Because olefins are incompletely removed when more than about l2'j& water is present in the sulfuric acid substrate, the mixture must be stored in a tight container. For some reason, absorbers prepared with Celite or firebrick in place of silica gel did not a l n x ~remove ~ the olefins quantitativelj..
the support is critical. With an isoquinoline column, the selectivity of the column increases as the concentration of liquid phase is increased-the retention times for olefins, aromatics, and cycloparaffins increase faster than those for paraffins as the percentage of the liquid phase on the support is increased. Some of these effects of concentration have been observed (4, 9). The optimum of 10% isoquinoline leads to a minimum of superposition of olefins on paraffins or cycloparaffins. Procedure. T h e subtraction technique can be applied in two ways. I n the first, separate runs a r e made with and without t h e absorber. I n the second, t h e chromatograms with and without t h e absorber are obtained in a single run by using two detectors in series at t h e end of t h e column n i t h t h e absorber placed between the detectors. The first detector senses all the components eluted from the column, and the second senses the saturates only. T o simplify the calculations, the two detectors are adjusted to equal response for a given concentration.
Columns. Even though t h e sulfuric x i d absorber simplifies t h e chromatogram, a n efficient' column is required. h 34-foot column made from '/+-inch copper tubing containing 10% b y weight of isoquinoline on firebrick (1 part of isoquinoline to 9 parts of firebrick) did the best over-all job of separating three- t o six-carbon compounds n-ithout excessive elution times. The percentage of liquid phase on
With either application of the technique, about 5 111. of naphtha is added to the isoquinoline column through a n introduction zone heated t o 100" C., and two chromatograms of the type shown in Figure 1 are obtained. The abbreviations for components are identified in Table I. The gas flow rate was 60 ml. per minute, and the column temperature was 26" C. After the sixcarbon components are eluted, the column is backflushed to remove the heavier hydrocariJons.
RESULTS AND DISCUSSION
Data are given in Table I1 for six replicate analyses of a known synthetic blend prepared from available compounds in the five- and six-carbon range. The error, or difference between amount prepared and found, averaged only 1.3% of the latter. The standard deviations averaged 1.6% of the amount found for all components except the three olefins determined by subtraction; deviations for these three averaged 2.6%. Table I gives analyses for three- to six-carbon components in a catalytically cracked naphtha, a coker naphtha, and a commercial gasoline-all derived from Mid-Continent crude oil. The three differ widely in the total and relative amounts of the various components. They did not contain detectable amounts of diolefins or acetylenes. Cyclohexene and 1-methylcyclopentene are not included because they are retained well into the seven-carbon region, where separation is incomplete. Three pairs and one trio of olefins are not separated in the isoquinoline column and are determined together. Most, if not all, of these components could be determined individually by making additional runs on columns with different selectivities. I n this manner, the values for cyclohexane and benzene were obtained with a l-chloronaphthallene column (9) and are included in Table I. The time of analysis was about 1.5 hours for the chromatographic run and 2 hours for the calculations. I n the catalytically cracked naphtha (3, I O ) , branched compounds predominate over straight-chain compounds, and more so for paraffins than for olefins. For the completely resolved components in this naphtha, standard deviations from six runs averaged 1.9% VOL. 32, NO. 3, MARCH 1960
337
of the amount found. For olefins incompletely resolved or determined by difference, the deviations are between 2 and 9% of the amount found. I n the coker naphtha, on the other
Table 1.
Analyses
hand, straight-chain compounds predominate over branched compounds. If the olefins were t o be selectively hydrogenated, the hydrocarbon distribution, as was previously pointed out
of Three Olefinic Naphthas for Components of Three to Six Carbons
(% by weight)
Component Propane Isobutane n-Butane Isopentane n-Pentane Cyclopentane 2,2-Dimethylbutane 2,3-Dimethylbutane 2-Methylpentane 3-Methylpentane n-Hexane Methylcyclopentane Cyclohexane0 Benzene" Propylene Isobutene 1-Butene trans-2-Butene cis-2-Butene 3-Methyl-1-butene 1-Pentene 2-Methyl-1-butene trans-2-Pentene cis-2-Pentene 2-Methyl-2-butene Cyclopentene 3,3-Dimethyl-l-butene 4-Methyl-1-pentene 3-Methyl-1-pentene 2,3-Dimethyl-l-butene cis-4-Me thyl-2-pentene trans-4-Methyl-2-pentene 2-Methyl-1-pentene 1-Hexene trans-3-Hexene cis-3-Hexene 2-Ethyl-1-butene
Abbreviation R 2MR nB 2MB nP CP 22DMB 23DMB 2MP 3MP nH MCP CH R1
;YR1> tB2 cB2 3MB1 P1 2MB1 tP2 cP2 2MB2 CP 1
4MPll 3MP11 23DMB1 c4MP2 t4MP2 2MP1 H1 tH3
1
E
l
f
Catalytically Cracked Naphtha Av. of Std. dev. 6 runs 0.27 0.01 0.96 0.02 0.53 0.02 2.62 0.03 0.64 0.01 0.12 0.01 0.02 0.01 0.02 0.51 0.02 1.70 0.02 1.07 0.02 0.67 0.02 0.96 0.24 0.56 0.42 0.01 1.27 0.02 0.01 0.82 0.01 0.68 0.12 0.01 0.01 0.34 0.02 1.10 0.03 1.05 0.02 0.63 0.04 2.50 0.01 0.22 0.0 0.02 0.22 0.02 0.23 0.12 0.01 0.35 0.01 0.59 0.03 0.24 0.02 0.62
0.16 MCP3 hlCP4 0.06 4Methylcyclopentene 0.69 trans-2-Hexene g+2) 2-Methyl-2-pentene 1.41 cis-2-Hexene 0.62 trans-3-Methyl-2-pentene t3MP2 1.16 c3MP2 cis-3-Methyl-2-pentene 0.41 23DMB2 2,3-Dimethyl-2-butene 26.9 Total a Determined with 1-chloronaphthalene column. 3-Methylc yclopentene
Table
Coker Commercial Naphtha, Gaso1in e, Av. of Av. of 3 Runs 3 Runs 0.02 0.02 0.54 0.06 4.31 0.56 5.41 1.05 5.92 3.04 0.50 0.40 0.17 0.02 0.44 0.12 1.81 3.13 0.94 1.90 3.51 4.82 1.29 1.30 0.51 0.48 1.67 0.21 0.0 0.0 0.23 0.28 0.21 0.09 0.20 0.08 0.19 0.13 0.90 0.48 0.49 0.55 0.52 0.63 0.28 0.32 1.13 0.64 0.24 0.33 0.0 0.0 0.61 0.25 3.13 0.12 0.09 0.08 0.22 0.15 0.23 0.55 2.11 0.40
0.04
0.35
0.36
0.01 0.01 0.05 0.07 0.07 0.02 0.03
0.20 0.09 0.53 0.89 0.23 0.42 0.19 25.2
0.14 0.04 0.28 0.76 0.28 0.36 0.20 37.0
II. Analysis of Synthetic Blend
(Average of six determinations, yo by weight) Prepared Found 12.88 13.08 n-Pentane 10.45 10.62 1-Pentene* 2.23 2.20 2,2-Dimethylbutane 9.98 10.08 cis-2-Pentene 14.14 13.92 2-Methyl-2-butenen 6.05 6.10 2,3-Dimethylbutane 12.33 12.44 3-Methylpentane 11.74 11.67 1-Hexene 7.92 8.11 2-Methyl-2-pentene' 11.86 11.78 Methylcyclopentane 99.58 100.00 Determined by subtraction. 0
338
ANALYTICAL CHEMISTRY
Std. Dev. 0.18 0.25 0.05 0.15 0.28 0.10 0.17 0.18 0.24 0.18
for a thermally cracked naphtha (3), would resemble that in the virgin naphtha. The degree of branching is, however, somewhat less in the coker naphtha than in the virgin naphtha. [An analysis of the virgin naphtha is found in Table I, ( 9 ) . ] The isomer distribution of the commercial gasoline differs from that of the catalytic and coker naphthas in two major respects. Individual olefins are usually intermediate in amount between those present in the other two naphthas, and the amounts of butanes and pentanes are much higher than in either. It can, therefore, be surmised that the gasoline is a mixture of a catalytically cracked naphtha, and a thermally cracked naphtha like coker naphtha, fortified with a lighter saturated naphtha. CONCLUSION
Extension of the subtraction technique in several ways should be possible. Direct extension to seven and more carbons would necessitate g r o u p ing more of the isomers and would have greater interference from aromatics, Runs on additional columns of varying selectivities would minimize the groupings. Interference from aromatics could be eliminated by using a stationary phase that would selectively retain them beyond snturates and olefins. Other complex mixtures that, like olefinic naphthas, have resisted analysis may be made to yield to a similar subtraction technique. LITERATURE CITED
(1) Bloch, M. G., Second Symposium on Gas Chromatography, Instrument Society of America, East Lansing, Mich., June 11, 1959. (2) Bradford, B. IT'., Harvey, D., Chalkley, D. E., J . Inst. Petrol. 41, 80 (1955). (3) Cady, Itr. E., Marschner, R. F., Cropper, W. P., Ind. Eng. Chem. 44, 1859 (1952). (4) Eggertsen, F. T., Knight, H. S., ANAL.CHEM.30, 15 (1958). (5) Farbenfabriken Bayer AG, Belg. Patent 560,908 (Jan. 15, 1958). (6) Keulmans, A. I. M., Kwantes, A., Zaal, P., Anal. Chim. .Acta 13, 357 (1055). (7) Knight, H. s., AN.4L. CHEhl. 30, 9 ' (19587. ' (8) Lichtenfels, D. H., Fleck, S. 9., Burow, F. H., Coggeshall, N. D , Zbid., 28, 1376 (1956). (9) Martin, R. L., W'inters, J. C., Zbid., 31,1954 (1959). (10) Melpolder, F. W.,Brown, R. A., Young, W. S., Headington, C. E., Ind. Eng. ChenL. 44,1142 (1952). (11) Nerheim, A. G., patent applied for. (12) Tenney, H. If.,ANAL. CHEM.30, 2 (1958). (13) Van de Craats, F., Anal. Chim. Acta 14, 136 (1956).
RECEIVED for review August io, 1959. Accepted October 28, 1959. Pittsburgh Conference on Analytical Chemistry and Applied Spectroscopy, March 1959.
